Bug 1680

Summary: Sound has harsh noise in 1.0-0.122.20110227svn
Product: Fedora Reporter: sangu <sangu.fedora>
Component: mplayerAssignee: Dominik 'Rathann' Mierzejewski <dominik>
Status: RESOLVED FIXED    
Severity: normal CC: adamw, kwizart, mishu, mithi, pikachu.2014, rtguille
Priority: P5    
Version: 15   
Hardware: x86_64   
OS: GNU/Linux   
namespace:
Attachments: mplayer -v mp3file
mp3 decoded by mplayer/mp3lib/sse-optimized and recorded with -ao pcm:file=sample.wav

Description sangu 2011-04-01 09:37:52 CEST
While audio or video file with mplayer 1.0-0.122.20110227svn.fc15, sound has harsh noise.
Then, sound is good in gnome players (like totem, rhythmbox) and mplayer-1.0-0.119.20100703svn.fc14. 

kernel-2.6.38.2-9.fc15.x86_64
Comment 1 Dominik 'Rathann' Mierzejewski 2011-04-01 12:09:45 CEST
Please add full mplayer -v yourfile output as an attachment.
Comment 2 sangu 2011-04-01 13:51:41 CEST
Created attachment 603 [details]
mplayer -v mp3file
Comment 3 Thomas Jansen 2011-04-01 23:45:01 CEST
I get the same problem using the same version of mplayer when I try to listen to mp3. The expected sound is played (i.e., you can recognize the song) but mixed with a shrill, high-pitched and loud noise.

Watching a video with mplayer (mkv container, h264 video, aac audio, both using ffmpeg) works as expected. Therefore, I suspect the problem occurs only for mp3 files. The relevant decoder was mp3lib with SSE optimization:

==========================================================================
Opening audio decoder: [mp3lib] MPEG layer-2, layer-3
dec_audio: Allocating 4608 + 65536 = 70144 bytes for output buffer.
mp3lib: using SSE optimized decore!
MP3lib: init layer2&3 finished, tables done
mpg123: Can't rewind stream by 221 bits!
MPEG 1.0, Layer III, 44100 Hz 96 kbit Joint-Stereo, BPF: 313
Channels: 2, copyright: No, original: Yes, CRC: No, emphasis: 0
AUDIO: 44100 Hz, 2 ch, s16le, 96.0 kbit/6.80% (ratio: 12000->176400)
Selected audio codec: [mp3] afm: mp3lib (mp3lib MPEG layer-2, layer-3)
==========================================================================

This assumption is supported by the fact that playing a song with "mplayer -ao pcm:file=sample.wav" yields a very strange signal when viewed in audacity. I'll attach a short excerpt. Note that the bottom half (0 to -1) is at 100%.
Comment 4 Thomas Jansen 2011-04-01 23:47:11 CEST
Created attachment 604 [details]
mp3 decoded by mplayer/mp3lib/sse-optimized and recorded with -ao pcm:file=sample.wav
Comment 5 Thomas Jansen 2011-04-01 23:54:40 CEST
One more indication that it's a decoder problem and a possible workaround for users before a new rpm is released: "mplayer -ac ffmp3" works correctly on the same file.
Comment 6 Dominik 'Rathann' Mierzejewski 2011-04-04 17:29:27 CEST
*** Bug 1682 has been marked as a duplicate of this bug. ***
Comment 7 Dominik 'Rathann' Mierzejewski 2011-04-04 17:48:07 CEST
Known issue, upstream is aware. Looks like gcc-4.6-related problem.
Comment 8 Dominik 'Rathann' Mierzejewski 2011-04-11 15:35:41 CEST
*** Bug 1692 has been marked as a duplicate of this bug. ***
Comment 9 Peter Gückel 2011-04-12 01:14:36 CEST
I thought the new ffmpeg or libva update that I just got would solve the problem, but they didn't.
Comment 10 Thomas Jansen 2011-04-12 07:51:10 CEST
Since this is bug in mplayer code, updating libraries won't help. Mplayer brings its own code for mp3 decoding. You have to wait for a mplayer update, Peter.
Comment 11 Dominik 'Rathann' Mierzejewski 2011-04-12 18:34:30 CEST
I've just built mplayer-1.0-0.123.20110412svn with mp3lib decoder disabled as a workaround until upstream fixes the bug.
Comment 12 Peter Gückel 2011-04-14 19:40:10 CEST
(In reply to comment #11)
> I've just built mplayer-1.0-0.123.20110412svn with mp3lib decoder disabled as a
> workaround until upstream fixes the bug.
> 

It seems that mp3lib decoder is not required. I updated to the new mplayer and playing mp3 files works just fine now.

So, what is mp3lib decoder for, since it is not required for playing mp3?
Comment 13 Nicolas Chauvet 2011-06-29 21:17:14 CEST
Does the bug is actually fixed ?
Please close the bug in this case.
Thx for the report.