v3ga

The video above shows a two-layered tech demo of Vision Factory :
- one scene with webcam kaleidoscopic background (shameless plug from Flight404), distance to the plane+rotation are driven by midi device.
- that good ol’ ripple effect applied as a filter on the background.
- scene with rgb debug cube, driven by javascript. [see script here]

I just wrote a module for handling and sending midi messages. It is essentially a wrapper around the cool rtmidi library written by Gary P. Scavone. In Vision Factory, a module is a special plug-in that can extend core functionnalities of the main application, and is registered as a global object so that every other plug-ins can access its interface.
I also integrated the management of frame buffer objects (FBOs) for being able to draw in a texture. It is merely used by the core engine for main rendering, but the api allows their creation for directly drawing onto the graphic card.
Ah. I just discovered this week that I had important memory leaks within javascript … I have absolutely no idea what causes them, which is very annoying. I guess I’ll have to look more carefully at the documentation.

I am in the process of integrating Javascript in Vision Factory through the use of Spidermonkey engine. The primary idea was to have an easy way to bind properties in an animation in order to avoid the fastidious (and often frantic) common recompilations every time one single parameter/formula had to be changed. It turned out that the integration went nice and smooth, so I decided to use scripting for the application set-up instead of dealing with configuration files. It actually makes things a lot clearer (I would dare to say ‘human-readable’), and most important easily maintainable.

Now, I am at a point where I am tempted to write a Javascript interface for every part of Vision Factory core engine. That is, not only scenes getters/setters would be accessible at runtime, but also basic services like sending/receiving network messages for example. Hmm, I have to think more about it as I want to keep things simple and flexible…

Anyway ! I just grabbed a little video showing how scripting integrates in the main application. The main component of Vision Factory is the layer manager, which allows to stack and blend scenes (with optional associated 2D filters) for rendering. Like every classes in the core engine, it exposes some properties that can be modified through user interface / network / scripting. I thought it would have been really cool to have a simple string property that would be interpreted as a script. Thus, it would be possible to inject javascript code in real-time into the application through network or UI. I sat down and started to wire the different parts together and … it worked !

The video shows a very basic rendering with two layers (a square and a rgb cube) whose parameters are live controlled by javascript. All is running on my local machine, with a processing server pumping up the script from hard-drive every second and sending it to the main application through network. Fun :-) With this configuration, one possible application would be a kind of collaborative live coding, each ‘participant’ having an assigned layer to change properties on.

I’ll try to design a better rendering example this week-end with easing, textures and stuff. Thanks for reading.

I will have the pleasure to come back to Saint-Ouen (near Paris) at Le centre de Ressources Art Sensitif (Mains d’oeuvre) to animate a second Processing workshop.
It will last two days, from 7th to 8th of April. The targeted audience is mainly beginners who have very little or no prior knowledge of computer programming. Coding practices will be explored through the creation of interactive graphic sketches.
More informations can be found there.

J’animerai le 7 et 8 avril prochains un atelier Processing au Centre de Ressources Art Sensitif (Mains d’oeuvre) à Saint-Ouen, près de Paris. Il sera principalement destiné à des personnes ayant peu ou pas d’expérience en programmation. La découverte de l’outil ainsi que les pratiques de programmation se feront à travers la réalisation de programmes interactifs graphiques.
Plus d’informations (tarifs, plan) sont disponibles sur cette page.

I updated the page dedicated to projects using BlobDetection lib I wrote a while ago for Java/Processing.
Most recent ones feature a multi-touch table built by the hyperactive guys at r3nder.net, while the other projet - called Living Interface and made by Marc Nonnenmacher- is a study of mixing together a real single life form with computed boids to produce sounds and graphics (video here).

I set up a registration page so that it’s possible to be notified when the software is released. If everything goes smooth, a beta version should be out around May 2007 !

