2005-01-06 01:43:34 +00:00
|
|
|
/*-
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
* Copyright (c) 2005-2009 Ariff Abdullah <ariff@FreeBSD.org>
|
|
|
|
* Copyright (c) 1999 Cameron Grant <cg@FreeBSD.org>
|
1999-09-01 04:08:39 +00:00
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
* are met:
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
*/
|
|
|
|
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
#ifdef HAVE_KERNEL_OPTION_HEADERS
|
|
|
|
#include "opt_snd.h"
|
|
|
|
#endif
|
|
|
|
|
1999-11-20 16:50:33 +00:00
|
|
|
#include <dev/sound/pcm/sound.h>
|
1999-09-01 04:08:39 +00:00
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
#include "feeder_if.h"
|
|
|
|
|
2001-08-23 11:30:52 +00:00
|
|
|
SND_DECLARE_FILE("$FreeBSD$");
|
|
|
|
|
2011-11-07 06:44:47 +00:00
|
|
|
static MALLOC_DEFINE(M_FEEDER, "feeder", "pcm feeder");
|
2000-12-18 01:36:41 +00:00
|
|
|
|
2000-09-23 22:11:32 +00:00
|
|
|
#define MAXFEEDERS 256
|
2000-09-09 21:24:03 +00:00
|
|
|
#undef FEEDER_DEBUG
|
1999-09-01 04:08:39 +00:00
|
|
|
|
2000-08-20 22:18:56 +00:00
|
|
|
struct feedertab_entry {
|
|
|
|
SLIST_ENTRY(feedertab_entry) link;
|
2000-12-18 01:36:41 +00:00
|
|
|
struct feeder_class *feederclass;
|
2000-08-20 22:18:56 +00:00
|
|
|
struct pcm_feederdesc *desc;
|
|
|
|
|
|
|
|
int idx;
|
|
|
|
};
|
|
|
|
static SLIST_HEAD(, feedertab_entry) feedertab;
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
|
|
void
|
|
|
|
feeder_register(void *p)
|
|
|
|
{
|
2002-01-26 22:13:24 +00:00
|
|
|
static int feedercnt = 0;
|
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
struct feeder_class *fc = p;
|
2000-08-20 22:18:56 +00:00
|
|
|
struct feedertab_entry *fte;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (feedercnt == 0) {
|
2002-01-26 22:13:24 +00:00
|
|
|
KASSERT(fc->desc == NULL, ("first feeder not root: %s", fc->name));
|
|
|
|
|
2000-08-20 22:18:56 +00:00
|
|
|
SLIST_INIT(&feedertab);
|
2003-03-05 14:48:28 +00:00
|
|
|
fte = malloc(sizeof(*fte), M_FEEDER, M_NOWAIT | M_ZERO);
|
2002-01-26 22:13:24 +00:00
|
|
|
if (fte == NULL) {
|
2002-05-25 11:18:03 +00:00
|
|
|
printf("can't allocate memory for root feeder: %s\n",
|
|
|
|
fc->name);
|
2002-01-26 22:13:24 +00:00
|
|
|
|
|
|
|
return;
|
|
|
|
}
|
2000-12-18 01:36:41 +00:00
|
|
|
fte->feederclass = fc;
|
2000-08-20 22:18:56 +00:00
|
|
|
fte->desc = NULL;
|
|
|
|
fte->idx = feedercnt;
|
|
|
|
SLIST_INSERT_HEAD(&feedertab, fte, link);
|
|
|
|
feedercnt++;
|
2002-01-26 22:13:24 +00:00
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
/* initialize global variables */
|
|
|
|
|
Last major commit and updates for RELENG_7:
- Rework the entire pcm_channel structure:
* Remove rarely used link placeholder, instead, make each pcm_channel
as head/link of each own/each other. Unlock - Lock sequence due to
sleep malloc has been reduced.
* Implement "busy" queue which will contain list of busy/active
channels. This greatly reduce locking contention for example while
servicing interrupt for hardware with many channels or when virtual
channels reach its 256 peak channels.
- So I heard you like v chan ... O RLY?
Welcome to Virtual **Record** Channels (vrec, rec vchans, vchans for
recording, Rec-Chan, you decide), the ultimate solutions for your
nagging O_RDWR full-duplex wannabe (note: flash plugins) monopolizing
single record channel causing EBUSY. Vrec works exactly like Vchans
(or, should I rename it to "Vplay" :) , except that it operates on the
opposite direction (recording). Up to 256 vrecs (like vchans) are
possible.
Notes:
* Relocate dev.pcm.%d.{vchans,vchanformat,vchanrate} to each of its
respective node/direction:
dev.pcm.%d.play.* for "play" (cdev = dsp%d.vp%d)
dev.pcm.%d.rec.* for "record" (cdev = dsp%d.vr%d)
* Don't expect that it will magically give you ability to split
"recording source" (eg: 1 channel for cdrom, 1 channel for mic,
etc). Just admit that you only have a *single* recording source /
channel. Please bug your hardware vendor instead :)
- Bump maxautovchans from 4 to 16. For a full-fledged multimedia
desktop/workstation with too many soundservers installed (esound,
artsd, jackd, pulse/polypaudio, ding-dong pling plong mudkip fuh fuh,
etc), 4 seems inadequate. There will be no memory penalty here, since
virtual channels are allocate only by demand.
