Ramen Node Compositer
Ramen, an impressive standalone node compositer by Esteban Tovagliari is now available for testing! This new open-source project looks very promising! Check it out!
Tovagliari writes:
Hi!
I’m developing an open source node compositor called Ramen. It’s not finished yet, but development is going well and I think it can be an useful tool for Blender’s users.
The app runs on linux and OSX. A Windows port should be very easy to do, but I don’t have access to a Windows machine. It will be included in the openArtist Linux distribution.
Some features:
- Loads and saves OpenEXR, DPX, SGI and JPG files.
- Works in linear RGB color space with HDR precision.
- Nodes are animable.
- Basic nodes implemented, channel ops, color correction, blurs, …
- Rotoscoping tools with animation (not finished yet, but working).
- Integrated flipbook for previewing composites.
- All image processing is multithreaded.
There’s the main project page and there’s a development blog with the latest news and some video tutorials here.
I’m looking for testers, contributors, help with the Windows port and opinions in general about the app.
Thank you!
-Esteban Tovagliari
OK, It seems really nice.
But, I would prefer to see this effort applied to improve Blender’s node compositor.
It’s one of the problems (and advantages some times) in open source. Instead of
joining forces, developers tend to start from scratch or fork existing projects.
Because of this, there’s no video editor alternative to Final Cut, compositor that stands
against After FX, audio sequencer comparable to Cubase… and so on.
Thanks god (Ton, actually :D) Blender is one of the few OS packages that is really professional.
this look awsome!!
I personally wouldnt mind to have good OpenSource compositor, Jahshaka never really took of. Blenders compositor is a versitile tool but I believe a dedicated OpenSouce project would be a good idea.
P.S. You cant cram an intire production pipeline in one app, maybe you shouldnt want to. ;P
Nice! I think they did Rotoscoping in Revenge of the Sith… i know they used it with Blue/Green-screen, anyway… :-D
Somebody tell Tovagliari that developing in Blender is gonna be much easier with the new RNA/Automatic UI Layout functionality of Blender 2.50. Maybe he would direct his efforts into our beloved compositor and add it some very powerful functionality :), avoiding him the hassle of developing the internals of such an app (mem management, everything-animatable, etc…)
jedihe
Anyone else getting it build? Tried having a few troubles… with cmake
Very good compositor. Keep up the good work.
Awesome!!! I have been waiting for something like this for such a long time! I can’t wait for a Windows build, assuming there will be one.
Esteban
Thank you, good luck with Ramen, I wrongly criticized your hard work for ‘starting from scratch’ rather than improving Blenders compositor echoing ZareOnes comments here, but they are unfounded, Ramen stands on it’s own two feet. The 3D world extends much further than blenders reach and many would be more than happy to use an OSS Compositor for Modo or VRay or whatever produced exr’s.
I’m looking forward to trying Ramen, although it is not a competition thing between blenders compositor and your own hopefully it will give it a kick in the a$$.
As for Blender being one of the only ‘professional’ OSS I tend to disagree with reference to blenders very basic compositor alone it can’t import multilayer exr’s correclty and blender can’t produce a multilayer exr that is compatible with any external applications such as ILM’s own exr viewer, Photoshop, After Effects, Nuke, Shake, etc.
It also can’t read in and separate the layers of multilayer exr’s produced from any external renderers like VRay, Mental Ray or Modo.
Good luck Ramen.
Thank you for the support!
That video only shows a small
part of the app. There’s much more…
There are valid reasons I have chosen to develop
my own compositor and not work on blender’s.
And there are other valid reasons for users
to use a stand alone compositor.
It’s not a competition.
I see my app as a complement to blender and
blender as a complement to my app.
Same as Gimp, Krita, Inkscape, …
It’s one tool more for the pipeline.
Est.
Can’t build:
/usr/include/boost/mpl/at.hpp: In function ‘typename boost::gil::detail::channel_pointer_type::type boost::gil::interleaved_view_get_raw_data(const HomogeneousView&) [with HomogeneousView = boost::gil::rgba16f_view_t]’:
…../ramenhdr/ramen/flipbook/flipbook.cpp:528: instantiated from here
/usr/include/boost/mpl/at.hpp:42: error: ‘template struct boost::mpl::at_c’ is not a function,
/usr/include/boost/gil/color_base.hpp:386: error: conflict with ‘template typename boost::add_reference::type boost::gil::at_c(boost::gil::detail::homogeneous_color_base&)’
/usr/include/boost/gil/image_view_factory.hpp:105: error: in call to ‘at_c’
make[2]: *** [ramen/CMakeFiles/ramen.dir/flipbook/flipbook.cpp.o] Error 1
make[1]: *** [ramen/CMakeFiles/ramen.dir/all] Error 2
make: *** [all] Error 2
Using boost-dev-all 1.40 on 64bit Karmic
Maybe we should rethink things. Duplicated effort is not good either.
