Thanks for the hint. Ok, so I went ahead I installed libmagic-ocaml-dev and recompiled the whole thing.
The good news: The correct decoder is now chosen and playback works. So far, so good :) The bad news: I'm having the strange effect that the source fades out prematurely almost exactly half way into the track and then suddenly resumes again (all within the same track). This only happens with FLAC files. I'll try to explain what happens using the log: The playlist encounters a FLAC file and playback starts normally at first: 2010/07/05 19:36:01 [format.mp3:2] Mime type for test.flac is not valid: audio/x-flac 2010/07/05 19:36:01 [decoder:3] Decoder FLAC chosen for "test.flac". 2010/07/05 19:36:01 [mksafe:3] Switch to safe_blank. 2010/07/05 19:36:06 [ambient_test(dot)m3u8:3] Prepared "test.flac" (RID 1). 2010/07/05 19:36:06 [mksafe:3] Switch to src_6946 with transition. 2010/07/05 19:36:08 [root:2] We must catchup 2.08 seconds! Then, about half way into that track the source suddenly fades out. Shortly before that happens the following message appeared in the log: 2010/07/05 19:37:32 [decoder:3] Decoder MP3 chosen for "test.mp3". "test.mp3" is the next track in the playlist. As soon as the fade-out completes the same track "test.flac" suddenly resumes from where it had just been faded out. From here on the playlist proceeds normally, e.g. "test.flac" finishes without problem and is followed by "test.mp3". To me this almost looks like liquidsoap is interrupting the FLAC decoder while it's figuring out the decoder for the next track... Any idea what could be the problem and how to fix it? regards, Henry > Hi, > > On Mon, Jun 28, 2010 at 6:26 AM, Henry Favretto <[email protected]> wrote: >> My guess would be that the FLAC decoder is tried AFTER the mp3 decoder. >> And since the mp3 decoder doesn't complain about the input it ends up >> being chosen... > > Good guess. We changed the order and didn't finish the planned > modifications, which should in the end give more control over decoding > order. > >> If such is the case what are my options to get around this problem (if >> there are any)? > > The only solution for now is to recompile with support for mime > detection, i.e. ocaml-magic. (It is recognized at configure time if > installed; note that this package isn't provided by savonet.) > > Cheers, > > David > ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Savonet-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/savonet-users