- Nuke/Rework the entire statically created cdev entries. Everything is
clonable through snd own clone manager which designed to withstand many
kind of abusive devfs droids such as:
* while : ; do /bin/test -e /dev/dsp ; done
* jot 16777216 0 | while read x ; do ls /dev/dsp0.$x ; done
* hundreds (could be thousands) concurrent threads/process opening
"/dev/dsp" (previously, this might result EBUSY even with just
3 contesting threads/procs).
o Reusable clone objects (instead of creating new one like there's no
tomorrow) after certain expiration deadline. The clone allocator will
decide whether to reuse, share, or creating new clone.
o Automatic garbage collector.
- Dynamic unit magic allocator. Maximum attached soundcards can be tuned
using tunable "hw.snd.maxunit" (Default to 512). Minimum is 16, and
maximum is 2048.
- ..other fixes, mostly related to concurrency issues.
joel@ will do the manpage updates on sound(4).
Have fun.
2007-05-31 18:43:33 +00:00
|
|
|
if (snd_verbose < 0 || snd_verbose > 4)
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
snd_verbose = 1;
|
|
|
|
|
Last major commit and updates for RELENG_7:
- Rework the entire pcm_channel structure:
* Remove rarely used link placeholder, instead, make each pcm_channel
as head/link of each own/each other. Unlock - Lock sequence due to
sleep malloc has been reduced.
* Implement "busy" queue which will contain list of busy/active
channels. This greatly reduce locking contention for example while
servicing interrupt for hardware with many channels or when virtual
channels reach its 256 peak channels.
- So I heard you like v chan ... O RLY?
Welcome to Virtual **Record** Channels (vrec, rec vchans, vchans for
recording, Rec-Chan, you decide), the ultimate solutions for your
nagging O_RDWR full-duplex wannabe (note: flash plugins) monopolizing
single record channel causing EBUSY. Vrec works exactly like Vchans
(or, should I rename it to "Vplay" :) , except that it operates on the
opposite direction (recording). Up to 256 vrecs (like vchans) are
possible.
Notes:
* Relocate dev.pcm.%d.{vchans,vchanformat,vchanrate} to each of its
respective node/direction:
dev.pcm.%d.play.* for "play" (cdev = dsp%d.vp%d)
dev.pcm.%d.rec.* for "record" (cdev = dsp%d.vr%d)
* Don't expect that it will magically give you ability to split
"recording source" (eg: 1 channel for cdrom, 1 channel for mic,
etc). Just admit that you only have a *single* recording source /
channel. Please bug your hardware vendor instead :)
- Bump maxautovchans from 4 to 16. For a full-fledged multimedia
desktop/workstation with too many soundservers installed (esound,
artsd, jackd, pulse/polypaudio, ding-dong pling plong mudkip fuh fuh,
etc), 4 seems inadequate. There will be no memory penalty here, since
virtual channels are allocate only by demand.
- Nuke/Rework the entire statically created cdev entries. Everything is
clonable through snd own clone manager which designed to withstand many
kind of abusive devfs droids such as:
* while : ; do /bin/test -e /dev/dsp ; done
* jot 16777216 0 | while read x ; do ls /dev/dsp0.$x ; done
* hundreds (could be thousands) concurrent threads/process opening
"/dev/dsp" (previously, this might result EBUSY even with just
3 contesting threads/procs).
o Reusable clone objects (instead of creating new one like there's no
tomorrow) after certain expiration deadline. The clone allocator will
decide whether to reuse, share, or creating new clone.
o Automatic garbage collector.
- Dynamic unit magic allocator. Maximum attached soundcards can be tuned
using tunable "hw.snd.maxunit" (Default to 512). Minimum is 16, and
maximum is 2048.
- ..other fixes, mostly related to concurrency issues.
joel@ will do the manpage updates on sound(4).
Have fun.
