After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 395054 - MP3 timeline playback not working properly
MP3 timeline playback not working properly
Status: RESOLVED DUPLICATE of bug 308312
Product: GStreamer
Classification: Platform
Component: gnonlin
0.10.11
Other All
: Normal normal
: git master
Assigned To: GStreamer Maintainers
Edward Hervey
Depends on:
Blocks:
 
 
Reported: 2007-01-10 16:26 UTC by Mathis Hofer
Modified: 2007-02-08 20:36 UTC
See Also:
GNOME target: ---
GNOME version: 2.17/2.18


Attachments
Example programs in Python and C. (317.20 KB, application/x-compressed-tar)
2007-01-10 16:57 UTC, Mathis Hofer
Details

Description Mathis Hofer 2007-01-10 16:26:17 UTC
Please describe the problem:
When using GnonLin to play MP3 files, only the first file gets played, then EOS is sent. The same thing works without problem if one uses WAV files for example.

Steps to reproduce:
Create a GnonLin example where you put two or more MP3 files (i.e. gnlfilesource) one after another in the "GnonLin-timeline" (gnlcomposition), having the second one start at the duration of the first etc.


Actual results:
Only the first file gets played, then EOS is sent.

Expected results:
The files should be played one after another.

Does this happen every time?
Yes.

Other information:
Comment 1 Mathis Hofer 2007-01-10 16:57:16 UTC
Created attachment 79966 [details]
Example programs in Python and C.

Here is an example program in Python and C to simulate the described behavior.

You can decode the MP3 files to WAV files and change the filenames. It then works as expected (the two files get played after each other).
Comment 2 Mathis Hofer 2007-01-10 17:05:26 UTC
Works with OGG too, no problems with WAV or OGG files.
Comment 3 Jan Schmidt 2007-01-19 15:45:40 UTC
I don't know what behaviour of the decoder might cause this?

It might be related to Bug #308312, but I'm not sure.

Edward, can you comment?
Comment 4 Mathis Hofer 2007-02-08 16:08:51 UTC
Are there any news about this issue?

Who is Edward and what does he/do you say about this?
Comment 5 Edward Hervey 2007-02-08 20:36:44 UTC
The problem is that mad doesn't accept the seek event:
ERROR       gnlcomposition gnlcomposition.c:1213:no_more_pads_object_cb:<timeline> Sending seek event failed!

Which means that that source isn't set up properly for playing back in segment mode. And what happens in the end is that it emits EOS (which is the normal behaviour in non-segment mode).

The full reason for this is that the mad and fluendo mp3 decoder don't implement seeking accurately. Here the seeking is requested before the decoder can estimate the average bitrate, therefore can not do the time=>bytes conversion which is needed to do an upstream seek in bytes.

I'm marking this as a duplicate of bug #308312 .

*** This bug has been marked as a duplicate of 308312 ***