openframeworks

ofxComposer

gtoledo3's picture

I mentioned this tech awhile back, and it met a big yawn around the forum, though I had some pretty interesting back and forth with some forum members and other QC users offline.

The system is still fairly alpha but I like it very much, for these reasons -

-OpenFrameworks is a LARGE active community.

-It's an opensource project -ofx and ofxComposer

-It's GLSL shader based, and I think that there's a good future in GL and GL shader tech. Many of us are already well familiar with it.

-It immediately provides video input, and lets us route that to a shader. So much traditional QC stuff revolves around taking an image, giving it an effect, or calling up different "cool visual" scenes. This much is basically already possible.

-There are some great ideas like quad warping and masking on an actual patch level. It kind of has to be seen in action to be appreciated.

Last I used it, it was not truly production ready, it's very new. I could probably setup a file, do a few things on the fly live, but there were some issues with file restoration and loading shaders that have errors, and what subsequently happens.

I like it though, because, aside from the active ofx community, there are so many addon frameworks that lend themselves to the kind of things we do already (CV, motion detection, GUI's, all kinds of stuff).

It's not unimaginable for something like non-shader objects to be represented in ofxComposer, because hardware objects (like camera and kinect) already can be. Now... I am not crazy about some of the ofx methods and having to learn some new abstractions sometimes, but it's not rocket science stuff either. I find myself thinking it's worth it.

I'd recommend that anyone who wants to checkout ofxComposer takes a look at the project - I seem to remember Patricio having his own branch setup with dependent addons to make it compile, at least before the most recent ofx version update. I wrote Patricio tonight to see current status of this (beyond what I can grok from the github) because I'm interested in contributing on this one as time and competency allows. I mentioned this awhile back and didn't get much interest, and I can't blame anyone, but I think it would be helpful for us to jump on some opensource stuff that's similar to QC in function and that we can have a positive influence on.

I'm not recommending this or the Three Node.js techs as any kind of alternative to QC - far from it. I just think they have some overlap with the kind of stuff we do, and it would be cool if QC-ers could influence these technologies early on :-)

ofxComposer

gtoledo3's picture

https://github.com/patriciogonzalezvivo/ofxComposer

http://www.patriciogonzalezvivo.com/blog/?p=585

There's a glsl fragment shader oriented node system that's popped up for openframeworks. At the moment, it's really alpha, has some bugs (don't go using it for production!), but I think it has a great deal of potential, and I'm pretty excited about it. There are already some really great novel features, like being able to see the texture at each patch node in the editor, and being able to mask or warp that result in the editor, in an interaction-like way.

I'm going to try to make contributions to it in my spare time, and if anyone else sees some potential in making this system rock, I encourage them to dive in too!

Patricio is a cool developer, aware of the qc scene, and I think it's an awesome opportunity for the worlds of openframeworks and node systems like qc to collide, hopefully in a good way! It's on github and opensource, so anyone can dive in and tweak the project, fork it, etc.

Also, the somewhat related addon, ofFX, is very cool as well.

https://github.com/patriciogonzalezvivo/ofxFX

... and has some super easy syntax for chaining fragment shaders.

I'm going to try to fill in gaps in the filter selections, add some stuff for fractals and raymarching/raytracing, and help with some editor polish as time permits. I know there are some real whizzes out there on this forum, and it would be cool to see some contributions from people with people that know the benefit and drawbacks from the qc way of doing stuff. I'd kind of love to see the qc scene get behind something opensource like this, and rock it.