internal libfaad2 forked copy
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come. libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues. Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo. Diego
tis 2010-11-02 klockan 13:28 +0100 skrev Diego Biurrun:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad? As I have quite some tv recordings with LATM and have a friend who does a lot of recordings from DVB-t with LATM, we can test many files. Will try to do that during this and next week. Dan
On Tue, Nov 02, 2010 at 01:56:42PM +0100, Dan Oscarsson wrote:
tis 2010-11-02 klockan 13:28 +0100 skrev Diego Biurrun:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad?
-afm ffmpeg -ac ffaac Diego
Diego Biurrun <diego <at> biurrun.de> writes:
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad?
-afm ffmpeg -ac ffaac
Which sample did you test (with r32570, the latest version when you wrote this)? Carl Eugen
On Tue, Nov 02, 2010 at 02:19:49PM +0000, Carl Eugen Hoyos wrote:
Diego Biurrun <diego <at> biurrun.de> writes:
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad?
-afm ffmpeg -ac ffaac
Which sample did you test (with r32570, the latest version when you wrote this)?
I tested with r32568 the samples I had locally available, they are from http://samples.mplayerhq.hu/A-codecs/AAC/ though I now see that I'm missing the LATM subdirectory. Diego
tis 2010-11-02 klockan 14:01 +0100 skrev Diego Biurrun:
On Tue, Nov 02, 2010 at 01:56:42PM +0100, Dan Oscarsson wrote:
tis 2010-11-02 klockan 13:28 +0100 skrev Diego Biurrun:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad?
-afm ffmpeg -ac ffaac
I have tried on mpeg-ts streams from DVB-t broadcasts. It works with lavf demuxer, not with standard demuxer. Though with my "sync audio and video after seek"-patch gives better result with standard demuxer and faad due to seek then ends up with audio in front of video. May patch will then skip a few video frames to match audio resulting in half decoded frames being skipped. With lavf demuxer and ffmpeg audio and video is more in sync after a seek and you see the incomplete frames which is less nice. Dan
On Wed, Nov 10, 2010 at 02:18:56PM +0100, Dan Oscarsson wrote:
tis 2010-11-02 klockan 14:01 +0100 skrev Diego Biurrun:
On Tue, Nov 02, 2010 at 01:56:42PM +0100, Dan Oscarsson wrote:
tis 2010-11-02 klockan 13:28 +0100 skrev Diego Biurrun:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad?
-afm ffmpeg -ac ffaac
I have tried on mpeg-ts streams from DVB-t broadcasts. It works with lavf demuxer, not with standard demuxer.
Which revision and file have you tested? It works just fine for me and for what I know for everyone else who tested it, except for one special format which should be very rarely used (and wouldn't work with the lavf demuxer either).
ons 2010-11-10 klockan 18:15 +0100 skrev Reimar Döffinger:
On Wed, Nov 10, 2010 at 02:18:56PM +0100, Dan Oscarsson wrote:
tis 2010-11-02 klockan 14:01 +0100 skrev Diego Biurrun:
On Tue, Nov 02, 2010 at 01:56:42PM +0100, Dan Oscarsson wrote:
tis 2010-11-02 klockan 13:28 +0100 skrev Diego Biurrun:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
So if I get the current source snapshot I will get the new ffmpeg that supports LATM? What do I need to set to enable it instead of libfaad?
-afm ffmpeg -ac ffaac
I have tried on mpeg-ts streams from DVB-t broadcasts. It works with lavf demuxer, not with standard demuxer.
Which revision and file have you tested? It works just fine for me and for what I know for everyone else who tested it, except for one special format which should be very rarely used (and wouldn't work with the lavf demuxer either).
Hmm, I though everything was ready when I got the snapshot 5 days ago. Fetching todays snapshot it works fine without any options and it selects fflatm itself. Dan
Diego Biurrun <diego <at> biurrun.de> writes:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2.
No. Carl Eugen
On Tue, Nov 02, 2010 at 02:14:02PM +0000, Carl Eugen Hoyos wrote:
Diego Biurrun <diego <at> biurrun.de> writes:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2.
No.
Then cut down on the suspense and enlighten us. Diego
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
Yes, patience. First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default, then there's some waiting for bug reports, then would should maybe think about disabling both internal and external faad by default and then we're really close to remove the actual code.
