Pixar to give away 'Toy Story' 3D RenderMan software

[markarjun]

The 3D rendering software behind films such as Toy Story, Monsters Inc and Harry Potter is to be given away free for non-commercial use.

 

RenderMan, which is developed by Pixar, has faced increased competition from rival animation programmes such as VRay and Arnold.

 

Although Pixar, which is owned by Disney, produces its own films, it licenses RenderMan to rival studios.

 

The company has also cut the price of its software for commercial use.

 

In a statement, the firm said it would release a free version of RenderMan "without any functional limitations, watermarking, or time restrictions".

 

"Non-commercial RenderMan will be freely available for students, institutions, researchers, developers, and for personal use," it added.

 

Ian Dean, editor of computer graphics magazine 3D World, told the BBC the move "could be seen as a reaction to the rise of alternatives such as Arnold," but that Disney/Pixar are also looking to "build a community".

 

He added that RenderMan was "very important at the higher end of the entertainment, animation and visual effects industries".

 

The new version of the software is due to be released ahead of the Special Interest Group on Graphics and Interactive Techniques (Siggraph) conference in August.

 

http://www.bbc.co.uk/news/technology-27677712

 

Sounds great, is animation master easy to export to renderman? or would extra work be needed to export it to renderman. I know this is not a commercial license but it would be great to use it as another option.

[robcat2075]

Pixar Renderman is a powerful shading and rendering environment but it is a very difficult one to use.

 

Anyone thinking of using Renderman should study this manual and ask themselves, do i really want to have to do all that? It would be necessary to digest and understand that manual to get something going with Renderman.

 

Pixar Renderman Manual

 

The great results Pixar gets for their films is the result of hiring extremely capable Renderman gurus and paying them to spend lots of time on small details.

 

 

Substantial coding skill will be necessary to get the sort of custom results that make people admire Renderman...

 

The Shading Language is a C-like language with extensions for handling color and point data types. A large number of trigonometric and mathematical functions, including interpolation and noise functions, are provided...

 

 

The shading and texturing environment in A:M is dramatically easier to use and can already do 90% or more of what is typically done with Renderman anyway.

[Rodney]

Well said Robert.

 

I wish someone was interested enough to dive deep into the technicalities but that's a tall order and then there is the maintenance aspect on top of all that as well.

What works well today is not guaranteed to work in the future so these technical solutions have a long tail.

 

That was the problem with the previous RIB exporter in that the author wasn't invested in upgrading it as Renderman itself improved and added features.

This isn't a fault of the author but simply a result of the law of diminishing returns.

 

At first blush the use of Renderman seems like a great idea but implementation of great ideas is rarely easy.

There is a reason PIXAR is releasing the non commercial license to everyone... to encourage usage.

This in an age when increasingly even PIXAR is looking for better and more cost effective solutions.

 

Concerning export to RIB (or Renderman friendly formats) the process is fairly straightforward.

You just have to have a utility between A:M and Renderman that converts resources.

As A:M can export to a variety of formats many utilities can be used.

There are several methodologies that may not be considered by others as easy will prove trivial to those who speak Renderman.

If interested in this path, pull up a chair and get ready to dig deep.

Where there is will there is way.

 

If interested in Renderman, and while awaiting its non-commercial release, you may also be interested in CGkit.

[Rodney]

For the brave and bold I'll add this into the mix:

 

While I'm sure there are many ways to convert to RIB format here is fairly straightforward way via a program that should be updated again soon if it's author's recent news release is correct.

The program is Ayam: http://ayam.sourceforge.net/

 

There are some issues with the current release (which I believe is about two years old), especially in that it doesn't like large amounts of triangles so those should be tesselated first (there is a write up on that in their forum).

 

I recommend importing the model/scene then saving out to Ayam's native format then exporting that out again to RIB.

This seems to resolve some issues with writing directly out to RIB format from the original OBJ, etc. (i.e. sometimes it doesn't want to do it... period)

 

Disclaimer: I don't know any particulars about using Ayam with Renderman beyond that.

 

Concerning Ayam itself the author wrote in his documentation:

RenderMan Interface Bytestream (RIB) export is the most important export module of Ayam and in fact justifies its existence

 

So, it seems likely that the exported files will work well with Renderman.

 

Note that you don't need to use Renderman to render RIB files. The following can also be used: Aqsis, Pixie, Gelato, RenderDotC, 3Delight, Air, Angel (and of course PRMan).

[fae_alba]

Ya know, I spend a lot of time on the road with long nights stuck in a hotel room with nothing to do. Not practical to lug my animation laptop along with my work laptop, so I can't work on my A:M projects. But i can (and do) a lot of coding. I just might take a crack at renderman, just so we can say that A:M plays nice with Pixar.

