bad sound (clicks + low pitch / slower speed)

[goblin]

Since I've upgraded mplayer from debian squeeze's 2:1.0~rc3++final.dfsg1-1, sound in one of my .mkv files became weird - hearing lots of clicks and lower pitch and possibly slower speed than it should have (like if you played a vinyl record on too slow RPM).

In squeeze-backports' mplayer version 2:1.0~rc4.dfsg1+svn33713-2~bpo60+1 it was logging lots of messages similar to:

[aac @0x7f1e7b4acbe0]Not evaluating a further program_config_element as this construct is dubious at best.

With latest mplayer with latest ffmpeg, I instead see lots of this (and the problem is the same):

[aac @0x125c100]Evaluating a further program_config_element.

Here's the full output of latest ffplay (version git-2012-03-09-d07de6d) when trying to play the file:

ffplay version git-2012-03-09-d07de6d Copyright (c) 2003-2012 the FFmpeg developers

built on Mar 9 2012 18:52:30 with gcc 4.4.5
configuration:
libavutil 51. 42.100 / 51. 42.100
libavcodec 54. 10.100 / 54. 10.100
libavformat 54. 2.100 / 54. 2.100
libavdevice 53. 4.100 / 53. 4.100
libavfilter 2. 63.100 / 2. 63.100
libswscale 2. 1.100 / 2. 1.100
libswresample 0. 7.100 / 0. 7.100

[h264 @ 0x207be00] Increasing reorder buffer to 1
Input #0, matroska,webm, from 'file.mkv':

Metadata:

title : <cut>
creation_time : 2006-03-30 <cut>

Duration: <cut>, start: 0.000000, bitrate: 1811 kb/s

Stream #0:0(jpn): Video: h264 (Main), yuv420p, 1280x720, SAR 1:1 DAR 16:9, 23.98 fps, 23.98 tbr, 1k tbn, 47.95 tbc (default)
Metadata:

title : <cut>

Stream #0:1(jpn): Audio: aac, 48000 Hz, 5.1, s16 (default)
Metadata:

title : Japense Audio 6ch AAC

[aac @ 0x207c480] Evaluating a further program_config_element.1/0 f=0/0

1.70 A-V: -0.016 fd= 2 aq= 320KB vq= 1141KB sq= 0B f=1/0

Last message repeated 2 times

When playing the file with latest mplayer with -ac faad (as suggested by someone on IRC), the sound is fine.

[Carl Eugen Hoyos]

Please provide a sample.

[goblin]

Uploaded a sample to incoming/tkt_1049_low_pitch_sound.mkv

Couple of extra things to note:

• I have two sound cards, the built-in HDA Nvidia, and a PCI Sound Blaster Audigy
• The order of them is random after each boot - so yesterday when I tried ffplay I had the buggy sound OK out of the Nvidia card, but after today's boot it was all quiet because it chose the Audigy - even though I always have default as Audigy in my ~/.asoundrc. So today to reproduce the bug I had to use AUDIODEV=hw:0 envvar to force SDL to use the Nvidia card.
• Again please note that the low pitch sound / clicks occur only on the Nvidia card when using ffplay - the Audigy card is all silent (no sound). This doesn't happen with mplayer: with mplayer, I can use any card and still hear the bug occur on both. Unless I use mplayer -ac faad, in which case the sound is fine on both cards.
• Please also note that the mkv has 6-channel audio in 5.1 - both of my cards are only doing stereo (as I only have stereo speakers connected). It might be trying to play all 6 channels together through only 2 channels, which might explain why there's extra data and why the sound appears "slower".

[Carl Eugen Hoyos]

Please do not report problems with ffplay if they are also reproducible with ffmpeg (ffplay depends on hardware and an external library that is known to contain bugs, ffmpeg is mostly self-contained).

Works fine with mplayer -ac faad.

out.wav sounds bad:

$ ffmpeg -i tkt_1049_low_pitch_sound.mkv out.wav
ffmpeg version N-38686-gd07de6d Copyright (c) 2000-2012 the FFmpeg developers
  built on Mar 10 2012 13:28:13 with gcc 4.3.2
  configuration: --cc=/usr/local/gcc-4.3.2/bin/gcc --enable-gpl
  libavutil      51. 42.100 / 51. 42.100
  libavcodec     54. 10.100 / 54. 10.100
  libavformat    54.  2.100 / 54.  2.100
  libavdevice    53.  4.100 / 53.  4.100
  libavfilter     2. 63.100 /  2. 63.100
  libswscale      2.  1.100 /  2.  1.100
  libswresample   0.  7.100 /  0.  7.100
  libpostproc    52.  0.100 / 52.  0.100
[h264 @ 0x8ebda20] Increasing reorder buffer to 1
Input #0, matroska,webm, from 'tkt_1049_low_pitch_sound.mkv':
  Metadata:
    title           : [Shinsen-Subs] Ergo Proxy - 02 - Confession of a Fellow Citizen
    creation_time   : 2006-03-30 12:06:58
  Duration: 00:26:47.31, start: 0.000000, bitrate: 41 kb/s
    Stream #0:0(jpn): Video: h264 (Main), yuv420p, 1280x720, SAR 1:1 DAR 16:9, 23.98 fps, 23.98 tbr, 1k tbn, 47.95 tbc (default)
    Metadata:
      title           : [Shinsen-Subs] Ergo Proxy - 02
    Stream #0:1(jpn): Audio: aac, 48000 Hz, 5.1, s16 (default)
    Metadata:
      title           : Japense Audio 6ch AAC
Output #0, wav, to 'out.wav':
  Metadata:
    title           : [Shinsen-Subs] Ergo Proxy - 02 - Confession of a Fellow Citizen
    creation_time   : 2006-03-30 12:06:58
    encoder         : Lavf54.2.100
    Stream #0:0(jpn): Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, 5.1, s16, 4608 kb/s (default)
    Metadata:
      title           : Japense Audio 6ch AAC
Stream mapping:
  Stream #0:1 -> #0:0 (aac -> pcm_s16le)
Press [q] to stop, [?] for help
[aac @ 0x8ebdfa0] Evaluating a further program_config_element.
    Last message repeated 24 times
size=    7392kB time=00:00:13.14 bitrate=4608.0kbits/s
[aac @ 0x8ebdfa0] Evaluating a further program_config_element.
    Last message repeated 24 times
size=   14720kB time=00:00:26.19 bitrate=4603.0kbits/s
[aac @ 0x8ebdfa0] Evaluating a further program_config_element.
    Last message repeated 9 times
[matroska,webm @ 0x8eb6aa0] Read error
[aac @ 0x8ebdfa0] Evaluating a further program_config_element.
    Last message repeated 3 times
size=   19008kB time=00:00:33.79 bitrate=4608.0kbits/s
video:0kB audio:19008kB global headers:0kB muxing overhead 0.000349%