On Tue, 2 Nov 2010 20:39:46 +0100, Reimar Döffinger wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
Yes, patience. First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default, then there's some waiting for bug reports, then would should maybe think about disabling both internal and external faad by default and then we're really close to remove the actual code.
dont forget merging our latm faad stuff into upstream. btw we can help ffmpeg by getting more latm samples. maybe someone could post a news entry asking for people to test svn ffmpeg and report bugs/samples straight to roundup? -compn
On Tue, Nov 02, 2010 at 09:34:50PM -0400, compn wrote:
On Tue, 2 Nov 2010 20:39:46 +0100, Reimar Döffinger wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
Janne just committed LATM support to FFmpeg. AFAIK that fixes the main shortcoming of FFmpeg's AAC decoder compared to libfaad2. IOW, the time for us to remove the internal libfaad2 copy from MPlayer appears to have come.
libfaad2 is still faster than ffaac in fixed-point mode on my K6-III, which has a slow FPU, but I'm not aware of any other major issues.
Anything I'm missing? Otherwise libfaad2 shall go the way of the dodo.
Yes, patience. First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default, then there's some waiting for bug reports, then would should maybe think about disabling both internal and external faad by default and then we're really close to remove the actual code.
dont forget merging our latm faad stuff into upstream.
Feel free to try. I personally have little hope there. Diego
compn ha scritto:
dont forget merging our latm faad stuff into upstream.
Menno first committed my patch and then reverted it because it didn't work. I seriously doubt that Nero's CTO has any interest in keeping faad up2date :-)
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default,
All samples play now with -demuxer lavf, but afaict, both -aid and TAB are completely broken with -demuxer lavf. I tested gazeta-20090408-1456-aa.ts from samples/ffmpeg-bugs/roundup/issue1462/ Carl Eugen
On Thu, Nov 04, 2010 at 10:31:35PM +0000, Carl Eugen Hoyos wrote:
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default,
All samples play now with -demuxer lavf, but afaict, both -aid and TAB are completely broken with -demuxer lavf.
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay, so for now a working native demuxer still isn't optional, unless someone want to fix our lavf code (in addition to fixing it so seeking is non-dead-slow).
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
All samples play now with -demuxer lavf, but afaict, both -aid and TAB are completely broken with -demuxer lavf.
Fixed,
Thank you!
however -demuxer lavf can't display audio and sub languages for BluRay, so for now a working native demuxer still isn't optional, unless someone want to fix our lavf code (in addition to fixing it so seeking is non-dead-slow).
Dead slow is not so bad, but seeking to EOF is imo (roundup issue 1963). Carl Eugen
On Sat, Nov 06, 2010 at 04:00:52PM +0100, Reimar Döffinger wrote:
On Thu, Nov 04, 2010 at 10:31:35PM +0000, Carl Eugen Hoyos wrote:
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default,
All samples play now with -demuxer lavf, but afaict, both -aid and TAB are completely broken with -demuxer lavf.
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this... Aurel
On Sun, 5 Dec 2010 20:55:24 +0100, Aurelien Jacobs wrote:
On Sat, Nov 06, 2010 at 04:00:52PM +0100, Reimar Döffinger wrote:
On Thu, Nov 04, 2010 at 10:31:35PM +0000, Carl Eugen Hoyos wrote:
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default,
All samples play now with -demuxer lavf, but afaict, both -aid and TAB are completely broken with -demuxer lavf.
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this...