2007-05-31 18:43:33 +00:00
|
|
|
/* initialize unit numbering */
|
|
|
|
snd_unit_init();
|
|
|
|
if (snd_unit < 0 || snd_unit > PCMMAXUNIT)
|
2007-06-17 15:53:11 +00:00
|
|
|
snd_unit = -1;
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
|
|
|
|
if (snd_maxautovchans < 0 ||
|
|
|
|
snd_maxautovchans > SND_MAXVCHANS)
|
|
|
|
snd_maxautovchans = 0;
|
|
|
|
|
|
|
|
if (chn_latency < CHN_LATENCY_MIN ||
|
|
|
|
chn_latency > CHN_LATENCY_MAX)
|
|
|
|
chn_latency = CHN_LATENCY_DEFAULT;
|
|
|
|
|
|
|
|
if (chn_latency_profile < CHN_LATENCY_PROFILE_MIN ||
|
|
|
|
chn_latency_profile > CHN_LATENCY_PROFILE_MAX)
|
|
|
|
chn_latency_profile = CHN_LATENCY_PROFILE_DEFAULT;
|
|
|
|
|
|
|
|
if (feeder_rate_min < FEEDRATE_MIN ||
|
|
|
|
feeder_rate_max < FEEDRATE_MIN ||
|
|
|
|
feeder_rate_min > FEEDRATE_MAX ||
|
|
|
|
feeder_rate_max > FEEDRATE_MAX ||
|
|
|
|
!(feeder_rate_min < feeder_rate_max)) {
|
|
|
|
feeder_rate_min = FEEDRATE_RATEMIN;
|
|
|
|
feeder_rate_max = FEEDRATE_RATEMAX;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (feeder_rate_round < FEEDRATE_ROUNDHZ_MIN ||
|
|
|
|
feeder_rate_round > FEEDRATE_ROUNDHZ_MAX)
|
|
|
|
feeder_rate_round = FEEDRATE_ROUNDHZ;
|
|
|
|
|
|
|
|
if (bootverbose)
|
|
|
|
printf("%s: snd_unit=%d snd_maxautovchans=%d "
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
"latency=%d "
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
"feeder_rate_min=%d feeder_rate_max=%d "
|
|
|
|
"feeder_rate_round=%d\n",
|
|
|
|
__func__, snd_unit, snd_maxautovchans,
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
chn_latency,
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
feeder_rate_min, feeder_rate_max,
|
|
|
|
feeder_rate_round);
|
|
|
|
|
2002-01-26 22:13:24 +00:00
|
|
|
/* we've got our root feeder so don't veto pcm loading anymore */
|
|
|
|
pcm_veto_load = 0;
|
|
|
|
|
2000-08-20 22:18:56 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2002-01-26 22:13:24 +00:00
|
|
|
KASSERT(fc->desc != NULL, ("feeder '%s' has no descriptor", fc->name));
|
|
|
|
|
|
|
|
/* beyond this point failure is non-fatal but may result in some translations being unavailable */
|
2000-08-20 22:18:56 +00:00
|
|
|
i = 0;
|
2000-12-18 01:36:41 +00:00
|
|
|
while ((feedercnt < MAXFEEDERS) && (fc->desc[i].type > 0)) {
|
2001-04-09 21:33:47 +00:00
|
|
|
/* printf("adding feeder %s, %x -> %x\n", fc->name, fc->desc[i].in, fc->desc[i].out); */
|
2003-03-05 14:48:28 +00:00
|
|
|
fte = malloc(sizeof(*fte), M_FEEDER, M_NOWAIT | M_ZERO);
|
2002-01-26 22:13:24 +00:00
|
|
|
if (fte == NULL) {
|
|
|
|
printf("can't allocate memory for feeder '%s', %x -> %x\n", fc->name, fc->desc[i].in, fc->desc[i].out);
|
|
|
|
|
|
|
|
return;
|
|
|
|
}
|
2000-12-18 01:36:41 +00:00
|
|
|
fte->feederclass = fc;
|
|
|
|
fte->desc = &fc->desc[i];
|
2000-08-20 22:18:56 +00:00
|
|
|
fte->idx = feedercnt;
|
|
|
|
fte->desc->idx = feedercnt;
|
|
|
|
SLIST_INSERT_HEAD(&feedertab, fte, link);
|
|
|
|
i++;
|
|
|
|
}
|
|
|
|
feedercnt++;
|
|
|
|
if (feedercnt >= MAXFEEDERS)
|
2002-01-26 22:13:24 +00:00
|
|
|
printf("MAXFEEDERS (%d >= %d) exceeded\n", feedercnt, MAXFEEDERS);
|
2000-08-20 22:18:56 +00:00
|
|
|
}
|
|
|
|
|
2001-03-16 20:58:44 +00:00
|
|
|
static void
|
|
|
|
feeder_unregisterall(void *p)
|
|
|
|
{
|
|
|
|
struct feedertab_entry *fte, *next;
|
|
|
|
|
|
|
|
next = SLIST_FIRST(&feedertab);
|
|
|
|
while (next != NULL) {
|
|
|
|
fte = next;
|
|
|
|
next = SLIST_NEXT(fte, link);
|
|
|
|
free(fte, M_FEEDER);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
1999-09-01 04:08:39 +00:00
|
|
|
static int
|
2000-08-20 22:18:56 +00:00
|
|
|
cmpdesc(struct pcm_feederdesc *n, struct pcm_feederdesc *m)
|
1999-09-01 04:08:39 +00:00
|
|
|
{
|
2000-09-23 22:11:32 +00:00
|
|
|
return ((n->type == m->type) &&
|
|
|
|
((n->in == 0) || (n->in == m->in)) &&
|
|
|
|
((n->out == 0) || (n->out == m->out)) &&
|
|
|
|
(n->flags == m->flags));
|
1999-09-01 04:08:39 +00:00
|
|
|
}
|
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
static void
|
2001-03-24 23:10:29 +00:00
|
|
|
feeder_destroy(struct pcm_feeder *f)
|
1999-09-01 04:08:39 +00:00
|
|
|
{
|
2000-12-18 01:36:41 +00:00
|
|
|
FEEDER_FREE(f);
|
|
|
|
kobj_delete((kobj_t)f, M_FEEDER);
|
|
|
|
}
|
1999-09-01 04:08:39 +00:00
|
|
|
|
2001-03-24 23:10:29 +00:00
