Subsonic Forum

I've been a happy subsonic user for couple years now. With the recent updates I've ran into a problem I seem not to be able to do anything about. What practically happens it that the playback simply just stops somewhere in the middle of the song, despite the length being shown correctly in the standard web interface's flash player. It also does not proceed to the next song in the list but stalls completely. This is frequent and happens with majority of my songs now, always at the very same spot for each song. It's also impossible to hear the end of the track even by manually trying to skip past the point where the playback ends. Subsonic logs look clean, except for these HTTP GET's with Range-parameter, but I'm not sure if they're related to this or not.

Code: Select all

[5/8/15 11:06:29 AM EEST]    INFO   StreamController   Got HTTP range: 0-


I'm currently running 5.2.1 on a dedicated, powerful PC and Debian Jessie. The 10k+ flac files are mounted from a network disk via nfs and i'm transcoding to mp3 when streaming. The issue also occurs on standalone clients but so far I haven't experienced it with the mobile client. The computer I'm mostly using for playback is in a big and complex company network which leads me to wondering whether or not this could be something network related?

I see the same thing in 5.2.1 and it's really annoying. Playing the mp3 in any other media player works fine. Even windows media player of all things... :/ On a couple songs I did notice a small pop or crackle at the time code of subsonic failure. It looks like subsonic does NOT handle even the slightest imperfection in a mp3. Some of the crackle I hear is even in the original recording. Really a bad bug.

Yeah, it pretty much makes the whole thing unusable for me, for now. Help, anyone?

Okay, this definitely seems to be ffmpeg-related. I lowered the converted mp3 bitrate from 320k to 192k and at least the first few, previously stalling songs are now being played until the actual end.

So, after a day of listening it's now clear that lowering the converted mp3's bitrate definitely helps by making this less frequent, but the problem still occurs. Any help for actually fixing this, anyone?