http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts C:\>mplayerx -demuxer lavf http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts MPlayer SVN-r32659-4.5.1 (C) 2000-2010 MPlayer Team 161 audio & 353 video codecs Playing http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts. Resolving samples.mplayerhq.hu for AF_INET... Connecting to server samples.mplayerhq.hu[213.144.138.186]: 80... Cache size set to 320 KBytes Cache fill: 13.73% (45003 bytes) libavformat file format detected. [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [mpegts @ 020d5010] Too many streams [NULL @ 023fd010] start time is not set in av_estimate_timings_from_pts [NULL @ 0245e010] start time is not set in av_estimate_timings_from_pts [NULL @ 020babb0] start time is not set in av_estimate_timings_from_pts [NULL @ 0245e7e0] start time is not set in av_estimate_timings_from_pts [NULL @ 020bc9e0] start time is not set in av_estimate_timings_from_pts [NULL @ 021d4010] start time is not set in av_estimate_timings_from_pts [NULL @ 021d4760] start time is not set in av_estimate_timings_from_pts [NULL @ 02490210] start time is not set in av_estimate_timings_from_pts [NULL @ 024907e0] start time is not set in av_estimate_timings_from_pts [NULL @ 02329010] start time is not set in av_estimate_timings_from_pts [NULL @ 02329760] start time is not set in av_estimate_timings_from_pts [NULL @ 02291010] start time is not set in av_estimate_timings_from_pts [NULL @ 02291760] start time is not set in av_estimate_timings_from_pts [NULL @ 021f0210] start time is not set in av_estimate_timings_from_pts [NULL @ 021f07e0] start time is not set in av_estimate_timings_from_pts [lavf] stream 0: video (h264), -vid 0 [lavf] stream 1: audio (dca), -aid 0 [lavf] stream 2: audio (ac3), -aid 1 [lavf] stream 3: audio (ac3), -aid 2 [lavf] stream 4: audio (ac3), -aid 3 [lavf] stream 5: subtitle (pgssub), -sid 0 [lavf] stream 6: subtitle (pgssub), -sid 1 [lavf] stream 7: subtitle (pgssub), -sid 2 [lavf] stream 8: subtitle (pgssub), -sid 3 [lavf] stream 9: subtitle (pgssub), -sid 4 [lavf] stream 10: subtitle (pgssub), -sid 5 [lavf] stream 11: subtitle (pgssub), -sid 6 [lavf] stream 12: subtitle (pgssub), -sid 7 [lavf] stream 13: subtitle (pgssub), -sid 8 [lavf] stream 14: subtitle (pgssub), -sid 9 [lavf] stream 15: subtitle (pgssub), -sid 10 [lavf] stream 16: subtitle (pgssub), -sid 11 [lavf] stream 17: subtitle (pgssub), -sid 12 [lavf] stream 18: subtitle (pgssub), -sid 13 [lavf] stream 19: subtitle (pgssub), -sid 14 LAVF: Program 1 VIDEO: [H264] 1920x1080 0bpp 23.976 fps 0.0 kbps ( 0.0 kbyte/s) Resolving samples.mplayerhq.hu for AF_INET... Connecting to server samples.mplayerhq.hu[213.144.138.186]: 80... Opening video filter: [scale] Opening video filter: [eq] ========================================================================== Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264) ========================================================================== ========================================================================== Trying to force audio codec driver family ffmpeg... Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders AUDIO: 48000 Hz, 6 ch, s16le, 1536.0 kbit/33.33% (ratio: 192000->576000) Selected audio codec: [ffdca] afm: ffmpeg (FFmpeg DTS) ========================================================================== AO: [dsound] 48000Hz 2ch s16le (2 bytes per sample) Starting playback... A: 0.0 V: 0.0 A-V: 0.000 ct: 0.000 0/ 0 ??% ??% ??,?% 0 0 300% Cache not responding! Exiting... (Quit) C:\> -compn
compn <tempn <at> twmi.rr.com> writes:
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this...
http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts
Yes, this is the best sample we have (with a bad name, I just had to search it for an hour...) Carl Eugen
On Sun, 5 Dec 2010 20:36:34 +0000 (UTC), Carl Eugen Hoyos wrote:
compn <tempn <at> twmi.rr.com> writes:
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this...
http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts
Yes, this is the best sample we have (with a bad name, I just had to search it for an hour...)
btw there is an external .sup (bluray) subtitle, any plans to support it in ffmpeg? http://samples.mplayerhq.hu/sub/BluRay/Subpictures_20.sup -compn
On Sun, Dec 05, 2010 at 08:36:34PM +0000, Carl Eugen Hoyos wrote:
compn <tempn <at> twmi.rr.com> writes:
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this...
http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts
Yes, this is the best sample we have (with a bad name, I just had to search it for an hour...)
