11/12/2008

Tech Talk

So yeah, it's real and it's live. You want to know what's behind the scenes technically?

When it comes to the IT-hardware, here are the specs:

1 1.83Ghz Mac mini
2 Logitech QuickCam Vision Pro webcams
1 JJ-Labs BMC microphone
1 d'feldt Audioengine 2 speaker


Besides from the hardware directly visible in the cams, and the computer running them, there's a bunch of additional stuff needed for the rig:

Loads of acoustic sounddampening-foam, a 200w low-heat Lohuis Nurture Lamp, a lamp timer, some fans to keep the air circulating and the temperature regulated and a green nightlight. All young ones need a nightlight, right? Joking aside, the nightlight operates at waveslengths that don't interfere with the plants sleep-cycles, so that one's mostly for us (and you).

So you type in some text .. what happens then? The text is posted to the webserver (running a PHP application) and saved in a message queue. The mac computer running the experiment polls that queue constantly for any new messages, and relays them to a script which plays them using speech synthesis on the line-out port. Connected to the line-out port on the mac mini is the d'feldt speaker, which plays the speech synthesis for the experiment-plant.

Here's where the experiment ends and the experience begins. We could stop right there and still let the experiment run it's course, but we wanted to keep you guys involved in the experience, you are the ones contributing the loving encouragement after all. so..

..Right in front of the speaker, we have the JJ-Labs microphone. The mic picks up the sound from the speaker along with a bunch of noise from the fans e.t.c. (sorry about that, but we wanted this to be the actual sound and not a faked simulation .. so we have to live with it). The sound is then streamed through an icecast server, right to the flash-site in your web-browser.

The images are streamed separately. A new image of each plant is taken every two seconds, and sent to the webserver. Every hour, the server builds a 15 second timelapse clip from the images captured so far. The resolution of the timelapse will change dynamically through the experiment to cover the entire growth period of the plants. (6 weeks condensed to 15 seconds will naturally not have the same resolution in the time interval between images as 1 week of growth condensed to the same 15 seconds).

The timelapse clip is the one you see when you first enter the site, so don't forget to check back in for your daily 15 second summary of the experiment so far!



Stefan adjusting the camera.

No comments:

Post a Comment