See a sound. Hear a picture. Live in Photoshop.

To a computer, sounds and pictures are both just a series of numbers. What would it sound like if you could apply image processing to audio? What would it look like if you could apply audio processing to an image? With Synthesthesia every Photoshop tool is an audio effect, and every audio effect is a Photoshop tool.

This tiny panel is the nerve center controlling Synthesthesia. The waveform view automatically updates to show the active document and selection. Transport controls let you listen to your image. Import WAV, AIFF, MP3, M4A, and FLAC as images, and export any image as a WAV. Launch the effects processor and reflow dialogs. Utilities like Boustrophedon and Guide at Playhead can be reached from the flyout menu.

The Effects Processor filter provides eight pro-audio effects that you can apply to any image: pitch shift, lo-fi sampling, distortion, glitch, multimode filter, modulation delay, reverb and tape emulation. They are real audio processes equally adept at processing sounds and images. That makes this databending done deliberately: a real effects chain rather than a file opened in the wrong program to see what happens. Kayrock TileOps comes at the same idea from the other end, rewriting the bits of each channel in place instead of playing them.

The preview is live in both senses. Turn a knob and the picture updates in the dialog and on the canvas, and if the panel was playing, you hear the change as you make it. Settings save as presets, and the whole thing records and plays back as a Photoshop action, so a chain you like can run over a folder of images without you.

Reflow

Photoshop resizes two ways. Image Size resamples the picture into a new grid, and Canvas Size leaves every pixel where it is and changes the frame around it. Reflow is a third: it keeps every pixel and its place in the sequence, and changes where the rows break.

This is the same photograph at 1080 pixels wide and at 1082. Two pixels. Nothing was resampled, nothing was filtered, and not one sample of the audio changed: the rows simply break in a different place, so the picture shears into ribbons. That is the whole idea, and it is why width is a musical control here rather than a size. Set the sample rate, the tempo, or how many scanlines a second you want and it works out the width; it can find the tempo of the audio itself and make each row a beat, so a loop lines up with the edge of the frame.