 

Here's the caveat, if I am spending a lot of time taking the deep dive into renderman, I may need to ask assistance with my A:M projects to keep them on track. Does that sound a tad bit self-serving? You bet! Since renderman is largely written in C, it shouldn't a huge-ish learning curve, building an interface between A:M and renderman that can be easily maintained and kept up to date with renderman is another matter. It might be a good plan to build an interface between the two and release it as public domain, in keeping with Pixar's stance with renderman. This has the added advantage of visibility to A:M.

 

This will take time, and I really, really, really want to see Papa Bear come to life, so if I do this, I may be asking for some help with my animation projects. That being said, I'm on vacation for the next week so I will be somewhat out of touch as I cavort with the Mouse in Orlando.

[markarjun]

Hi Robcat, Rodney thanks for the advice! you're right i agree Renderman is a colossus to use. I use a lot of animation software but without doubt the best animation tools are those found in animation master.

 

They're simple and straightforward everything works as you feel it should (far better than maya or blender in my view. unfortunately clients and co-workers are generally biased towards what software they prefer to use. In an ideal world i could import the polygon models that my client gives me (which makes my co-workers happy) and animate as it is in animation master( which makes me happy) without reworking it as a patch , then export it to a third party render software(which makes my clients and co-workers happy).

 

I know the guys at soulcage have done something like this but I would love to rig a polygon model in animation master alongside my patch models..Dont hate me for it!

 

That doesn’t mean that I don’t believe animation master is great at modelling or rendering, I think that if it allowed more people to interchange and use whatever worked for them even more people would realise how really great a program it is and use it more.

 

Best of luck with the programming fae sounds promising!

[robcat2075]

...then export it to a third party render software(which makes my clients and co-workers happy).

 

what is that getting that you can't get from rendering in A:M?

[Rodney]

This will take time, and I really, really, really want to see Papa Bear come to life, so if I do this, I may be asking for some help with my animation projects.

 

If the priority is to see Papa Bear come to life I would forgo the commitment to Renderman.

There just isn't going to be enough of a return on investment to make it worth the effort in the short term.

 

The primary reasons I encourage others to extend out to Renderman is 1 ) it can be done 2 ) fully customizable rendering would be useful 3 ) it's nice to have options.

Beyond that it's something of a fool's errand.

 

I would love to rig a polygon model in animation master alongside my patch models

 

If the polygon model is created in a piece-wise fashion we can do some of this now... even better in some cases as the polygon parts can be combined with the spline meshes to form a mix of hard (poly) and soft (spline) forms that can then be animated (it can help to think of the polys as bones and the splines as muscles). Exporting these new creations out however... um... that's something that has yet been done to store such a thing because there really is no file format for such outside of A:M... and barely within A:M. Then there is the rendering of this new creature as well... something only A:M could do at present. But if you prefer to live within the A:M ecosystem, life is good. And all of this is steadily getting better.

 

And the best news of all (for those tethered to polys) is that over time (perhaps lifetimes) the polygon world is expanding to overcome obstacles within it's own ecosystem... increasingly moving toward splines. This brings other options and compatibility to light.

 

This then is the real impetus for investigating what can be done through Renderman; to take advantage of what millions of lines of code and years of work have (at least potentially) wrought.

It's that potential that intrigues... often against better judgement... and forms the premise of a potentially pleasing outcome.

[robcat2075]

Like Rodney, I don't see a lot of advantage in Renderman. Someone could devise a way to seamlessly export A:M models and materials and lights to Renderman but the rendered result would look exactly like it does in A:M because the mathematical shaders available in A:M wont' look any different if they are computed in Renderman.

 

For example, "Lambert" shading is available in A:M. It's one of many particular ways of calculating the brightness of a surface as it turns away from a light source. Just about every Renderer supports it. It's math and math doesn't change (it shouldn't) because someone else did it.

 

 

 

Now if the goal is to get access to shading methods that are not in A:M now... that's a valid goal, but how would the A:M user use them? How many of them could once it is available?

 

In A:M? Possibly as a new set of materials that one uses in A:M and are recognized only by going to Renderman. I can imagine someone getting that to work but that would be A LOT of new stuff to program and debug. A material to control RM's version of subsurface scattering would be just one example.

 

I suspect most of these additional things RM does would be too many parameters for most A:M users to become good at.

 

I don't believe that anyone's A:M project will look better by being rendered in Renderman without a very serious reworking to use the additional stuff that RM offers and most A:M users are just scratching the surface of what A:M does already.

 

Using RM well is something that very devoted lighting and shading professionals make careers out of. By design, A:M is built to be not as demanding to get a result out of.