|
|
|
static struct pcm_feeder *
|
2000-12-18 01:36:41 +00:00
|
|
|
feeder_create(struct feeder_class *fc, struct pcm_feederdesc *desc)
|
|
|
|
{
|
2001-03-24 23:10:29 +00:00
|
|
|
struct pcm_feeder *f;
|
2000-12-18 01:36:41 +00:00
|
|
|
int err;
|
|
|
|
|
2003-03-05 14:48:28 +00:00
|
|
|
f = (struct pcm_feeder *)kobj_create((kobj_class_t)fc, M_FEEDER, M_NOWAIT | M_ZERO);
|
2002-01-26 22:13:24 +00:00
|
|
|
if (f == NULL)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
f->data = fc->data;
|
|
|
|
f->source = NULL;
|
|
|
|
f->parent = NULL;
|
|
|
|
f->class = fc;
|
|
|
|
f->desc = &(f->desc_static);
|
|
|
|
|
|
|
|
if (desc) {
|
2000-12-18 01:36:41 +00:00
|
|
|
*(f->desc) = *desc;
|
2002-01-26 22:13:24 +00:00
|
|
|
} else {
|
2000-12-18 01:36:41 +00:00
|
|
|
f->desc->type = FEEDER_ROOT;
|
|
|
|
f->desc->in = 0;
|
|
|
|
f->desc->out = 0;
|
|
|
|
f->desc->flags = 0;
|
|
|
|
f->desc->idx = 0;
|
2000-08-20 22:18:56 +00:00
|
|
|
}
|
2002-01-26 22:13:24 +00:00
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
err = FEEDER_INIT(f);
|
|
|
|
if (err) {
|
2001-04-08 20:20:52 +00:00
|
|
|
printf("feeder_init(%p) on %s returned %d\n", f, fc->name, err);
|
2000-12-18 01:36:41 +00:00
|
|
|
feeder_destroy(f);
|
2002-01-26 22:13:24 +00:00
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
return NULL;
|
2002-01-26 22:13:24 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return f;
|
1999-09-01 04:08:39 +00:00
|
|
|
}
|
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
struct feeder_class *
|
|
|
|
feeder_getclass(struct pcm_feederdesc *desc)
|
1999-09-01 04:08:39 +00:00
|
|
|
{
|
2000-08-20 22:18:56 +00:00
|
|
|
struct feedertab_entry *fte;
|
|
|
|
|
|
|
|
SLIST_FOREACH(fte, &feedertab, link) {
|
2000-12-18 01:36:41 +00:00
|
|
|
if ((desc == NULL) && (fte->desc == NULL))
|
|
|
|
return fte->feederclass;
|
|
|
|
if ((fte->desc != NULL) && (desc != NULL) && cmpdesc(desc, fte->desc))
|
|
|
|
return fte->feederclass;
|
2000-08-20 22:18:56 +00:00
|
|
|
}
|
|
|
|
return NULL;
|
1999-09-01 04:08:39 +00:00
|
|
|
}
|
|
|
|
|
2000-09-23 22:11:32 +00:00
|
|
|
int
|
2001-03-24 23:10:29 +00:00
|
|
|
chn_addfeeder(struct pcm_channel *c, struct feeder_class *fc, struct pcm_feederdesc *desc)
|
2000-09-23 22:11:32 +00:00
|
|
|
{
|
2001-03-24 23:10:29 +00:00
|
|
|
struct pcm_feeder *nf;
|
2000-09-23 22:11:32 +00:00
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
nf = feeder_create(fc, desc);
|
|
|
|
if (nf == NULL)
|
2002-01-26 22:13:24 +00:00
|
|
|
return ENOSPC;
|
2000-12-18 01:36:41 +00:00
|
|
|
|
2000-09-23 22:11:32 +00:00
|
|
|
nf->source = c->feeder;
|
2000-12-18 01:36:41 +00:00
|
|
|
|
2003-07-07 17:40:02 +00:00
|
|
|
if (c->feeder != NULL)
|
|
|
|
c->feeder->parent = nf;
|
2000-09-23 22:11:32 +00:00
|
|
|
c->feeder = nf;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
1999-09-01 04:08:39 +00:00
|
|
|
int
|
2001-03-24 23:10:29 +00:00
|
|
|
chn_removefeeder(struct pcm_channel *c)
|
1999-09-01 04:08:39 +00:00
|
|
|
{
|
2001-03-24 23:10:29 +00:00
|
|
|
struct pcm_feeder *f;
|
1999-09-01 04:08:39 +00:00
|
|
|
|
2000-12-18 01:36:41 +00:00
|
|
|
if (c->feeder == NULL)
|
2000-08-20 22:18:56 +00:00
|
|
|
return -1;
|
2000-12-18 01:36:41 +00:00
|
|
|
f = c->feeder;
|
|
|
|
c->feeder = c->feeder->source;
|
|
|
|
feeder_destroy(f);
|
2002-01-26 22:13:24 +00:00
|
|
|
|
2000-08-20 22:18:56 +00:00
|
|
|
return 0;
|
1999-09-01 04:08:39 +00:00
|
|
|
}
|
|
|
|
|
2001-03-24 23:10:29 +00:00
|
|
|
struct pcm_feeder *
|
|
|
|
chn_findfeeder(struct pcm_channel *c, u_int32_t type)
|
2000-09-23 22:11:32 +00:00
|
|
|
{
|
2001-03-24 23:10:29 +00:00
|
|
|
struct pcm_feeder *f;
|
2000-09-23 22:11:32 +00:00
|
|
|
|
|
|
|
f = c->feeder;
|
|
|
|
while (f != NULL) {
|
|
|
|
if (f->desc->type == type)
|
|
|
|
return f;
|
|
|
|
f = f->source;
|
|
|
|
}
|
2002-01-26 22:13:24 +00:00
|
|
|
|
2000-09-23 22:11:32 +00:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
/*
|
|
|
|
* 14bit format scoring
|
|
|
|
* --------------------
|
|
|
|
*
|
|
|
|
* 13 12 11 10 9 8 2 1 0 offset
|
|
|
|
* +---+---+---+---+---+---+-------------+---+---+
|
|
|
|
* | X | X | X | X | X | X | X X X X X X | X | X |
|
|
|
|
* +---+---+---+---+---+---+-------------+---+---+
|
|
|
|
* | | | | | | | | |
|
|
|
|
* | | | | | | | | +--> signed?