Maybe node compositor backend should it move out into a shared library. Node design processing would be good in krita gimp…. Basically anything handling images or videos.
Lets focus on reducing effort and getting a broader range of powerful tools.
http://clam-project.org/ is working on 3d audio processing kinda node based.
http://code.google.com/p/libmv/ is working on movement tracking.
These are also working on integrating with blender. Yet still being shared to use with other programs.
Node design is visual programming. We are also getting node designed logic system. Basically lots of this node tech could be beneficial to other projects.
Not to criticize… it looks great (at least the output does) and maybe I’ve just spent too much time around blender’s UI but I think a single window interface would be more “normal” ish for a compositor. (After FX, Motion, Combustion, Blender[can’t forget it] etc.)
Maybe I’m having a “Blender’s UI Rocks!” moment. But I think rather than 4 windows a non-overlapping would be nice.
There will be a windows version.
Please use gcc 4.3
gcc 4.4 is not supported by boost.
boost 1.41 will be out in a few days with the fix for this.
Est.
That’s great to hear that there will be a Windows version!
Meshweaver: Rotoscoping is used in alot of movies, not only star wars. But yes, rotoscoping was done in ROTS, the lightsabers, getting rid of the cameraman reflection from C-3PO’s body, compositing characters in, ect.
Good idea but I hate the separate windows. Maybe take a lesson from GIMP’S shortcomings and use non overlapping windows instead. That is why Blender’s UI is so great.
Looks promising! Hopefully a windows port comes up soon.
This is fantastic…Would love to have an opensource compositor to the arsenal…However I cant help wish these additions were coming to blenders compositor.
I second roofoo suggestions try and take the blender route of none overlapping windows. I am one of those that is so glad that GIMP will soon have the option to do things in one window.
I am not one of those that believe all things under the sun should be added to Blender, a focused approach such as this if it attracts more developers could possibly develop quicker.
Goodluck with you project
Fantastic work Esteban! I’m thrilled to see such a promising looking tool emerging. I see a few similarities with Nuke’s interface – a good source of inspiration, I believe.
I too feel that the choice of a non-overlapping window layout with the option to “tear out” windows would add greatly to the workflow. Stored layouts which can be recalled with a hotkey are also really useful additions to the compositing workflow.
I’m installing the dependencies now. I’ll see whether I can get a working OSX build.
there are lots of compositing application available as free. but none of them are competitive with the proprietary softwares currently available. Blender 3D is the one of the professional tool which can get professional quality out put. My suggestion is that it would be great if you can contribute to the existing blender compositing application and make that one better.
Good luck to your project!
Esteban, good job on the compositor. I think it’s good to have a standalone compositor, I support you on this. I only ask, that when you finnish and work the bugs out, if you could help out with blender. We need a hand on the roto tools. Keep up the good work.
Craig, if you get a working Mac OS X build, could you post it somewhere?
IF( APPLE)
# use bundle utilities here
Sort of stopped me from going ahead with modifying the CMakeList.txt file.
Also, if you could post the modifications you did, thanks in advance.
Thank you Esteban for this tool. The Demo is impressing.
For my own, I have no problem using standalone programs, and this one is a thing I really need for ages. It is the reason why I worked on a Roto & Paint setup for Blender some time ago (a blend file using the available tools, not a software because I’m not a coder at all !).
As I’m a Windows user, I hope that the Windows version will be available.
I tried “Silhouette V3” some time ago, and I found it very bugged… even if it is supposed to be one of the best on the market.
A thing that would ve very useful to include in your compositor, is Garbage painting.
Most of the compositors allow to mix Roto and Paint to make masks or erase some unwanted things in the images. Painting on the images is often faster for small corrections here and there, and for wires removal in stunts shots, for example.
