This seems like the place to share this.
I made a short python script that converts wav files to pngs. It's kinda slow, due to my inexperience with coding, but it works and the quality of the sound files that come from the pngs are good enough. https://github.com/mashrow/png-audio
You need scipy, openCV2, and numpy libraries and need to put the script in a directory with a subfolder named "output", that itself has a subfolder named "output" and then run it.
The attached file is audio in png form from https://www.youtube.com/watch?v=Gz5K43W8oYY
As far as size is concerned the png is generally smaller than uncompressed audio but not other audio containers. The wav file I used to make this png is 35 MB while, the ogg file i made from that wav is only 1.5 MB