Can I use software to paint with my voice?

October 1, 2009 12:41 AM

Can I use software to paint with my voice? I'm a High School teacher, but not an art teacher, and I've had an idea for an artwork project in but I don't know whether it is possible.

The background is that there is an art event coming up when pupils and staff in the school are being asked to produce original and varied ways of making your mark or art.

I got this inspiration yesterday, but don't know if it is feasible. Please advise me.


It is in 3 parts:

1) Software on a computer which captures the human voice (or at least sound of the voice).

2) Software which takes sound a provides a "visualization" based on the pitch or beat of the sound.

3) A way to combine the two so that it feels like that you are painting or making pictures purely based on your voice. This could be screen shots of a visualization, but it may seem better to come out as a print out or jpg.

I am aware that this may be too advanced and complicated to succeed, but thought I would ask.

posted by ndaguiar to computers & internet (14 comments total)

The first thing I think of is using Sound Recorder on Windows or Garage Band on a Mac and then bringing that file into software that does visualizations (Windows Media Player, Winamp for Win / iTunes or iTune-offshoots for Mac). I'm not sure if there are any "paint-like" visual add-ons or if there is any software to emulate as much.

There is software out there that creates sound based on art/shapes, though.
posted by june made him a gemini at 12:49 AM on October 1

Ze Frank made a flash widget called "Voice Draw" that does this.
posted by slightlybewildered at 12:54 AM on October 1 [2 favorites]

He also has some examples of produced pictures (and user links to more examples in the comments) in an earlier post.
posted by slightlybewildered at 12:58 AM on October 1 [1 favorite]

Check this out!

http://blog.makezine.com/archive/2009/07/tablet_guitar_creates_art_and_music.html

They offer schematics also.

http://www.workinprogress.ca/projects/guitare-a-crayon/

He's looking to sell it although the article is from July 2009 - who knows, maybe he'll sell it to you.
posted by Aegean at 1:07 AM on October 1

My colleague Susumu Harada developed a research project called VoiceDraw that allows you to paint using vowel sounds. It was originally designed for people with motor impairments who can't use a mouse or stylus to draw.

I don't believe that it's been widely released, and there's no downloadable version on the web site, but if you contact Susumu he may be able to hook you up.
posted by shaun at 1:23 AM on October 1

Engineers use a variety of tools to visualize signals with periodic content (audio being a good example). A spectral analyzer shows you a signal's the distribution of energy (volume) across a spectrum of frequencies. You could see the differences between different instruments or human voices singing the same note. The musician Aphex Twin has drawn pictures in the spectral plots of his songs through clever software techniques. Baudline is great for this, but it only runs on Linux (find a nerdy kid in your school and I'm sure she'll be able to set this up for you in an hour or two tops).

Another interesting tool is an oscilloscope, which lets you plot signal amplitude against a repeating linear sweep to get something like this. The effect is very striking in real life, because of the oscilloscope's practically-instantaneous reaction time. You can also set it up to plot two signals against one another, and see their relative phase. Here is a really groovy demo of this kind of setup being driven by a modular synthesizer. You could just as easily do it with two microphones - which would be particularly fun if you can some well-trained singers who could try to hit certain intervals (unison, perfect fifth, octave, etc) and stay in phase with one another.
posted by phrontist at 1:29 AM on October 1

if you can find some well trained singers
posted by phrontist at 1:59 AM on October 1

This kind of thing can be done with something like max particularly with jitter. I am pretty sure supercollider has graphics capability but I have not messed with it... regardless it can interface with cocoa.

All this stuff is programming oriented... either through code in sc or connecting boxes in max. But they allow you to take sound input, analyze it and process it in whatever way you can imagine.
posted by sundri at 2:53 AM on October 1

The most basic way of drawing with the voice would be to send a binary signal which was interpreted by a machine or process at the far end. This reminds me of some of the semaphore signalling systems which were used as the precursors of a fax. Whilst you could use a computer to assist with all this it might be more fun and instructive to do it without one.

The steps are:
1. Select an image to be sent and break it down into pixels.
2. For each have the sender establish whether it will be a high or low value (use as many bits as you choose really).
3. Send the signal to the receiver by whatever means you choose - shouting over a wall perhaps.
4. Receiver re-creates the image.
In this case the voice is being used as a tool of transmission rather than directly as a tool of expression. In that sense it is no different from a digital camera - but the "art" is introduced by having a low resolution system with interesting artifacts.
posted by rongorongo at 3:05 AM on October 1

It's probably too expensive and has too much of a learning curve to be a real option unless you have a lot of time, but for the sake of putting the information out, there's always Max.

A very rough way of thinking about it is that Max lets you more or less do anything with anything. It'll happily let you do things like feed an audio stream straight into a video processor so you can "see" what it sounds like.

There's also an iPhone app called Meshmerizer that has some audio-reactive functions. Not a lot of specific control, though. It's more of a visual toy.
posted by Su at 3:05 AM on October 1

Aaaand sundri beats me to it.
At least I had a secondary item *grin*
posted by Su at 3:06 AM on October 1

If you want to do something custom, I reccomend either Max/MSP with Jitter or Supercollider and this interface to Processing.
posted by phrontist at 3:26 AM on October 1

Thanks for all the advice. I will be meeting with a couple of "nerdy" kids from school tomorrow to see what they think about Baudline, and will probably look into Ze Frank and Susumu Harada.

Will try and post back how I get on.
posted by ndaguiar at 9:58 AM on October 1

Why not hook volume/pitch controls into tablet-style pressure/tilt readings, and use a regular input device (mouse, touchscreen, wii remote etc) to move the cursor?
posted by Tzarius at 6:19 AM on October 3