pyTivo Discussion Forum

[wmcbrine]

Here's something I've been thinking about for a while... but new contributor David Rein went ahead and did it: Transcoding of audio files for the music plugin. It's checked into my repository now.

This uses the same ffmpeg as the video plugin. However, you may need to change yours -- for Windows, the one that's currently bundled with pyTivo segfaults immediately when you tell it to output an mp3. Instead, I've been using one from here:

http://arrozcru.no-ip.org/ffmpeg_builds/

I'm using the latest, ffmpeg-r11213-gpl-shared-win32.tar.bz2 . I do want to replace the one in pyTivo, but then again I'm looking at taking it out altogether... but that's for another thread.

For Linux, I suggest building your own from the SVN source; configure with at least "--enable-libmp3lame". You'll probably want "--enable-libfaac" and "--enable-gpl", too.

I haven't looked into the Mac situation yet.

So what does this mean? You can play any audio file that ffmpeg supports. For that matter, you can even play the audio tracks of video files through the music plugin. I've played unencrypted AAC, WMA, WAV and AU, as well as MPG and FLV.

As with the extension-based video filter, we need a fairly exhaustive list of extensions to make this work well; currently there's only a very limited one (and I omitted the extensions for video). The list of extensions is in the TRANSCODE tuple near the top of music.py.

[wgw]

Amazingly enough, a new ffmpeg r11870 was just released which works with the music transcoding. Or one can revert to r10464.
http://ffdshow.faireal.net/mirror/ffmpeg/

Seems to also work fine for video transcoding with my initial quick tests. Still reassigns channels on mkv's though like previous versions.

[wmcbrine]

[wgw]

I think it was compiled without a couple speech libs.