SpritesMind.Net

So I'm getting to the point where I want to get presets to use in my YM2612 VSTi and JS test area. My main question is WHEN in a VGM to take register values and dump to a file. I'm definitely dumping on note on events, but should I bother dumping if a song SOMEHOW changes a parameter while a note is on? If I read Eke's code correctly, it seems that the YM2612 supports this, and it may almost certainly happen in practice w/the VSTi, but what's the likelihood of this happening in a VGM?

The most likely events are the TL getting changed with a mid-note volume change and the frequency changing if the sound driver is instructed not ot retrigger notes. Any register can be changed at any time... Have you tried running Shiru's VGM2TFI to rip voices? The TFI format is documented in the old TFM Music Maker help file. (Alternatively, TFM comes with a bunch of voices you can also use...)

Note, any register can be changed at any time, but not all the changes are affect sound immideately, some only take effect after next key on. This fact could be used in drivers.

i have not yet used vgm2tfi; shiru gave me specs for other preset formats some time ago as well, and i'd want to use one of the simpler ones of those if possible.

if i write my own ripper, i pretty much only need to rip a preset if a note on takes place or if it's in the middle of a note and one of the few parameters that actually do affect the sound mid-way are changed (i think it's AR, algorithm/feedback, TL, and maybe one more like SSG or something; i don't have the code in front of me at the moment), right? with this i can still reserve the option to ignore middle-of-note parameter changes.

if i use vgm2tfi, i still need to either write *.tfi loading into my VSTi or i need to convert the *.tfi to something usable by the VSTi (VST's native preset format is *.fxp (one preset) and *.fxb (a bank of presets)).

AR/DR/SR/RR aren't take any effect before next keyon, TL, MULT, DT, and probably SL (not sure) affect sound immideately.

That set of parameters is better than I was expecting! Now I'll only need to watch two or three registers per slot instead of five or so, thanks!

Keep in mind that modifying TL on an output slot results in volume change, so if a driver changes volume mid-note, you're going to wind up with a bunch of duplicate voices with only differing volume. If you want that, then go for it, I guess =P