nesdev.org

I'm having trouble with higher notes not sounding like guitars. Do I not have the right ADSR envelopes or the right kind of vibrato?

psycopathicteen wrote:I'm having trouble with higher notes not sounding like guitars.

Why would that be a problem?

I'm trying to make guitars sounds.

Using FM synth ? Complicate to replicate, i think that requires very high frequency modulation.

Yeah, I'm stupid, I misread it as sounding like, not not sounding like. That joke sucked.

You can hear how the first part it does a fairly good guitar sound, but when you get to the second part it sounds really weird.

That's a very bizarre FM guitar. I was expecting something along the lines of this: https://www.youtube.com/watch?v=rlmbkKtjfSc

That's because I'm not an FM synth expert.

Found what was causing it to sound crappy at high frequencies. The way I calculated note frequencies was too inaccurate.

I've been thinking of ways to do 4 op fm synth on either the CPU or SPC700 (or both at the same time), and then I thought maybe I can make a 2612 VGM player if I can get all 6 channels to work.

Would be awesome to have a 2612 VGM player on SNES, honestly i though from years now about doing the opposite (SPC player on Sega Genesis ^^).

I think I'll try doing HDMA SPC700 loading. So, I guess I just have to time the spc700 loop to be exactly 65 cycles, but somehow make sure it starts at the beginning of a scanline too. Like checking for a header byte, followed by 32 nops, followed by the loop.

psycopathicteen: have you considered making an python IT/XM/etc tracker conversion script etc for this that exports to SPC, ASM code or SFC binary?

The n00bz over at http://battleofthebits.org could have some fun with this.

psycopathicteen wrote:I think I'll try doing HDMA SPC700 loading. So, I guess I just have to time the spc700 loop to be exactly 65 cycles, but somehow make sure it starts at the beginning of a scanline too. Like checking for a header byte, followed by 32 nops, followed by the loop.

Won't you waste many previous SPC700 cpu time doing that ? I believe having a fast loading loop once per frame would probably be probably lighter on SPC side at least.

I just tried doing it the normal way, and I counted cycles and it appears that it takes just as long for the spc700 either way. I did it 2 bytes at a time, and it added up to 33 spc700 cycles per 16 bits, which would be 66 cycles per 32 bits. I could also try doing 4 bytes at once, but I doubt there will be much improvement.