Rockbox Technical Forums
Support and General Use => Audio Playback, Database and Playlists => Topic started by: porktree on October 31, 2007, 03:07:52 PM
-
I've had this happen on two different mp3's (that are long 40 minutes plus). They begin playing normally, but after about 10 minutes or so (it varies) they stop outputing sound. The time counter and progress indicator for continue to move, and I can fast forward (although once they've stopped making sound no amount of fastforwarding causes sound to return). If I reverse they will play up to the point output ceased and cease output at that point. If I play a different mp3, and return to the problem mp3, I get the same results. The mp3 plays in the OF, and if I boot they Sansa a couple of times, it will play as expected in Rockbox.
I'm using a Sansa e270, and the build I'm on is r15365-071029.
This issue is intermittent. I can listen to Major Nelson's full podcast on my way home from work, but when I start Kevin Smith's smodcast, it dies. The next time it may happen on Major Nelson's podcast, and let me listen to all of Wait Wait Don't Tell Me.
-
I suspect that this may have something to do with rebuffering getting stuck somehow. The buffering system recently experienced a large revision with the inclusion of Metadata on Buffer. r15305 introduced Metadata on Buffer and there have been several revisions related to this since then. You can help to narrow down this problem by testing some older builds to see when this problem appeared. Using a binary search pattern would probably be a good idea. You can find a list of daily builds from the Current Builds page. Unfortunately this only gives coarse graining since many commits occur each day. Even better would be to download the source from SVN and compile builds to test which revisions exhibit this behavior.
For the record, I have not noticed such problems on my Gigabeat F40.
-
I'll try and start working backwards. I first noticed this in r15136-071016, which was the first Rockbox I installed (coincidently enough). I blew it of as user/file/encoding error, and it wasn't until I saw it again in later versions with different files that I thought it might not be me. It's intermittent enough that it may not happen for days.
-
If you can find a reliable way of reproducing this issue then you should file a bug report (provided one has not already been filed) for this.