cool 
can you tell the folks who code the iphone app mill colour that a vignette would be awesome to have my mails just bounces when mailing their feedback email 
regardig the compositng in blender, i hope someone like esteban or someone who can optimize the code. because it goes kinda slow now working with fullHD material is impossible , my blender crashes.
I’m pretty confident thats going to be optimized for durian as they are planning to work with 4k plates. 
If you’ll do a windows build I really give it a try.
I’m a bit bored by the linux dependencies problems,you are obliged to keep the system always updated.
@est77: a good and open compositing alternative is the thing i’m missing the most, so thanks for working on a possible solution!
have you heard about lumiera: http://lumiera.org
it’s a foss NLE/compositor in early development state. afaik it’s also going to be node based. if you read their brainstorm result, it looks like those guys really put a thought into their plans. interresting stuff!
i’m not ordering/begging you to team up with them (not at all), but it would make sense to pull on one string. at least those projects could inspire themselfs 
i’m not doing much compositing by now, but i once tried fusion 5 at a workshop and i love it!
anyone noticed: vfxDesktop’s reels looks 100% like autodesk flame coincidence out of question.
but since he doesn’t show a sign to probably open it some day, i won’t spend to much time thinking about it.
so long
@est77: Please reconsider the option to unite to Blenders dev team and do some great stuff for Blender compositor. I think, in this way, will be much easier and faster to have a free compositor to Nukes level.
I don’t see why everything in the OSS CG community has to be associated with Blender. A good OSS node-based compositor will benefit everyone, not just Blender. There may be other reasons he chose to go with QT as opposed to Blender, familiarity with the toolkit, likes it better, etc. Right now Blender’s compositor is a mess, its not easily extensible, the devs have other issues to work through such as a node api to make it easy for devs to work with, and there is no real masking tools in blender. He may not want to have to fix other peoples issues that he is not familair with, deal with politics that he is not privy to, and chose to build his own compositor because he feels that he can do a better job and not have to be affiliated with any one OSS project. Maybe he just doesn’t feel like having to learn how Blender works and instead wanted to focus on something he knows. So no I don’t think he should have to fix Blender’s compositor just because he himself is interested in creating a compositor. His long term goals may not be the same as what the BF wants or needs, he may want to make some money off of his work one day, he may want to create a node system that is easily extensible (as stated in one of his goals). He may want to include other libraries that Blender may not want to deal with. All things that should be taken into consideration before the Blender “faithful” jump down his throat for not conforming to our community needs. I say let him work on his project and if he decides to join Blender, great, but its not a requirement and we shouldn’t discourage his work just because he doesn’t fix our mess.
We’ve had luck before with the sculpt tools but that doesn’t always have to be the case. Besides I think its time Blender focuses on core groups of much needed features (modeling, rendering, materials) instead of worrying about the shoving in the kitchen sink at the moment.
Who’s discouraging it? I think everyone wants him to continue this whether or not he decides to work on Blender’s compositor. Plus, it’s his project. He can do what ever he wants with it. I don’t think anyone “jumped down his throat”, I just think that they were curious as to why he wouldn’t just work on Blender’s compositor. It’s not that serious, lol.
But why would that even need to be asked. Like I said not every project related to cg has to be affiliated or concerned with Blender’s welfare. If I were to create my own video sequencer and posted a thread here letting people know about this interesting new video sequencer, I wouldn’t want people coming at me with “yeah that nice, but what about making Blender better because I really need Blender to be better”. Blender is great but its not the end all be all. Maybe his goal is to write something on the level of Nuke or Shake, and if he were to include his compositor in Blender that may never happen as people see Blender as first and foremost a 3D application. XSI has a pretty damn nice compositor but that hardly gets mentioned or used when people speak of it. Maya used to have a compositor when it first shipped but it was dropped because compositors for the most part aren’t interested in firing up a 3D application to do what they do. A compositing application is enough if its any good.
Now there are benefits to blender having a nice compositor. Its 3D, most compositing apps have 3d features in them for simple tracking layering, etc. Its a one stop shop for Blender.
I love all the application features of Blender (the UI!) and the new stuff with 2.5, i really like to edit video and composite in Blender rather than ''Round Tripping out and back. But if it’s worth it to go out then I guess I’m there.
Hi!
Sorry for taking so long. My original reply got lost.
First, thank you for the responses.
I removed many dependencies from the code. Now you only need OpenGL, GLEW, Intel TBB,
OpenEXR, Boost and FFTW3. Most of them are in modern linux disributions. I think I can also include boost files and glew in my code. I built a Mac OS X version, but it needs some more work.
I started writing the app before blender had a compositor. In fact, this release is going to be
the third major version. The previous ones were based in FLTK and were internal tools for my job.
I decided to open source it because as an end user of compositing software I’m a bit tired of politics,
marketing, apps dying without replacement… the key is freedom. Microsoft or Apple could buy The Foundry tomorrow and kill linux versions of Nuke or replace it with something you don’t like.
As a side effect I would like to find help to develop it. The ideas behind ramen and blender’s compositor are very different and it would be difficult for me to contribute.
I don’t see how some things I want to do can be integrated inside blender.
Also I don’t start from zero, I have too much code written.
I don’t intend to replace Nuke or Shake. The idea is more a practical system that is extensible and
perform well the most common tasks. And I think I’m not far from that. Animation and ImageIO are
the places that currently need more work.
Writing new nodes in ramen is easy. For example, simple processing nodes takes minutes to write.
The other important goals for the program is that it should be able to share data with other apps ( roto curves, luts, color corrections, …) and the HDR color precision.
Est.
love to see the development of your software… :)… didn’t see a good compositor for linux…
what, shake and nuke aren’t good enough for you? 
talking here about open source… err… forgot to mention Good Open source compositor in my previous post… sorry.
Esteban is still going strong and making great progress on his compositor. Here’s his latest post and how it interacts with Blender.
http://ramenhdr.blogspot.com/2009/12/multichannel-exr-support.html#comments