|
|
|
|
* | | | | | | | |
|
|
|
|
* | | | | | | | +------> bigendian?
|
|
|
|
* | | | | | | |
|
|
|
|
* | | | | | | +---------------> total channels
|
|
|
|
* | | | | | |
|
|
|
|
* | | | | | +------------------------> AFMT_A_LAW
|
|
|
|
* | | | | |
|
|
|
|
* | | | | +----------------------------> AFMT_MU_LAW
|
|
|
|
* | | | |
|
|
|
|
* | | | +--------------------------------> AFMT_8BIT
|
|
|
|
* | | |
|
|
|
|
* | | +------------------------------------> AFMT_16BIT
|
|
|
|
* | |
|
|
|
|
* | +----------------------------------------> AFMT_24BIT
|
|
|
|
* |
|
|
|
|
* +--------------------------------------------> AFMT_32BIT
|
|
|
|
*/
|
|
|
|
#define score_signeq(s1, s2) (((s1) & 0x1) == ((s2) & 0x1))
|
|
|
|
#define score_endianeq(s1, s2) (((s1) & 0x2) == ((s2) & 0x2))
|
|
|
|
#define score_cheq(s1, s2) (((s1) & 0xfc) == ((s2) & 0xfc))
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
#define score_chgt(s1, s2) (((s1) & 0xfc) > ((s2) & 0xfc))
|
|
|
|
#define score_chlt(s1, s2) (((s1) & 0xfc) < ((s2) & 0xfc))
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
#define score_val(s1) ((s1) & 0x3f00)
|
|
|
|
#define score_cse(s1) ((s1) & 0x7f)
|
1999-09-01 04:08:39 +00:00
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
u_int32_t
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
snd_fmtscore(u_int32_t fmt)
|
2006-01-22 15:06:49 +00:00
|
|
|
{
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
u_int32_t ret;
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
if (fmt & AFMT_SIGNED)
|
|
|
|
ret |= 1 << 0;
|
|
|
|
if (fmt & AFMT_BIGENDIAN)
|
|
|
|
ret |= 1 << 1;
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
/*if (fmt & AFMT_STEREO)
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
ret |= (2 & 0x3f) << 2;
|
|
|
|
else
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
ret |= (1 & 0x3f) << 2;*/
|
|
|
|
ret |= (AFMT_CHANNEL(fmt) & 0x3f) << 2;
|
2006-01-29 01:32:37 +00:00
|
|
|
if (fmt & AFMT_A_LAW)
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
ret |= 1 << 8;
|
|
|
|
else if (fmt & AFMT_MU_LAW)
|
|
|
|
ret |= 1 << 9;
|
|
|
|
else if (fmt & AFMT_8BIT)
|
|
|
|
ret |= 1 << 10;
|
|
|
|
else if (fmt & AFMT_16BIT)
|
|
|
|
ret |= 1 << 11;
|
|
|
|
else if (fmt & AFMT_24BIT)
|
|
|
|
ret |= 1 << 12;
|
|
|
|
else if (fmt & AFMT_32BIT)
|
|
|
|
ret |= 1 << 13;
|
|
|
|
|
|
|
|
return ret;
|
2006-01-22 15:06:49 +00:00
|
|
|
}
|
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
static u_int32_t
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
snd_fmtbestfunc(u_int32_t fmt, u_int32_t *fmts, int cheq)
|
2006-01-22 15:06:49 +00:00
|
|
|
{
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
u_int32_t best, score, score2, oldscore;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (fmt == 0 || fmts == NULL || fmts[0] == 0)
|
|
|
|
return 0;
|
|
|
|
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
if (snd_fmtvalid(fmt, fmts))
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
return fmt;
|
2006-01-22 15:06:49 +00:00
|
|
|
|
|
|
|
best = 0;
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
score = snd_fmtscore(fmt);
|
2006-01-22 15:06:49 +00:00
|
|
|
oldscore = 0;
|
|
|
|
for (i = 0; fmts[i] != 0; i++) {
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
score2 = snd_fmtscore(fmts[i]);
|
|
|
|
if (cheq && !score_cheq(score, score2) &&
|
|
|
|
(score_chlt(score2, score) ||
|
|
|
|
(oldscore != 0 && score_chgt(score2, oldscore))))
|
|
|
|
continue;
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
if (oldscore == 0 ||
|
|
|
|
(score_val(score2) == score_val(score)) ||
|
|
|
|
(score_val(score2) == score_val(oldscore)) ||
|
|
|
|
(score_val(score2) > score_val(oldscore) &&
|
|
|
|
score_val(score2) < score_val(score)) ||
|
|
|
|
(score_val(score2) < score_val(oldscore) &&
|
|
|
|
score_val(score2) > score_val(score)) ||
|
|
|
|
(score_val(oldscore) < score_val(score) &&
|
|
|
|
score_val(score2) > score_val(oldscore))) {
|
|
|
|
if (score_val(oldscore) != score_val(score2) ||
|
|
|
|
score_cse(score) == score_cse(score2) ||
|
|
|
|
((score_cse(oldscore) != score_cse(score) &&
|
|
|
|
!score_endianeq(score, oldscore) &&
|
|
|
|
(score_endianeq(score, score2) ||
|
|
|
|
