Howdy, Whoops, I think I may have just fired off an empty reply. Sorry about that. Anyway, if I may step in for a second...
At 11:22 AM 1/5/2002 +0100, Christoph Reichenbach wrote: >Hi, > >> COMMENT & QUESTION - Set the tempo thing aside for a moment. It seems as if >> many notes don't want to be let go of at the proper time. The note continues >> to hold for a much longer period of time than it should or even forever >> (until the program is closed). Could it be this problem is affecting the >> polyphony count? > >That's quite possible, yes. Just as a reference point, I haven't noticed any notes held indefinitely or obscenely long using the same GM mapping that FreeSCI had around August 2000 when put in a driver and used with Sierra's interpreter. I also haven't noticed any dropped notes. IIRC, the only difference between the mapping code in the GM driver and the mapping code from old FreeSCI is that FreeSCI silenced unmapped patches whereas the driver plays them with a piano. I'm using the MIDI device on an SB Live. Also, I gather from this discussion that the polyphony choke (although disabled at the moment) is currently set at 32 and (when enabled) will prevent new notes from being played when full. But since the MT-32 has a maximum polyphony count of 32, and often has less in practice, FreeSCI might want to cut off old notes to pass new notes through, since that's more likely to be what was intended. Cheers, Ravi.
