Kyatric : having dealed with WebRTC support in browser those last weeks for plugin dev, I can say that it's urgent to wait ;). Not everything is working, FF support isn't on par, etc...

birdlantern : I have done some research for you ! You can apparently use a Flash file to speak to an OSC server. I'm finishing a plugin now for a client, and working on Spriter plugin also, so I won't be able to help you with that for the moment, but there's something there : benchun.net/projects/flosc. It's a Java server, bridging with a Flash client to exchange OSC messages. I don't think there's a way to send OSC message directlty from the client, because you need to forge packet that aren't websocket or HTTP, so outside of the Javascript's range (and I don't know if somebody tackled the work to forge packet with Flash's Socket lib...)

Ashley : thanks a lot !

birdlantern : you need to use a bridge, because JS is only going to speak HTTP/HTTPS or Websocket. So you need a PHP/Flash/Java/whatever proxy, to send informations to, which in turn is going to send OSC packets.

kiyoshi : the problem with WebGL is that it's not supported on mobile platform, while CSS3 3D transform are.

Kyatric : I'm surprised, because when checking on the MDN, I see that you can call .pause() in FF on media elements.

Rukiri : I have some code laying around to have a mode 7, but since I'm sinking in the amount of work I'm currently comitted to, I think you are going to finish it before me.

My suggestion : don't do any calculation xith the Canvas itself. It's going to be too slow. Use a CSS3 3D transform, by extracting live the graphics from a Sprite, pasting it in a <div> setting the Perspective projection, and having the Sprite skewed and sheared automatically by the GPU.

Is it possible to have work at least on Firefox & Chrome ?

When looking through the code of the "Audio" plugin, I saw a

etSuspendedfunction.

Is it possible to use it as a "pause" function, or is it for something else ?

Savvy001 : here's the link of the whole thing : http://dl.dropbox.com/u/1412774/AudioSpriteExample/AudioSpriteExample.zip

I'm really sorry I didn't answered you before, the workload is heavy !

aurum34 : there's a function "Delete gesture of type" + name_of_the_gesture.

If you check my original demo, you can see the name of the 16 basic gestures implemented in the library (the names on the screenshot in the .capx).

For the moment, you can't get the direction of a gesture. Just know that there's a gesture. If I have the time, I'm going to see if I can add that.

Just to say : you can have a per-pixel collision detection by using a webgl shader. The GPU boost can help you lower down the time needed for the calculation.

However, when I saw the last poll, I realized that for a lot of people, webGL shaders weren't a prority...

steenosterby: I'm currently using speech.js to do something like that. When I'm dug myself out of the work I'm doing, I'll release the plugin here...

Edit : did you saw that, in the link you gave, it's not TTS, but speech recognition ?