spectrogram to sound
Hello,
I'm more a Jitter user than an MSP one and I searched the forums but surprisingly I couldn't find any "simple" information on this.
I need to synthesize a sound from it's spectrogram. So to say, I slice a row of pixels from an image, convert it to grayscale (and in case to a list) and than feed it to.. ? Is there such an object which recreate a sound starting from its "composing frequencies"? If that make sense at all..
Thanks,
Giorgio
maybe not exactly waht you are looking for, but I've been doing something like this...http://giorgiomagnanensi.com/spunspectrums/
...starting from this patch by Alexander Refsum Jensenius: http://www.uio.no/english/research/groups/fourms/software/imagesonifyer/index.html
Giorgio
Not sure if it's the right solution, but you could experiment with oscbank~ by mapping pixel values to frequencies of individual oscillators. If you have lots of pixels this would of course eat up your processor capacity, but you could give it a try.
Grazie Giorgio, nice patch, but I can't really get why (and how) ImageSonifyer use the pink (or white) noise to generate the sound. My limit.
Thanks Benjisugar, yes oscbank˜ seems the way to go. I finally found this as a starting point.
https://cycling74.com/forums/image-to-sound-a-la-aphex-twins-equation/
Still exploring the patch though..
Also check the "forbidden planet" patch in the Examples/fft-fun folder for a subtractive synthesis method based on FFT.