Fm2

[beserge]

Still in need of testing, but at first glance it seems to work OK.

[andrewikenberry]

fm2 DaisySP FM Pair Osc

[andrewikenberry]

fm2 DaisySP FM Pair Osc

[stephenhensley]

Alright! sorry for the delay on this.

This definitely already sounds cool, but I think since this is already going to be a bit of a hard-coded module (one modulator/one carrier), with the expectation of most often, if not exclusively, being used with sine waves.

I believe the breakdown of the FM becomes incorrect in this context because the base-oscillator class does not handle negative frequencies (you avoided this by taking the absolute value of the frequency at all times), but I don't think that has the same effect since the frequency would be PI radians out of phase when frequency is negative, likely changing the expected side-bands.

We might as well do it with PM instead of actual FM with stuff as it is. Might simplify some of the interface a bit, too. (Only dual sines). Pretty sure our existing oscillator class will suit this fine, and if it can't we should fix it to support it, though I'm pretty sure the PhaseAdd function should be able to do what you need.

[stephenhensley]

Alright. Sounding gnarly.

Few things I noticed:

• Ratio does not work -- notes on this below
• Once fixed, ratio doesn't track unless updated every time. This is what should get stored as a private variable, and the frequency should be updated either within process, or else where. This has some implications on the other setter, and may require some changes -- notes below
• The range for index should probably be scaled internally to respond well between 0-1:
  • Maybe just multiply by (0.2 or something on the input -- kind of hard to decide where it gets too rough, but even 0.4 becomes harsh/noise, not that we don't want that stuff to be possible, but it should be beyond 1 in my opinion.

Regarding the ratio tracking: it seems that its not really possible to set/forget this as the SetRatio is what sets the modulator frequency, depending on the carrier frequency. So trying to just have a Fm2 with ratio set to 2x, and a fixed index doesn't have the expected output because the modulator just gets stuck on the initial frequency.

After a brief discussion the path forward (since this is a module with a pretty specific use-case) is to:

• Remove the SetModulatorFrequency function
• Change SetCarrierFrequency to SetFrequency
• store ratio_ as private variable, and update mod. freq. whenever car. freq. setter is called.

If we still want some semblance of independent control we could do a carrier ratio and modulator ratio, but

That'll simplify the overall module, and should fix any of the inconsistencies related to ratio.

Future improvements:

• use lookup table or other fast sin implementation.