Converting spectrogram images to sound

ARSS is an opensource commandline program that can produce high quality black&white spectrograms, but more importantly: it can chew up images and synthesize sounds treating these images as spectrograms.

https://sourceforge.net/projects/arss/

It has some limitations - the input image has to be a specific type of BMP file, and it’s a rather barebone commandline tool - however it does the job and allows users to convert image files into sound files.

I used that tool when to design some sound effects for games, and it is a very unique tool, giving you abilities that no synthesizer provides.

I think it might be a great idea to use the code and make this program a part of Audacity’s toolset. Importing image files and synthesizing sound out of them would be a super cool thing. Being able to export the spectrograms out can be great too, because you can convert a sound into an image, then process the image with image-specific tools and then resynthesize the sound back. Have you ever tried pixelating a spectrgram or sharpening it to listen to the results?

Using the commandline program is a bit tedious, and also the need to convert all your images prior to the resynthesis is also tiresome.
Adding a capability to read major image formats (JPG, PNG) and convert them to grayscale automatically would be great.

Also R,G, B and A channels of an image file could be treated as separate audio channels this way - color of the pixels would translate to the stereo image of the sound (pun intended).

What do you think?

See this topic: https://forum.audacityteam.org/t/databending-picture-files-loaded-under-u-law-not-working/43620/1