It all started with a simple question: what if a page of literature can be used as the source to create some sort of dance music?
Wouldn’t it be great to have dance music 100% created using pages from Hemingway, J. D. Salinger or Jack Kerouac?
In that case, what kind of mapping could be used to translate words, paragraphs, punctuation and even text mood into electronic dance music?
For this project several parts had to be combined. From the hardware side, a camera or scanner, a computer and a custom structure. From the software side, at least Optical Character Recognition software to recognize the text, mapping and text manipulation routines, libraries to create multi-track sound files and this was just the beginning.
After running some tests in a Linux desktop, I have decided to create a portable machine– 4x5x2 incxhes -named BookSound using a Raspberry Pi, a Raspberry Pi camera, an Oled display, a button, a switch and a custom case and camera arm, both made with Fusion360 and Annet A8 3D Printer.
1. Open a book and place a page under the machine arm
2. Press the black button on the front
3. Wait for the machine to scan, run OCR and map the music - it could take up to 2 minutes :( -
4. Start to dance
The OCR and mapping parts takes around 60 seconds, so the procedure is far from immediate.
During the different steps, the OLED screen display messages about what the machine is doing including words recognized, drum kits to be used, BPM, etc.
The machine determines number of words, words/paragraph, commas/words, sound of phonemes, and it also detects - with a subjective approach - the general mood of the page. All this is combined to create dance music.
Future plans? Make this machine faster, maybe with the new Raspberry Pi and works on the mapping to create more complex and subtle dance tracks, using not only drums but also simple melodies.
More future plans? Present the machine at a literature festival (?) so readers will be able to dance their favorite books.
That's all. Peace and thanks for your interest in the project.
Roni Bandini - Buenos Aires - Argentina @ronibandini
Below you can find a new demo with a page from JD Salinger Catcher in the rye. I have updated BookSound software and now the machine is able to incorporate words from the text into the music track. In this case the high pitch "Ready" was extracted from the page.