Vision Factory is a multiplatform software (MacOSX / Windows) written in C++. It is aimed at creating real-time interactive visual animations by the programming of reusable and sharable plugins. Thanks to its modular architecture, one can quickly build complex scenes by using data resources (images, webcam feed, sounds, osc/midi messages, web services, …) through the use of dedicated built-in libraries.
The software relies on several third-party libraries and has numerous facets making it a rich framework for the experimentation and the creation of visuals. It is well suited for installation and/or vj shows for example and was primarly thought for easily handling a multi-computers rendering system by providing native network capabalities.

The image above is courtesy of Hudson-Powell using Vision Factory.

So this is it. I finally got my hands on a brand new MacBookPro notebook, being my first Macintosh computer I have ever owned. The switch from PC/Windows was less difficult than I first thought, though I am still feeling a bit uncomfortable with the keyboard and shortcuts (” Where the hell are the ‘{}’ chars ? “). Of course, this is very little pain compared to the difficulties I have been encountering for some time now with my slowly-dying three years old laptop : broken keyboard and dvd device, usb ports switching on an off randomly, unbearable fan noise and heat, degrading Windows operating system, and on and on … Anyway, I installed BootCamp, so I am still able to develop Windows applications which was very important for continuing developping VisionFactory on that platform.

I brought it to the second Processing workshop I animated in Craslab (Paris) last week-end. It went super well, again with lovely participants and organizers. Cheers to them ! For the occasion, I had coded a very little application based on Dan Shiffman SMS library and directly inspired from his particles example. It’s basic and brute force codeing, but still it is fun. By the way, it produced small Oooohs and Aaaahs when I showed it to the workshop audience, guess it was time to show how to install and use libraries in Processing.

Download source code.

Photo and video were shot by Mike, who just released his new website yesterday, exposing his masterpiece paintings.

I will be animating a workshop in Saint-Ouen (near Paris) at Le centre de Ressources Art Sensitif (Mains d’oeuvre) and it will last two days, from 2nd to 3rd of December. The targeted audience is beginners who have very little or no prior knowledge of computer programming. Coding practices will be explored through the creation of interactive graphic sketches.
More informations can be found there.

J’animerai le 2 et 3 Décembre un atelier Processing au Centre de Ressources Art Sensitif (Mains d’oeuvre) à Saint-Ouen, près de Paris. Il sera principalement destiné à des personnes ayant peu ou pas d’expérience en programmation. La découverte de l’outil ainsi que les pratiques de programmation se feront à travers la réalisation de programmes interactifs graphiques.

Plus d’informations (tarifs, plan) sont disponibles sur cette page.

I was randomly browsing the web when I came across this video. This is a very dynamic and funny talk given by Luis von Ahn showing how the processing power of a human brain can be used to improve image search on the web by playing fun games (ESP game and Peekaboom).

This brilliant idea is currently used in Google image labeler.

If only my college lessons had looked like this !

Some of my interactive works will be exposed in the concert hall named Florida in Agen, the town where I live and work. This will be part of a global installation focused on the use of free and open-source softwares for creating music, animations, films, …

Among other, Banjamin Cadon from Labomedia will talk about Pure Data, and my friend SevenFive will play live. All is free and opened to public so if you happen to be in the area, just come and stop by.

I just released a new website for BlobDetection lib (for Processing / Java). Sources code were released after a bit of cleaning and a page featuring various interactive projects using it was added. I’ve been recently working with Philip Worthington to enhance the library core functionnalities and fix some sneaky issues, it will be put online very soon.

Lately I’ve been quite busy with refactoring and redesigning the core application of Vision Factory. I’ve come up with several ideas that would have been dirty patches rather than clean implementations in the actual environement… So, at some point, I decided to rebuild the whole stuff from scratch. I’ve been studying design patterns which have been a great help for seeing things “out of the box”. I plan to release a beta version of the application for the beginning of next year, which will ideally come out with a SDK (Win+Mac) and an API documentation.

At the moment, I’m pretty excited to put all of this together and see where it goes. Time will tell.