Rename it or add a symlink :-) Diego
On Sun, Dec 05, 2010 at 03:41:29PM -0500, compn wrote:
On Sun, 5 Dec 2010 20:55:24 +0100, Aurelien Jacobs wrote:
On Sat, Nov 06, 2010 at 04:00:52PM +0100, Reimar Döffinger wrote:
On Thu, Nov 04, 2010 at 10:31:35PM +0000, Carl Eugen Hoyos wrote:
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
First the issues that Carl reported against the FFmpeg decoder must be analyzed and/or fixed, then FFmpeg one must be made default,
All samples play now with -demuxer lavf, but afaict, both -aid and TAB are completely broken with -demuxer lavf.
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this...
http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts
Huh? Are you sure this has languages with the native demuxer? For Bluray the languages comes from other files in the structure, and the reason it doesn't work for lavf is probably because it can't match them up because of the different channel numbering. Since demux_lavf doesn't export the native stream number at all, it currently isn't fixable.
Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:
Fixed, however -demuxer lavf can't display audio and sub languages for BluRay,
Do you have any sample ? It shouldn't be hard to add support for this...
http://samples.mplayerhq.hu/sub/BluRay/sample-dts-not-valid-dca-frame.m2ts
Huh? Are you sure this has languages with the native demuxer?
We were just searching for the sample with the most subtitles, misunderstanding what you wrote. Carl Eugen
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I just tested most of the samples from http://samples.mplayerhq.hu/A-codecs/AAC/ AFAICT they all work fine with both internal FAAD2 and fflatm, both with native and lavf demuxer. The only exception is http://samples.mplayerhq.hu/A-codecs/AAC/freetv_aac_latm.ts which works with no combination of demuxer or decoder (no sound output). ffplay also fails, but vlc plays it fine. So what's missing for making ffaac default? Diego
On Sun, 07 Nov 2010 22:23:19 +0100, Diego Biurrun wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I just tested most of the samples from
http://samples.mplayerhq.hu/A-codecs/AAC/
AFAICT they all work fine with both internal FAAD2 and fflatm, both with native and lavf demuxer. The only exception is
http://samples.mplayerhq.hu/A-codecs/AAC/freetv_aac_latm.ts
which works with no combination of demuxer or decoder (no sound output). ffplay also fails, but vlc plays it fine.
So what's missing for making ffaac default?
benchmarks. also testing all aac files in http://samples.mplayerhq.hu/archive/audio/aac/ -compn
On Sun, Nov 07, 2010 at 05:42:37PM -0500, compn wrote:
On Sun, 07 Nov 2010 22:23:19 +0100, Diego Biurrun wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I just tested most of the samples from
http://samples.mplayerhq.hu/A-codecs/AAC/
AFAICT they all work fine with both internal FAAD2 and fflatm, both with native and lavf demuxer. The only exception is
http://samples.mplayerhq.hu/A-codecs/AAC/freetv_aac_latm.ts
which works with no combination of demuxer or decoder (no sound output). ffplay also fails, but vlc plays it fine.
So what's missing for making ffaac default?
benchmarks. also testing all aac files in http://samples.mplayerhq.hu/archive/audio/aac/
Go for it :) I don't think benchmarks are really necessary, AFAIR ffaac is 2-3 times faster than libfaad2. Diego
On Sun, Nov 07, 2010 at 10:23:19PM +0100, Diego Biurrun wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I just tested most of the samples from
http://samples.mplayerhq.hu/A-codecs/AAC/
AFAICT they all work fine with both internal FAAD2 and fflatm, both with native and lavf demuxer. The only exception is
http://samples.mplayerhq.hu/A-codecs/AAC/freetv_aac_latm.ts
which works with no combination of demuxer or decoder (no sound output). ffplay also fails, but vlc plays it fine.
So what's missing for making ffaac default?
fflatm already is, if ffaac isn't I think it's ok to make it the default.
On Sun, 07 Nov 2010 22:23:19 +0100 Diego Biurrun <diego@biurrun.de> wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I just tested most of the samples from
http://samples.mplayerhq.hu/A-codecs/AAC/
AFAICT they all work fine with both internal FAAD2 and fflatm, both with native and lavf demuxer. The only exception is
http://samples.mplayerhq.hu/A-codecs/AAC/freetv_aac_latm.ts
which works with no combination of demuxer or decoder (no sound output). ffplay also fails, but vlc plays it fine.
