Image sonification

General Discussion Other

caffeinec..

in General Discussion • Other • 10 months ago

Hi all,

I'm quite new in the area of image processing/sonification. I have been doing a lot of research in colours and image processing and mapping

I was wondering if it is possible - from an image I take the rgb values and convert them into hsv... and from there I plot histogram and generate sound from the histogram?

My main objective is to map color - hues specifically into different sound frequencies.

Replies(5)

velvetkev..

It's certainly possible: I did a piece a while back which mapped the RGB values of an image to synthesiser parameters (pitch, filter cutoff and something else... it's been a while). It would be easy to convert that to HSV; I'm not sure how you mean to apply the histogram but I'm sure it's doable. Have a look at the MInim library that comes with Processing.

caffeinec..

Re: Image sonification

10 months ago

mapping colour histogram to sound was just an idea..I can't say I know much about histogram. I've only recently learned histogram related to image in C++ image processing

Another idea I was thinking, is that to implement mapping between hue values and sine waves. So it takes input an image file and then returns a sine wave sonified version of the input image.

so I will convert image rgb into HSV first and then from I could extract the H.

I was thinking I could create a mapping between hue (which will have values between 0 and 1) and frequency (play around with different frequency ranges and mappings..) but still not sure I can go around about it.

velvetkev..

Re: Image sonification

10 months ago

On a furious deadline right now but I'll try and dig the stuff I did out for you- probably be a couple of days. Should give you a place to start.

caffeinec..

Re: Image sonification

8 months ago

Any idea?

nose_1.1

Re: Image sonification

4 months ago

I would also like to see whatever code you have made in the past, velvetkevorkian.

I also have a project that involves sonifying onscreen pixels, but probably a different take than most... Any hints for doing so would be awesome.

Top Reply