Bug 242365 - wrong track length and positioning for ogg files with Phonon vlc backend
Summary: wrong track length and positioning for ogg files with Phonon vlc backend
Status: RESOLVED UPSTREAM
Alias: None
Product: phonon-backend-vlc
Classification: Frameworks and Libraries
Component: general (show other bugs)
Version: unspecified
Platform: openSUSE Linux
: NOR normal
Target Milestone: ---
Assignee: Martin Sandsmark
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-06-21 15:14 UTC by Cyril Brosch
Modified: 2012-05-31 01:59 UTC (History)
7 users (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments
shown track length is shorter than actual track length (466.88 KB, image/png)
2010-06-21 15:16 UTC, Cyril Brosch
Details
shown track length is longer than actual track length (341.79 KB, image/png)
2010-06-21 15:17 UTC, Cyril Brosch
Details
actual jump-to time (see tooltip) diverges from mouse pointer position (402.18 KB, image/png)
2010-06-21 15:19 UTC, Cyril Brosch
Details
screenshot showing all the problems (track length, position) in vlc (479.71 KB, image/png)
2010-06-22 12:18 UTC, Cyril Brosch
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Cyril Brosch 2010-06-21 15:14:21 UTC
Version:           4.4.1 (KDE SC 4.4.4) (using KDE 4.4.4) 
OS:                Linux

I noticed that the videolan backend passes wrong track lengths for ogg files (but not for mp3), which causes several problems.
Often the shown length is too short (1st screenshot; the progress slider reaches its end, but Amarok keeps playing up to the actual length), but sometimes it's too long (2nd screenshot). In this cases Amarok stops playing after the actual end is reached.
Additionally jumping to a position doesn't work as it should: See the third screenshot: The mouse pointer (not visible in the shot, I've drawn the approximate position) points to a position temporally after the actual slider position, but as the tooltip says, clicking there will make the pointer jump to a position 20 seconds earlier, before the actual playing time.

The aberration often seems to be proportional to the overall track length, it adds or chops (I can't predict what of both it does) about 9%, but there are sometimes bigger deviations.

I tested ogg and mp3 file, in Amarok and in Dragonplayer, I don't have flac or wave ones. The xine backend doesn't have this problem.
My libvlc version is 1.1 rc4, phonon-vlc is the git version from 2010-06-01.

I attach a sample file: The actual length is 4:09, but it is shown as 2:09.

Reproducible: Always
Comment 1 Cyril Brosch 2010-06-21 15:16:48 UTC
Created attachment 48189 [details]
shown track length is shorter than actual track length
Comment 2 Cyril Brosch 2010-06-21 15:17:27 UTC
Created attachment 48190 [details]
shown track length is longer than actual track length
Comment 3 Cyril Brosch 2010-06-21 15:19:44 UTC
Created attachment 48191 [details]
actual jump-to time (see tooltip) diverges from mouse pointer position
Comment 4 Cyril Brosch 2010-06-21 15:26:52 UTC
OK, the sample file is top big for b.k.o., look here: http://www.cyrilbrosch.net/comme_d_habitude.ogg (some song from youtube)
Comment 5 Myriam Schweingruber 2010-06-22 11:58:26 UTC
Cyril, could you please test this with the latest git version only? The VLC backend is under development, so a version from 20 days ago is a bit old.
Comment 6 Cyril Brosch 2010-06-22 12:12:47 UTC
Arrg, I forgot to test the most obvious thing: The problem is an issue of vlc, which can't determine the length of ogg tracks correctly, not of the backend.

See: http://trac.videolan.org/vlc/ticket/1859 and http://trac.videolan.org/vlc/ticket/1098 (open for years now!).
So nothing to do for the KDE guys, except one thing: Make Amarok not stop playing, if the actual length is shorter than shown by the vlc backend.
Should I file a new bug report for that?
Comment 7 Cyril Brosch 2010-06-22 12:18:22 UTC
Created attachment 48217 [details]
screenshot showing all the problems (track length, position) in vlc
Comment 8 Myriam Schweingruber 2010-06-22 12:28:59 UTC
Thank you for the update, Cyril, I will ping the VLC guys about that.
Comment 9 Cyril Brosch 2010-06-23 15:04:42 UTC
It's still in 1.1 final and of course with phonon-vlc 0.2, I've filed a separate report for the stopping problem, see https://bugs.kde.org/show_bug.cgi?id=242597 .
Comment 10 Nikos Chantziaras 2010-08-28 19:37:23 UTC
I can confirm this bug.

Using latest phonon-vlc from Git 
(git://gitorious.org/phonon/phonon-vlc.git).
Amarok 2.3.1.90.
KDE 4.5.0.
Phonon 4.4.2.
VLC 1.1 branch from Git 
(git://git.videolan.org/vlc/vlc-1.1.git).
Comment 11 Myriam Schweingruber 2010-08-29 13:00:24 UTC
Setting to confirmed, based on comment #10
Comment 12 Cyril Brosch 2010-08-30 11:12:55 UTC
@Nikos

Could you please confirm it in the bug report I linked to in comment #9, because I don't know whether the Amarok guys can do anything upstream (I hope I use this word correctly...) about the vlc bug, but they surely can make a workaround that Amarok doesn't stop when the track is shorter than shown (see the extra bug report).
Comment 13 Nikos Chantziaras 2010-08-30 11:19:46 UTC
(In reply to comment #12)
> @Nikos
> 
> Could you please confirm it in the bug report I linked to in comment #9,
> because I don't know whether the Amarok guys can do anything upstream (I hope I
> use this word correctly...) about the vlc bug, but they surely can make a
> workaround that Amarok doesn't stop when the track is shorter than shown (see
> the extra bug report).

That bug is pretty much a duplicate of this one :-/  But anyway, there you go.
Comment 14 Jean-Baptiste Kempf 2010-10-17 20:26:12 UTC
The bug of duration still exists, and is a VLC bug, not a VLC-Phonon bug.

But it should at least go 'next' correctly.
Comment 15 Myriam Schweingruber 2010-10-17 20:40:46 UTC
Closing as upstream, based on comment #14
Comment 16 Jean-Baptiste Kempf 2010-10-17 20:45:31 UTC
Upstream bug is http://trac.videolan.org/vlc/ticket/1859
Comment 17 Myriam Schweingruber 2011-04-27 21:32:38 UTC
Reassigning to the new bugzilla product for better bug tracing of the various
backends. Sorry for the noise.