I understand how FM synthesis works, but I was wondering how to FM arbitrary samples or part of samples. I.E any frame of audio samples.

Further more the modulator can itself be an arbitrary frame of samples itself.

How is this done in an algorithmic manner?


closed as unclear what you're asking by Marcus Müller, Matt L., MBaz, lennon310, A_A Nov 28 '18 at 16:58

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ sorry, it's unclear what you mean with "FM arbitrary samples or part of samples": Samples of what? What is a "part of a sample"? (that's kind of a thing that makes no sense, in discrete signal processing, which deals with samples) And no, your Fourier Transform-based description doesn't read right in any form; I'm really not that sure that you've 100% understood FM, sorry. $\endgroup$ – Marcus Müller Nov 22 '18 at 0:07
  • $\begingroup$ I will not argue about my understanding of FM. I meant simple forms. Back to the matter: I mean Instead of The carrier being a sine/square/etc oscillator, it is any wave shape you give it, and in turn it is modulated by standard oscillators, or, with yet another non standard frame of samples. That's how samplers with FM Support work somehow. As long as it is tuned. I was wondering what is the algorithm change for that. that's it. $\endgroup$ – user192234 Nov 22 '18 at 0:29
  • 2
    $\begingroup$ Frequency Modulation specifically is mapping an amplitude of a waveform of interest to a frequency of a carrier (thus the carrier's frequency is modulated by the waveform). That said it is not very clear if what you are asking for would be FM at all. Could you clarify how what you are considering is still to be FM when you say you modulate one arbitrary waveform with another? Maybe it would help if you could describe the end result you are trying to achieve (for what purpose)? $\endgroup$ – Dan Boschen Nov 22 '18 at 4:37
  • 1
    $\begingroup$ I really meant it when I said everything is unclear to me, and just like Dan I work at uni teaching DSP; if the both of us don't understand your description, I'd say there's a high likelihood that it's not 100% purely my fault (definitely not saying I understand everything, I'm not overly smart, but I'd like to point out that it'd really help addressing Dan's questions – maybe even one by one – in your question). I see in your profile that you're probably a physicist, so I'm fairly optimistic you can express what you mean in formulas! That'd be the optimum; math is much less ambiguous. $\endgroup$ – Marcus Müller Nov 22 '18 at 8:44
  • $\begingroup$ @MarcusMüller I'll do some thinking before I formulate it mathematically. :) $\endgroup$ – user192234 Nov 22 '18 at 11:16

I have, in the past, used delay modulation on arbitrary audio signals. but the modulation waveform itself was not arbitrary, but was a sinusoid. but we could speed this modulation oscillator up to an audio rate.

Now that I think of it, I tried hooking that modulation oscillator up to a pitch detector. This is a quarter century ago, so I don't remember everything.

I s'pose you could derive that delay modulation signal from the input and likely LPF it before applying it to the instantaneous delay of the signal going out.

  • $\begingroup$ This brings up to mind one or two good ideas to have a go at it. Thank you! $\endgroup$ – user192234 Nov 22 '18 at 10:44
  • 1
    $\begingroup$ Using Delay modulation for this has a great sound :) $\endgroup$ – user192234 Nov 22 '18 at 12:01
  • $\begingroup$ @user192234, are you interpolating between samples when you apply delay modulation? because some instantaneous delays will be have fractional sample precision. you gotta be able to get between samples. $\endgroup$ – robert bristow-johnson Nov 23 '18 at 8:58
  • $\begingroup$ Now I do, yeah. Removes the distortion. $\endgroup$ – user192234 Nov 28 '18 at 15:09

Not the answer you're looking for? Browse other questions tagged or ask your own question.