fourier question

  Awepittance said:

Some french dude i forgot his name made a bunch of Jitter patches designed to read the audio of a wave form then convert its FFT/spectral print to a video or still image. Then you can manipulate this still image with video effects like blur, distort, pixelate.

I totally had the opposite idea! You know how you can open any file format like a raw wave file in Audacity, was thinking it would be sweet to like put a flanger on a jpeg and then open it back up as an image, never figured out how to get the file working afterwards though. I wonder if there are any image editors that will let you open a music file like raw data.

  spraaaa said:

  Awepittance said:

Some french dude i forgot his name made a bunch of Jitter patches designed to read the audio of a wave form then convert its FFT/spectral print to a video or still image. Then you can manipulate this still image with video effects like blur, distort, pixelate.

 

I totally had the opposite idea! You know how you can open any file format like a raw wave file in Audacity, was thinking it would be sweet to like put a flanger on a jpeg and then open it back up as an image, never figured out how to get the file working afterwards though. I wonder if there are any image editors that will let you open a music file like raw data.

all the versions of photoshop i've used let you open raw data, the problem is that it's degrades the sound quality in some way when you open it up as a normal image. there is actually a blog online i saw a couple months ago all about one man's experiments with editing sounds in photoshop and vice versa and im sure you can do it the other way around. I've never tried it myself.

I've always wanted to have a sound printed out on really high quality paper using photoshop, then manipulate that paper either by crumpling it or burning /wetting parts of it so it smears then rescanning it into the computer at the exact same dimensions to see what it would sound like. I've always been too lazy to try though

but seriously if anyone knows a method to 'match' up similar spectral analysis of tons of sounds like in that scrambled hackz video i will suck your dick for life

  Jonas said:

scrambled hacks dude promised he release the sources years ago, will never happen.

yeah and i've emailed the guy probably 5 times never getting a single response.

it makes me wonder if the program itself is a hoax at worst or unfinished and very very difficult to use and requires a shit load of manual intervention at best, i mean the concept seems relatively straight forward and is technically possible with computers today, but he has not once given his software for others to try or shown the source to anyone as far as i know. very frustrating!

i know nothing about this subject. Goes over my empty head.

does anybody here remember 'scrambledhackz' , the spectrum reading thread made me think of it. This piece of software was essentially a beat slicer, that mapped a fingerprint/spectral energy of every slice and then put them into a database. The goal of the program was to 'trigger' different similar sounds by matching up spectrally real-time incoming sounds to the one on the database. A great example of it was taking tiny micro slices of a michael jackson talking, and the dude was able to speak into a microphone and make michael jackson say (in a really glitchy way) anything he was saying into the mic

the reason i bumped this thread, its been about 4-5 years since they showed this (potentially vaporware) piece of 'live software' in a video, but i havent seen anything even remotely close to what it does come out commercially or in any form. Have i missed it?

The Michael Jackson bit was astonishing. If I fester on the idea that there is a secret software which could duplicate what you say, with video footage of someone else speaking, on the fly I will just get depressed. This dude has performed a crime against my hopes and dreams.

http://www.soundspotter.org/ this is supposed to be similar

  Quote

Soundspotter - real-time concatenative audio matching and retrieval

 

Soundspotter is a library for feature extraction and similarity/dissimilarity-based matching designed

for real-time use in music performance.

 

Written by Michael Casey,

I haven't looked into it yet.

i tried this a few times, forever trying to get the patch to work on my mac version of max/msp, kept missing some avlib.dyn file or something

got it opened and running on the PC max/msp and i could never get it to function properly, im guessing i was just doing something wrong

soft push. good intro to ft