(!score_signeq(score, oldscore) &&
|
|
|
|
score_signeq(score, score2)))))) {
|
|
|
|
best = fmts[i];
|
|
|
|
oldscore = score2;
|
|
|
|
}
|
2006-01-22 15:06:49 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return best;
|
|
|
|
}
|
|
|
|
|
|
|
|
u_int32_t
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
snd_fmtbestbit(u_int32_t fmt, u_int32_t *fmts)
|
2006-01-22 15:06:49 +00:00
|
|
|
{
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
return snd_fmtbestfunc(fmt, fmts, 0);
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
}
|
2006-01-22 15:06:49 +00:00
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
u_int32_t
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
snd_fmtbestchannel(u_int32_t fmt, u_int32_t *fmts)
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
{
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
return snd_fmtbestfunc(fmt, fmts, 1);
|
2006-01-22 15:06:49 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
u_int32_t
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
snd_fmtbest(u_int32_t fmt, u_int32_t *fmts)
|
2006-01-22 15:06:49 +00:00
|
|
|
{
|
|
|
|
u_int32_t best1, best2;
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
u_int32_t score, score1, score2;
|
|
|
|
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
if (snd_fmtvalid(fmt, fmts))
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
return fmt;
|
2006-01-22 15:06:49 +00:00
|
|
|
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
best1 = snd_fmtbestchannel(fmt, fmts);
|
|
|
|
best2 = snd_fmtbestbit(fmt, fmts);
|
2006-01-22 15:06:49 +00:00
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
if (best1 != 0 && best2 != 0 && best1 != best2) {
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
/*if (fmt & AFMT_STEREO)*/
|
|
|
|
if (AFMT_CHANNEL(fmt) > 1)
|
2006-01-22 15:06:49 +00:00
|
|
|
return best1;
|
|
|
|
else {
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
score = score_val(snd_fmtscore(fmt));
|
|
|
|
score1 = score_val(snd_fmtscore(best1));
|
|
|
|
score2 = score_val(snd_fmtscore(best2));
|
2006-01-29 01:32:37 +00:00
|
|
|
if (score1 == score2 || score1 == score)
|
|
|
|
return best1;
|
|
|
|
else if (score2 == score)
|
2006-01-22 15:06:49 +00:00
|
|
|
return best2;
|
2006-01-29 01:32:37 +00:00
|
|
|
else if (score1 > score2)
|
2006-01-22 15:06:49 +00:00
|
|
|
return best1;
|
|
|
|
return best2;
|
|
|
|
}
|
|
|
|
} else if (best2 == 0)
|
|
|
|
return best1;
|
2006-02-23 19:23:55 +00:00
|
|
|
else
|
2006-01-22 15:06:49 +00:00
|
|
|
return best2;
|
|
|
|
}
|
|
|
|
|
2003-07-07 17:40:02 +00:00
|
|
|
void
|
|
|
|
feeder_printchain(struct pcm_feeder *head)
|
|
|
|
{
|
|
|
|
struct pcm_feeder *f;
|
|
|
|
|
|
|
|
printf("feeder chain (head @%p)\n", head);
|
|
|
|
f = head;
|
|
|
|
while (f != NULL) {
|
|
|
|
printf("%s/%d @ %p\n", f->class->name, f->desc->idx, f);
|
|
|
|
f = f->source;
|
|
|
|
}
|
|
|
|
printf("[end]\n\n");
|
|
|
|
}
|
|
|
|
|
2000-09-23 22:11:32 +00:00
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
|
|
static int
|
2001-03-24 23:10:29 +00:00
|
|
|
feed_root(struct pcm_feeder *feeder, struct pcm_channel *ch, u_int8_t *buffer, u_int32_t count, void *source)
|
2000-09-23 22:11:32 +00:00
|
|
|
{
|
2001-03-24 23:10:29 +00:00
|
|
|
struct snd_dbuf *src = source;
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
int l, offset;
|
2000-09-23 22:11:32 +00:00
|
|
|
|
2001-03-24 23:10:29 +00:00
|
|
|
KASSERT(count > 0, ("feed_root: count == 0"));
|
2000-09-23 22:11:32 +00:00
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
if (++ch->feedcount == 0)
|
|
|
|
ch->feedcount = 2;
|
|
|
|
|
2001-03-24 23:10:29 +00:00
|
|
|
l = min(count, sndbuf_getready(src));
|
|
|
|
|
2003-02-03 15:02:28 +00:00
|
|
|
/* When recording only return as much data as available */
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
if (ch->direction == PCMDIR_REC) {
|
|
|
|
sndbuf_dispose(src, buffer, l);
|
2003-02-03 15:02:28 +00:00
|
|
|
return l;
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
}
|
2003-02-03 15:02:28 +00:00
|
|
|
|
2001-03-24 23:10:29 +00:00
|
|
|
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
offset = count - l;
|
|
|
|
|
|
|
|
if (offset > 0) {
|
|
|
|
if (snd_verbose > 3)
