Been thinking about ways of making use of video data combined with sensor data in a way that scales to lots and lots of different sensor nodes at once. I haven’t really thought much about this design space before, so it’s been a fun mental exercise. I threw this together this weekend partly to see if I could make actionscript do it (answer: yes) and partly to see if it was interesting (answer: maybe).

You’ll need a webcam/microphone for this to do anything. Works great with the built in Apple ones. Make sure to make some noise to see what happens.

Known issues:

• Sometimes the wrong camera is selected by default. For Macbook Pros, you want “USB Video Class Video”. You might need to experiment.
• Sometimes clicking “allow” on the security dialog box doesn’t trigger the “you now have camera access” event. If you select “allow” and “remember” and then refresh the page it should work.

For people without cameras + microphones, here’s what it looks like with no sound.

With medium sound:

With lots of sound! (I’m leaning into the mic here to try and get it this loud)