Maowei Yu

SIGGRAPH ASIA

[flashvideo filename="video/Siggraph.flv" /]

Programming

jm04js
jm04patch
The patch combine all my tests. It is the basic structure of “Speechfighter”.
The white points are audience’s position. Whenever they touch the good or bad words, it will draw a gray sphere and the text will disappear. If the text is not in the data, it will change the font size and still stay on the screen.

Video Demo:
[flashvideo filename="video/speechfighter01.flv" /]

Programming, Thesis Project

This step, I combined javascript, my data analyze patch and an extra jitter patch called “cv.jit.blobs.centroids” .
The extra patch is for getting data from live video to make my work really interact to the audience.
Below are the patch and script:
jm03js

Programming, Thesis Project

This is the new development. I combine text input patch to my javascript.
Whenever I type a new text it will show up on the screen.
The tricky part is writing a counter to separate texts to different element.
function settext(t)
{
jtext2ds[counter].text(t);
counter++;
}
Below are the patch and script:

jm03js

Special thanks to Peter! I had a create time discussing programming with you

Programming, Thesis Project

If you saw my crazy patch in the “Data to jit.gl.text2d 02″, you will agree why I want to use Javascript to coding multiple text fields.
Below are the patch and coding that I tried to create texts with javascript and MaxMSP:

jmfile
maoweitext01.js
Basically, I created jit.gl.text3d, the jitter object, inside of javascript and this time I also create a array of string.

var jtext3d = new JitterObject(”jit.gl.text3d”, “sfighter”);
.
.
var texts = ["Why", "does", "never", "work", "?", "how", "it", "because", "nothing", "is", "easy"];
then I get the data of people’s movement and sound for MaxMSP and sent to javascript to control the texts movement.

Here is the result:

Programming, Thesis Project

Ok, I arranged the patch to sub patch and create a gate for the different words.
Arranged Patch:

Whenever you type a text and press space key, it will switch the direction to the next object (p text).

Programming, Thesis Project

I decided to use jit.gl.text2d as the main object to draw my texts on the screen.
Here are some test about the object.
In this moment, I would like to connect my data analyze patch to the jit.gl.text2d patch to see if it’s work.
It is not complicate as it looks like.
data to jit.gl.text2d patch:

Fortunately, It works. see!!! I typed good. Did you see that!
Result:

Next step, I realized that one text object can only content one text ,then how to make multiple texts on the screen?
That’s easy! As a normal person, I will try to duplicate the patch……I won’t attach it here, just imagine “data to jit.gl.text2d patch” duplicate 100 times. I’ll tell you the result: it works.

Result:

Programming, Thesis Project

This is the patch about how to identify the good words and bad words in my project.

The upper part of this patch just is for nature speaking or nature typing. I set space key to send a bang, so we don’t need to press enter to send the texts to the database.

The “coll object” construct my data inside(see right image). Whenever the text information go through it, it can send the different texts to different receiver. For example, when I type happy, according to the data in the “coll object”, the program will recognize the word happy and send any related information such as texts, numbers and bang which I already setup to the positive receiver.

Now, the tricky part is that I should figure out how the treat the words which is not in the “coll object” database. It seems no this kind of outlet….

Programming, Thesis Project