|
|
|
|
printf("%s: (%s) %spending %d bytes "
|
|
|
|
"(count=%d l=%d feed=%d)\n",
|
|
|
|
__func__,
|
|
|
|
(ch->flags & CHN_F_VIRTUAL) ? "virtual" : "hardware",
|
|
|
|
(ch->feedcount == 1) ? "pre" : "ap",
|
|
|
|
offset, count, l, ch->feedcount);
|
|
|
|
|
|
|
|
if (ch->feedcount == 1) {
|
2007-03-16 17:15:33 +00:00
|
|
|
memset(buffer,
|
|
|
|
sndbuf_zerodata(sndbuf_getfmt(src)),
|
|
|
|
offset);
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
if (l > 0)
|
|
|
|
sndbuf_dispose(src, buffer + offset, l);
|
|
|
|
else
|
|
|
|
ch->feedcount--;
|
|
|
|
} else {
|
|
|
|
if (l > 0)
|
|
|
|
sndbuf_dispose(src, buffer, l);
|
2007-03-16 17:15:33 +00:00
|
|
|
memset(buffer + l,
|
|
|
|
sndbuf_zerodata(sndbuf_getfmt(src)),
|
|
|
|
offset);
|
|
|
|
if (!(ch->flags & CHN_F_CLOSING))
|
|
|
|
ch->xruns++;
|
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes
in every sense.
General
-------
- Multichannel safe, endian safe, format safe
* Large part of critical pcm filters such as vchan.c, feeder_rate.c,
feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that
using them does not cause the pcm data to be converted to 16bit little
endian.
* Macrosses for accessing pcm data safely are defined within sound.h in
the form of PCM_READ_* / PCM_WRITE_*
* Currently, most of them are probably limited for mono/stereo handling,
but the future addition of true multichannel will be much easier.
- Low latency operation
* Well, this require lot more works to do not just within sound driver,
but we're heading towards right direction. Buffer/block sizing within
channel.c is rewritten to calculate precise allocation for various
combination of sample/data/rate size. As a result, applying correct
SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar
to what commercial 4front driver do.
* Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not
result long delay.
* Eliminate sound truncation if the sound data is too small.
DIY:
1) Download / extract
http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz
2) Do a comparison between "cat state*.au > /dev/dsp" and
"for x in state*.au ; do cat $x > /dev/dsp ; done"
- there should be no "perceivable" differences.
Double close for PR kern/31445.
CAVEAT: Low latency come with (unbearable) price especially for poorly
written applications. Applications that trying to act smarter
by requesting (wrong) blocksize/blockcount will suffer the most.
Fixup samples/patches can be found at:
http://people.freebsd.org/~ariff/ports/
- Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42)
due to closer compatibility with 4front driver.
Discussed with: marcus@ (long time ago?)
- All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been
moved to their own dev sysctl nodes, notably:
hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans
Bump __FreeBSD_version.
Driver specific
---------------
- Ditto for sysctls.
- snd_atiixp, snd_es137x, snd_via8233, snd_hda
* Numerous cleanups and fixes.
* _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme.
This was intended for pure debugging and latency measurement, but proven
good enough in few unexpected and rare cases (such as problematic shared
IRQ with GIANT devices - USB). Polling can be enabled/disabled through
dev.pcm.0.polling. Disabled by default.
- snd_ich
* Fix possible overflow during speed calibration. Delay final
initialization (pcm_setstatus) after calibration finished.
PR: kern/100169
Tested by: Kevin Overman <oberman@es.net>
* Inverted EAPD for few Nec VersaPro.
PR: kern/104715
Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp>
Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman,
those at #freebsd-azalia @ freenode and others for testing.
Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
|
|
|
}
|
|
|
|
} else if (l > 0)
|
|
|
|
sndbuf_dispose(src, buffer, l);
|
2000-09-23 22:11:32 +00:00
|
|
|
|
|
|
|
return count;
|
|
|
|
}
|
2000-12-18 01:36:41 +00:00
|
|
|
|
|
|
|
static kobj_method_t feeder_root_methods[] = {
|
|
|
|
KOBJMETHOD(feeder_feed, feed_root),
|
Sound Mega-commit. Expect further cleanup until code freeze.
For a slightly thorough explaination, please refer to
[1] http://people.freebsd.org/~ariff/SOUND_4.TXT.html .
