Please excuse me if my terminology is wrong. I'm from a music production background and have no experience in signal processing.

I was wondering if it was possible to stretch out the overtones (partial tones) of sound through digital signal processing. In the natural harmonic series or the overtone series, there is the fundamental frequency. Then you have overtones at integer multiples of the fundamental frequency. Is there a software which would let me alter the ratio between the different overtones of the sound? For example, If I have an audio file of 0-20,000 hertz. How can I take the first 10,000 hertz of the audio spectrum and stretch it out to 20,000 hertz so that all the overtones are spaced wider and the ratio is doubled.

It's called Frequency Scaling or Pitch Scaling I believe. I do understand the difference between Frequency shifting and Frequency Shifting and I'm curious to experiment with Frequency Scaling and not Frequency shifting.

I'm looking to do something like this.

Thank you.

  • $\begingroup$ If you play your audio at twice the reproduction sampling rate that it was recorded with, without doing anything else to it, wouldn't that do exactly what you want? $\endgroup$ – Marcus Müller Aug 2 at 12:13
  • $\begingroup$ @MarcusMüller I have tried up-sampling the audio file. But It sounds the same to me. Maybe the up-sampling algorithm in my software is doing other things to the audio to preserve the frequencies? How could I double the sample rate without preserving the frequencies? Is there existing software to achieve this? Thank you. $\endgroup$ – ab97 Aug 2 at 12:23
  • $\begingroup$ as said, don't upsample it. Play it twice as fast. Simply tell your system the audio is not at sampled at say, 48 kHz, but at 96 kHz. $\endgroup$ – Marcus Müller Aug 2 at 12:30
  • $\begingroup$ @MarcusMüller Thank you for your inputs. I will surely experiment with what you have suggested. I have found another solution using a plugin called MTransformer. You can see what it can do here. Thank you. $\endgroup$ – ab97 Aug 2 at 15:27

Everything is possible with advanced technology, due diligence and perseverance. To discuss the means to reach your goals requires even more detailed description of target scenarios.

If you have a pure tone -- I mean, a musically pure single note performed by a good musician with a quality instrument or an opera singer -- you can analyze this sound by spectral methods and even real-time-process this sound in order to shift its frequency or change its overtones any way you like. Then, you can implement these transformations with a digital signal processor or conventional computer.

The task is more difficult, if you have a musical piece recorded and want to process individual parts of the performance through submitting them to a "pitch scaling" as defined in your Correlogram Museum reference. Even to "pitch scale" a recorded or live-performed triad, not a single note, is an algorithmic and methodological challenge.

In the Correlogram Museum examples, the performer plays on a synthesized piano-like instrument. The DSP unit most certainly takes part in producing the sounds of this instrument, but the pipeline of sound production is inverse w.r.t. your scenario: the sound is generated with "scaled" overtone frequencies, whereas you have a sound with overtones prepared early and want to "post-process" the sound in order to change its overtone content.

I admit there are design projects pertinent to the scenario of your question. Ask colleagues, browse the literature, search the Internet.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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