filters: cutoff is scaled by CUTOFFRATIO, but hcutoff and bandf aren't

[claudeha]

cutoff (for low pass filter, aka vcf effect) is in units that are similar to Hz at typical sample rates, while hcutoff (for high pass filter, aka hpf effect) and bandf (for band pass/reject filter) are not scaled and are expected to be less than 1

server.c:  sound->cutoff = cutoff / CUTOFFRATIO;
server.c:  sound->hcutoff = hcutoff;
server.c:  sound->bandf = bandf;

Neither is truly sample rate independent, but I wasted 30mins today wondering why hcutoff 0.001 was doing what I wanted but cutoff 0.1 wasn't.

Should we make hcutoff and bandf also scale by CUTOFFRATIO?

Should we add an additional factor of 44100.0/g_samplerate to try to make them more independent of sample rate?

These changes would change how things sound, might need to be behind a compatibility flag?

[yaxu]

Oh that's odd that they're scaled differently. SuperDirt expects them all in Hz, so perhaps we should standardise on that?

Agreed making them independent of samplerate would be best.

I don't think a compatibility flag is needed.