Summary of changes includes:
1 Volume Per-Channel (vpc). Provides private / standalone volume control
unique per-stream pcm channel without touching master volume / pcm.
Applications can directly use SNDCTL_DSP_[GET|SET][PLAY|REC]VOL, or for
backwards compatibility, SOUND_MIXER_PCM through the opened dsp device
instead of /dev/mixer. Special "bypass" mode is enabled through
/dev/mixer which will automatically detect if the adjustment is made
through /dev/mixer and forward its request to this private volume
controller. Changes to this volume object will not interfere with
other channels.
Requirements:
- SNDCTL_DSP_[GET|SET][PLAY|REC]_VOL are newer ioctls (OSSv4) which
require specific application modifications (preferred).
- No modifications required for using bypass mode, so applications
like mplayer or xmms should work out of the box.
Kernel hints:
- hint.pcm.%d.vpc (0 = disable vpc).
Kernel sysctls:
- hw.snd.vpc_mixer_bypass (default: 1). Enable or disable /dev/mixer
bypass mode.
- hw.snd.vpc_autoreset (default: 1). By default, closing/opening
/dev/dsp will reset the volume back to 0 db gain/attenuation.
Setting this to 0 will preserve its settings across device
closing/opening.
- hw.snd.vpc_reset (default: 0). Panic/reset button to reset all
volume settings back to 0 db.
- hw.snd.vpc_0db (default: 45). 0 db relative to linear mixer value.
2 High quality fixed-point Bandlimited SINC sampling rate converter,
based on Julius O'Smith's Digital Audio Resampling -
http://ccrma.stanford.edu/~jos/resample/. It includes a filter design
script written in awk (the clumsiest joke I've ever written)
- 100% 32bit fixed-point, 64bit accumulator.
- Possibly among the fastest (if not fastest) of its kind.
- Resampling quality is tunable, either runtime or during kernel
compilation (FEEDER_RATE_PRESETS).
- Quality can be further customized during kernel compilation by
defining FEEDER_RATE_PRESETS in /etc/make.conf.
Kernel sysctls:
- hw.snd.feeder_rate_quality.
0 - Zero-order Hold (ZOH). Fastest, bad quality.
1 - Linear Interpolation (LINEAR). Slightly slower than ZOH,
better quality but still does not eliminate aliasing.
2 - (and above) - Sinc Interpolation(SINC). Best quality. SINC
quality always start from 2 and above.
Rough quality comparisons:
- http://people.freebsd.org/~ariff/z_comparison/
3 Bit-perfect mode. Bypasses all feeder/dsp effects. Pure sound will be
directly fed into the hardware.
4 Parametric (compile time) Software Equalizer (Bass/Treble mixer). Can
be customized by defining FEEDER_EQ_PRESETS in /etc/make.conf.
5 Transparent/Adaptive Virtual Channel. Now you don't have to disable
vchans in order to make digital format pass through. It also makes
vchans more dynamic by choosing a better format/rate among all the
concurrent streams, which means that dev.pcm.X.play.vchanformat/rate
becomes sort of optional.
6 Exclusive Stream, with special open() mode O_EXCL. This will "mute"
other concurrent vchan streams and only allow a single channel with
O_EXCL set to keep producing sound.
Other Changes:
* most feeder_* stuffs are compilable in userland. Let's not
speculate whether we should go all out for it (save that for
FreeBSD 16.0-RELEASE).
* kobj signature fixups, thanks to Andriy Gapon <avg@freebsd.org>
* pull out channel mixing logic out of vchan.c and create its own
feeder_mixer for world justice.
* various refactoring here and there, for good or bad.
* activation of few more OSSv4 ioctls() (see [1] above).
* opt_snd.h for possible compile time configuration:
(mostly for debugging purposes, don't try these at home)
SND_DEBUG
SND_DIAGNOSTIC
SND_FEEDER_MULTIFORMAT
SND_FEEDER_FULL_MULTIFORMAT
SND_FEEDER_RATE_HP
SND_PCM_64
SND_OLDSTEREO
Manual page updates are on the way.
Tested by: joel, Olivier SMEDTS <olivier at gid0 d org>, too many
unsung / unnamed heroes.
2009-06-07 19:12:08 +00:00
|
|
|
KOBJMETHOD_END
|
2000-12-18 01:36:41 +00:00
|
|
|
};
|
|
|
|
static struct feeder_class feeder_root_class = {
|
2003-08-05 07:23:35 +00:00
|
|
|
.name = "feeder_root",
|
|
|
|
.methods = feeder_root_methods,
|
|
|
|
.size = sizeof(struct pcm_feeder),
|
|
|
|
.desc = NULL,
|
|
|
|
.data = NULL,
|
2000-09-23 22:11:32 +00:00
|
|
|
};
|
2000-12-18 01:36:41 +00:00
|
|
|
SYSINIT(feeder_root, SI_SUB_DRIVERS, SI_ORDER_FIRST, feeder_register, &feeder_root_class);
|
2001-03-16 20:58:44 +00:00
|
|
|
SYSUNINIT(feeder_root, SI_SUB_DRIVERS, SI_ORDER_FIRST, feeder_unregisterall, NULL);
|