I don’t know if you have already thought about that, if it is difficult to code, or if it is already included in your software or on your To Do list, but it would really put the compositor at professionnal level.
does anyone have .deb I am using the lastest karmic 64 and tried to install gcc 4.3 but still wont compile.
is there a chance to get a pre-compiled build? i was never able to compile anything on ubuntu because of the get-this and get-that dependency-hell..
All for having a separate app…
either that or we could always get the MyPaint developers to hack on blender too: Enter the Blender OS!
Blender is a 3D app, not a Swiss Army Knife.
+1 for a single-window UI though!
nice work!
If you are a good coder, join forces with blender compositing team!
Simple! And obvious…
Est is right. There *are* valid reasons for other projects. For one thing, the free/open-source community is not homogeneous – lots of people start lots of projects for lots of different reasons. While duplication of effort seems wasteful, it can actually be positive. For example, if Est were to stop working on this project, there is no guarantee he would find it interesting or enjoyable to be involved with Blender. In that case, neither project would get any attention.
We (the Blender community) should encourage developers to get involved, but we should be careful not to criticize those who wish to invest their own time on a different project.
awesome man!!!! i’ve been waiting for a long time for something like this. i hope someone will make a windows build soon.
Spamagnet it is why duplication happens is important. When something is being duplicated it always time to look at what you are doing and if that is causing the problem.
KDE is great examples of this. Most Image editors in KDE share the same plugin system. Lot of browsers share the same plugins. Best of them all in KDE is Kontact. Multi applications glue together to create one powerful tool. Advantage if a person does not want all of Kontact they can just load up the one part they want only.
Blender has always been a single stand alone program. Very little sharing with other programs. Maybe its time to reconsider that position. Shared engine could make Est job simpler while allowing him to develop his own interface.
Basically what Est has done should have us asking a few questions if blender need to change its design. Now a internal struct change is not a fast process. It could be years before node based systems in blender could be exported. So Est work is good at the current time.
Remember these questions if parts of blender should be exported for other projects to use don’t come up often. The exported lib for reading blender files was a great step forward itself.
ibkanat: I’m having the same issue, but apparently boost 1.41 will be released soon, and then we can complie with gcc 4.4
Duplication usually happens for a few reasons
ego – you get more glory and recognition if you create your own project than if you are a contributor to a big project
code – understanding others code is more difficult than your own so developers tend to use the reason that the other code was ‘messy’
language and toolkit – some people prefer working in java, others in c, others in python; some want a kde or gnome or other toolkit bindings – if a project is not in your preferred language or toolkit you are unlikely to contribute
culture – a coder might feel unwelcome or that his patches are not reviewed quickly enough
architecture – the ideas the individual has might not be compatible with the architecture of the project
LetterRip
@ eguru
There are only a few free or OpenSource projects, not alot.
There’s Blenders compositor and now Ramen. There used to be an app called Jahshaka (or something) thats now called CineFX, but its more of a player in its current state.
If you know any more i’d like to hear about it.
Niels: good point.
There’s also Cinelerra, but honestly, try using it. It really is nowhere near where it should be in terms of usability.
You can pretty much consider the whole Jahshaka/CineFX project as vaporware. It’s been years and nothing stable (let alone usable) has been produced. (I hope they prove me wrong!)
I reckon we need several alternatives for each application. It makes for a more robust suite of tools. So far professional standard video editing and compositing has been lacking in the open source world, with Blender being the exception.
Duplication is a good thing.
tom: Valid points. If the duplication is due to those reasons its natural and fine.
If the duplication is due to parts not being effective to share that can be made effective to share at not much cost is not fine. Reason why I say every time something like this happens that area should be looked at to see if we can share code with the other project. Long term it will assist blender more people looking over the code.
But question still has to be answered should particular parts be spun of into libs.
Freen points out something important there are not many good compositors out there. Duplication creates competition and experimentation. But for the competition and experimentation to bare fruit it has to work.
Now key thing here is if you want to create something like Cinelerra Jahshaka or CineFX. There is no shared backend. If you go to create a new music player you can use a shared backend. If you want to create a new audio editor shared backends.
The lack of multi project in this area that works is one of the major reasons why I am saying it worth the time for blender to look at its self and see if its sane to consider break this part off in the lib.
I am for competition. Duplication is good. Duplication of Effort is bad if the area could be shared without causing pain to either project Ie breaking the points tom wrote up.
Results of duplicating effort is getting a stack of unstable programs with very few working ones.
Of course we all would prefer having a stack of programs that works competing with each other and giving us new features. Remember projects cannot get onto new features until the basics work.