Fourier Transform of orbiting objects, translated to OSC wave forms and oscillation.

[dlarseninclusive]

I am starting some ground work on capturing an orbit, like Mercury for instance; then translating that into parameters that a synthesizer can understand. Likely using Fast Fourier Transformation (FFT) which is also common in digital filter algorithms.

Looks like this:

Fourier Square.jpg

With an orbit, I found an example of one with two binary stars:

Binary Star.jpg

With one waveform per oscillator, and modulation in amplitude. I will get more complex with 3D and 4D plots if this works.

I am still working on data collection, and the math functions to get the output I am looking for.

I think Zebra is a likely winner to accept the output as far as capabilities.

Interested in anyone's thoughts on how I might apply that to Zebra, or is there a better candidate?

One of the handful of gaps to fill in this plan is how to take the output waveforms, and apply them to synth parameters, hopefully in an automated way.

Thanks

[Winstontaneous]

Awesome, what are you using for the math functions & plots? I work with R for my day job and am learning Python which has better crossover to DSP. I've done some similar basic experiments towards making a "sonic scale model of the solar system." I don't know much about Zebra's h2p file format and am not sure how you would map this to, say, an MSEG (perhaps the binary data at the bottom of h2p presets holds this)?

[dlarseninclusive]

Paraphrasing a bit and I am still working on it.
(DFT)Fourier Transform
X(k) = integral(x(t)e -jkt )dt
For each discrete waveform, and add amplitude modulation A for each unique cycle.
f(t) = (Sin,n(Theta) + n(i)A dt)n(i)

Beyond that, I have a few other gaps to fill. I am trying to use universe sandbox for data collection and it does not output data. I need to figure out how to get measurements or find another data source.

And yes, printing that data should be possible just outputting the parameters to the h2p file. Translating that data to the correct parameters is another unknown. (the first couple might be manual tuning with the oscilloscope to match a graph)

I was starting to look at Matlab, if this can get generalized better than not guessing the answer for 20-200 hours I think python would work fine.

I am still trying to wrap my head around the whole "machine". I expect to have a better answer soon.

[dlarseninclusive]

R is an interesting idea. I have more time with Octave, and only played with Matlab from time to time.

Python is much more approachable if it ends up being fast enough.

[dlarseninclusive]

Steve Duda had an interesting idea, prototyping on supercolider, which looks worth checking out.

https://en.wikipedia.org/wiki/SuperCollider

[r_v]

If you would find it useful to have a Python module for turning data into Zebra oscillators: https://github.com/harveyormston/osc_gen

It supports fft-based resynthesis, so should do a good job of handing any kind of periodic data.

[dlarseninclusive]

That looks like the right stuff.. Thanks!

I will check it out.

[CrystalWizard]

This looks like a cool project. Sonification ! Definitely keep this thread updated as it progresses.

[V0RT3X]

Might be worth picking up Serum.

That way you could try using the math functions in serums wavetable editor to create your particular waveform, and then load that waveform into something like wav2zebra.

It’s what I’d try

[dlarseninclusive]

Might be worth picking up Serum.

That way you could try using the math functions in serums wavetable editor to create your particular waveform, and then load that waveform into something like wav2zebra.

It’s what I’d try

I asked Steve Duda what he thought about Serum's functions and he pointed me towards SuperCollider.

[dlarseninclusive]

This looks like a cool project. Sonification ! Definitely keep this thread updated as it progresses.

Will do thanks! I intend to post progress.

[dlarseninclusive]

I am thinking the new uhm script might be worth investigating for this project.

Again, I am shooting for what the orbit of Mercury sounds like.

Transform.jpg

Fourier Transform Sine Reference Link:

Mercury Precesion Modulation.jpg

Reference to paper:

I am not sure how to describe an integral as trigonometry, yet... might have to make some assumptions.

[EvilDragon]

Those integrals seem fairly complex and not at all easy to break down. Good luck

[dlarseninclusive]

Yeah for real, Thanks.

Those integrals seem fairly complex and not at all easy to break down. Good luck

[Howard]

Not a sonification-music fan - too "arty" for its own good. But I'll say "good luck" too!