Ultrasonic FM?

[zerocrossing]

Has there ever been a synthesizer that uses FM like FM radio, where two high frequency ultrasonic waves are used, to make an audio signal to of the beat frequencies?

[gambero]

The basis of FM radio is the same as FM synthesizers, with the added difference that the carrier signal is demodulated away, so you ideally just get the modulator signal back. Long story short, if done correctly it would just sound like the modulator.

[Tone2 Synthesizers]

Has there ever been a synthesizer that uses FM like FM radio, where two high frequency ultrasonic waves are used, to make an audio signal to of the beat frequencies?

Modulating high frequency content usually will result in lots of aliasing.

Tone2 Nemesis is the only synth I know that is able to do real FM (not PM) synthesis with high frequencies and any waveform and without aliasing.

[zerocrossing]

The basis of FM radio is the same as FM synthesizers, with the added difference that the carrier signal is demodulated away, so you ideally just get the modulator signal back. Long story short, if done correctly it would just sound like the modulator.

Ah, good point. I sort of figured that might be the case.

[zerocrossing]

Has there ever been a synthesizer that uses FM like FM radio, where two high frequency ultrasonic waves are used, to make an audio signal to of the beat frequencies?

Modulating high frequency content usually will result in lots of aliasing.

Tone2 Nemesis is the only synth I know that is able to do real FM (not PM) synthesis with high frequencies and any waveform and without aliasing.

Nemesis is one of my favorites!

[justin3am]

I think what you are referring to is called heterodyning. I've played around with those concepts with my hardware modular and test equeipment but not with software. Using ringmod, FM and sample&hold on two ultra-sonic signals to create child tones within the audible spectrum.
With sufficient oversampling or some kind of perdictive model, I assume it would be possible to ringmod a sine at 30kHz and another at 31khZ and get an audible 1k tone. Things get more interesting with more complex signals and feedback, as harmonics start interacting with each other in chaotic ways.

Without oversampling, those ultrasonic frequencies will cause aliasing but if you can adjust the sample rate dynamically (for example by using a sample&hold clocked at audio rates, then you can have some manner of control of the frequencies at which the processed sounds generate aliasing.