Hi, On Mon, Dec 6, 2010 at 3:06 PM, <kro...@aolohr.com> wrote: > a = audio_to_stereo(a) > Starting liquidsoap. > At line 30, char 22: unbound symbol audio_to_stereo.
It means that your liquidsoap doesn't have any operator of that name. Are you not using liquidsoap SVN or 1.0 beta? > [Btw, I did not use the 'a = mean(audio_to_stereo)request.queue(...)))' > first, beacuse it barfed, and second, since absolutely none of my files > are in stereo, nor will they ever be. I took your phrase 'unless all > your input files are mono' as meaning I wouldn't need this. ] To clarify: if all your files are mono, you don't need anything (no audio_to_stereo) if you also want a mono output, and audio_to_stereo alone is needed if you want a stereo output. It's all written in the types of those operators: A mono output wants 1 mono channel, a stereo wants 2. A request source that must play mono and stereo files must have type *+1 (meaning at least one channel, but variable). To go from variable to fixed number of channels you have audio_to_stereo which has takes *+1 and yields 2. Now of course there's another ingredient that is hard to understand: liquidsoap has some flexibility in assigning types, and it guesses for you. But already having a rough idea of what channel types you want is useful. Here are some examples illustrating what I said: $ liquidsoap --no-pervasives -c -i 'r = request.queue() output.icecast(%mp3(mono),mount="foo",fallible=true,r)' r : source(audio=1,video=0,midi=0) Here the source r must produce 1 channel. $ liquidsoap --no-pervasives -c -i 'r = request.queue() output.icecast(%mp3(mono),mount="foo",fallible=true,mean(r))' r : source(audio=?#A,video=0,midi=0) Now, it's okay if the source produces any number of channels, but fixed. The precise number is unknown at the time of printing the type, but will be forced to 2 by default. $ liquidsoap --no-pervasives -c -i 'r = request.queue() output.icecast(%mp3(mono),mount="foo",fallible=true,mean(audio_to_stereo(r)))' r : source(audio=*+1,video=0,midi=0) By inserting the audio_to_stereo, we've fixed the number of channels in input of the mean() operator: two. More importantly, our source can now produce a variable number of channels, but at least one (*+1). Cheers, -- David ------------------------------------------------------------------------------ _______________________________________________ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