Actually, the audio frames inside don't look like LATM - it's raw AAC with extradata passed through the IOD descriptor in PMT table. -- Jindrich Makovicka
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I'm not aware of any issues blocking libfaad2 removal anymore. AFAIK all problems with LATM were resolved and ffaac is the default AAC decoder since some time now. Speak your minds soon or be silent forever :) Diego
On Wed, 22 Dec 2010 12:55:25 +0100, Diego Biurrun wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I'm not aware of any issues blocking libfaad2 removal anymore.
AFAIK all problems with LATM were resolved and ffaac is the default AAC decoder since some time now. Speak your minds soon or be silent forever :)
external libfaad work ok ? anyone test it on mac/mingw/cygwin/bsd/os2/amiga or other OS that mplayer supports but most of our forked libs do not ? imo we should just keep it in a seperate repo like svn.mplayerhq.hu/libfaad2-outdated/ and if anyone needed it to test without playing with external they could pull it that way. i guess svn diff -r x:x > file.patch and then applying it reversed would be about the same. ffaac is still missing the small aac subtypes like ssr: http://roundup.ffmpeg.org/issue662 -compn
On Wed, Dec 22, 2010 at 08:05:42AM -0500, compn wrote:
On Wed, 22 Dec 2010 12:55:25 +0100, Diego Biurrun wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I'm not aware of any issues blocking libfaad2 removal anymore.
AFAIK all problems with LATM were resolved and ffaac is the default AAC decoder since some time now. Speak your minds soon or be silent forever :)
external libfaad work ok ?
Last I checked, yes.
anyone test it on mac/mingw/cygwin/bsd/os2/amiga or other OS that mplayer supports but most of our forked libs do not ?
Who cares? And most of our forked libs support these OSes. If we have local changes to support them, the solution is to port them upstream.
imo we should just keep it in a seperate repo like svn.mplayerhq.hu/libfaad2-outdated/ and if anyone needed it to test without playing with external they could pull it that way. i guess svn diff -r x:x > file.patch and then applying it reversed would be about the same.
This is what old revisions are for.
ffaac is still missing the small aac subtypes like ssr: http://roundup.ffmpeg.org/issue662
These are irrelevant. Testing with external libfaad is of course welcome nonetheless. Diego
On Wed, 22 Dec 2010 14:37:51 +0100, Diego Biurrun wrote:
On Wed, Dec 22, 2010 at 08:05:42AM -0500, compn wrote:
anyone test it on mac/mingw/cygwin/bsd/os2/amiga or other OS that mplayer supports but most of our forked libs do not ?
Who cares? And most of our forked libs support these OSes. If we have local changes to support them, the solution is to port them upstream.
i meant that mplayer's configure can handle such os's but maybe the external lib's cannot. its not that we have changes but that our configure is superior. not to mention autotools being required to compile faad. more testing would be nice , but i'm not blocking its removal. -compn
On 22 dec 2010, at 12:55, Diego Biurrun <diego@biurrun.de> wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I'm not aware of any issues blocking libfaad2 removal anymore.
AFAIK all problems with LATM were resolved and ffaac is the default AAC decoder since some time now. Speak your minds soon or be silent forever :)
IMO: remove it from configure ASAP. Then remove the actual code in January. I know there's history so it might be quite pointless to split it this way but the diffs are less messy and I don't know if restoring a full subdir wouldn't be a mess.
On Wed, Dec 22, 2010 at 05:59:29PM +0100, Reimar Döffinger wrote:
On 22 dec 2010, at 12:55, Diego Biurrun <diego@biurrun.de> wrote:
On Tue, Nov 02, 2010 at 01:28:18PM +0100, Diego Biurrun wrote:
[...]
I'm not aware of any issues blocking libfaad2 removal anymore.
AFAIK all problems with LATM were resolved and ffaac is the default AAC decoder since some time now. Speak your minds soon or be silent forever :)
IMO: remove it from configure ASAP. Then remove the actual code in January.
I'll disable it in configure by default and then kick it out come 2011. Nobody touch this code before me, I've wanted to get rid of libfaad2 for far too long :-) Diego
participants (8)
-
Aurelien Jacobs -
Carl Eugen Hoyos -
compn -
Dan Oscarsson -
Diego Biurrun -
Jindrich Makovicka -
Nico Sabbi -
Reimar Döffinger