Stephen Hazel skrev: > From what I can grok, the call progression for rendering seems to be: > fluid_synth_write_s16() calls
If you want to have 16-bit output, that's correct. If you want 32-bit output, you will start with another call. Anyway if I didn't overlook something, as fluid_synth_write_s16() calls fluid_synth_one_block() with "do_not_mix_fx_to_out" to 0, which means that fx_right_buf is never used and fx_left_buf is used as a mono reverb/chorus buffer. > It's good, but maaaan, when I see all those funky symbols and such > i start to freak out :( > If somebody rattles off "bresenhem line algorithm" (or whatever), I freak. Yeah, sometimes it's easier to understand things just reading the source, instead of trying to understand all the math behind it :-) > I'd like to tack on a .mid + .sf2 => .wav bit of functionality. > (not .SF2, actually, I split sf2 presets into .WAVs and an ascii file > that describes each preset.) Fluidsynth almost has render-to-wave functionality. Current status of that is: In version 1.0.9 you can do render to raw with the -a file option, but it will be just as slow as playing the tune in real-time. In the upcoming version (and present in SVN), there is the -F switch that will render at the pace that your computer can handle. What's left to implement is to render to another file format than raw (wave / ogg / mp3 / etc), which will probably be done by linking with libaudiofile or libsndfile. // David _______________________________________________ fluid-dev mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/fluid-dev
