Okay... how DO you make OPN voices?

andlabs · Post by **andlabs** » Tue May 17, 2011 4:39 am

This is something I don't quite get. I know we have ample documentation on how the YM2612 works, and I have a basic understanding of how FM synthesis works, but I don't understand how one creates a OPN voice based around a specific waveform... Is there some guide or some old document used by MD/PC-88/PC-98 developers that explained the process — ore even X1 or X68000 programmers who used OPM chips, since the voices appear to be mostly compatible? I did look quickly at soundshock and found http://www.gorenfeld.net/lou/fmguide.html but that doesn't really help much — I'm looking for something different (I'm pretty sure but...); something along the lines of technical that talks about the correspondence between operators and the effect their parameters have? Thanks, and I hope that didn't confuse anyone =P

TmEE co.(TM) · Post by **TmEE co.(TM)** » Tue May 17, 2011 6:40 am

The closer TL is to zero the harsher tone you get, the closer to max is Feedback the more the output looks like sawtooth on operator 1. Ratios of multipliers make up a big chunk of tonal qualities of the instrument.
When you mix in ADSR to modulator operators you can get some cool sounds happen ^^
Just keep messing with it and you'll learn to predict what sounds you'll be getting.

GreenLine · Post by **GreenLine** » Tue May 17, 2011 7:44 am

With respect to FM sounds I found this information very useful:

http://noisesculpture.com/synthesizer-boot-camp

There are several videos that show the process of forming sounds with lots of examples. The author also publishes a free e-book with good info:

http://noisesculpture.com/how-to-make-a-noise

This has helped me much getting introduced in the FM synthesis approach used by OPN chips.

Gigasoft · Post by **Gigasoft** » Tue May 17, 2011 8:27 am

A technique I sometimes use is to load a recording of the intrument up in a frequency analyzer, and look for groups of evenly spaced lines around a central line. Then, set an operator to the middle frequency and another with a frequency equal to the spacing, and let the second modulate the first, and play with the total level of the modulator until it sounds right. You can also compare the resulting sound with the original recording in the analyzer and look for differences.

You could also just play around with different settings until you find something that sounds interesting, then adjust it until you are satisfied.

andlabs · Post by **andlabs** » Sun May 22, 2011 6:23 am

I've read parts of the book/watched the videos and played around some more — that helped me get my head around it a bit so far. Until then, I'll keep playing around and reading more and I will keep everything said here in mind. Thanks