hu, cd, scd, acd, supergrafx discussions.
Post Reply
Posts: 88
Joined: Mon Jun 23, 2008 1:58 pm


Post by tomaitheous »

Did some tests with LFO this pass week.

One quick note, initialize *all* registers to 00 on startup for *each* register select. That includes stuff that would *not* normally apply to it, i.e. the noise registers and LFO. LFO had some strange mode that if these were not initialized, for me on the SGX. Drove me nuts. It treated the LFO data like 1bit values. I'm not sure what the threshold was, but it wasn't as simple as ANDing source samples and getting 00/$1F. Maybe I'll test it more this weekend to find out. Interesting but not useful - imo.

Ok, on to the results. Register $809 shift value is 1=0, 2=2, 3=4. Not 0,4,8 as the PSG doc and the patent says. The patent *figure* is it correct though - go figure. The other thing is that the signed 1's compliment PCM data effects all 12bits of channel 0's period value(frequency) - shifted or not. The addition *can* cause the period value to wrap back to 0x1000 (which is the slowest period value) if the base period value is close enough to 0x000 - say like 0x034. Makes for some interesting sounds. The channel 1 can also output sound artifacts from it's own DAC in LFO mode, so it's a good idea to turn the volume down to 0 unless you want this effect. Also, the volume control doesn't not change the timbre of the frequency modulation - that's too bad. The other details about the LFO control are correct. There are some other quirks that I'll test for this weekend ;)
User avatar
Site Admin
Posts: 402
Joined: Sun Jun 22, 2008 3:19 pm
Location: Lvl 3


Post by MooZ »

I consider the LFO to be the 4th dimension of the PSG. I don't even know how it's supposed to be used and if any game had ever used it :)
Let's go for a mad LFO week end !
User avatar
Posts: 9
Joined: Fri Aug 29, 2008 3:28 am
Location: Ohio (USA)


Post by Arkhan »

The LFO was put there to let the PSG achieve FM-like sounds, on a non FM chip!

Post Reply