 

If the A:M rendering environment is too hard for a user, Renderman is not the answer for them.

 

 

 

If RM is somehow dramatically faster at the same tasks as A:M is, that would also be a reason to investigate it even if it were just to directly transfer A:M projects to have their math done in RM. But I don't think there is any claim that it is a dramatically fast renderer.

 

 

 

All of the stuff I've said above applies to other 3rd party renderers: How do you get it to work with A:M without requiring the user to redo all the lights and cameras and textures and materials in that other rendering environment?

[robcat2075]

The ideal would be that a 3rd party renderer appears like a choice in the options panel. You pick it and A:M uses it for final renders and on-screen renders and you don't need to export anything anymore than you do now to do a final render.

 

It would faithfully interpret all the native A:M materials and lights and whatever when those are what you wanted anyway and somehow there would be additional interface elements in A:M so you could set and control and keyframe the other things the renderer offers when you want those.

 

It would be a hell of a lot of work though.

[jakerupert]

It might be interesting just for one or two users around here, but far more usefull would be a connection to or a similar thing to a realtime rendering solution like unreals udk or marmoset offers.

We could spill out animations so fast, that it would be fun to do "movies" finally.

And the quality would be sufficient for you tube pay channel .

If you do land a "boxoffice hit " that way, you could still let rerender it in topnotch quality later on.

 

Just my two cents.

 

Dont look for the grapes that hang too high, but look for smaller possibilities in reach .

[Rodney]

Jake reminds us of yet another reason why a Renderman solution won't work.

A streamlined Renderman pipeline would only satisfy a very small part of the 'need everything' solution.

 

With regard to Unity3D and Marmoset (and the Skyshop portion which requires U3D 4.2 or greater) what isn't clear to me is why more folks aren't already publicly using Unity3D with A:M.

That would be a good step in the right direction.

 

Regarding why one would investigate Renderman over U3D/Marmoset the primary issue with pressing toward Renderman over U3D/Marmoset at this stage is one of price.

PIXAR has just revealed that for Renderman non-commercial usage (which covers a wide swath of A:M Users) there isn't any.

 

With all the talented folk who frequent the A:M Forum I do hope that someone is pressing into U3D (at least the free version). If not they should be. We've got a gaming forum area set aside for it.

[robcat2075]

If I had a good idea for a game I'd investigate the game engine stuff but I'm not much on that front.

[Rodney]

If I had a good idea for a game I'd investigate the game engine stuff but I'm not much on that front.

 

While creating a game is ideal there are other things that can be accomplished with A:M and Unity3D.

Something as simple as telling a story for instance.

This would represent an alternative way of 'rendering' our assets from A:M.

 

Hey, back on topic!

 

Seriously though, it is fairly trivial to get A:M assets into Unity (as either 2D or 3D assets).

In the not too far flung future perhaps we can host a few courses in using other programs in conjunction with A:M.

A basic course (the first of many if there were to be sufficient interest) might not have to cover much more than getting resources from A:M into .

While I could be wrong, and often have been, I don't think we'll see much progress on developing a course for using Renderman with A:M until at least 2015.

The free non-commercial license hasn't even been released yet.

[markarjun]

...then export it to a third party render software(which makes my clients and co-workers happy).

 

what is that getting that you can't get from rendering in A:M?

 

 

If I am working on a project in its entirety then of course animation master is my choice, the rendering is fast and decent and most of all i'm used to it. But if i'm working with others then I don’t often get a choice. the models are created by someone else and the rendering and composting is completed by someone else. While i do love the all in one nature of animation master, enabling people to bring in polygon assets easy and giving people more options to render in external programs would mean more people can experience and use animation master.

[robcat2075]

If I am working on a project in its entirety then of course animation master is my choice, the rendering is fast and decent and most of all i'm used to it. But if i'm working with others then I don’t often get a choice. the models are created by someone else and the rendering and composting is completed by someone else. While i do love the all in one nature of animation master, enabling people to bring in polygon assets easy and giving people more options to render in external programs would mean more people can experience and use animation master.

 

Not being able to use polygon models, I understand that problem, but that's long before you even get to a render.

 

But the render... what is that lacking that some else can't work with it? Once they are rendered image files are image files.

[frosteternal]

....Like Rodney, I don't see a lot of advantage in Renderman.

In A:M? Possibly as a new set of materials that one uses in A:M and are recognized only by going to Renderman. I can imagine someone getting that to work but that would be A LOT of new stuff to program and debug. A material to control RM's version of subsurface scattering would be just one example.

.....

I suspect most of these additional things RM does would be too many parameters for most A:M users to become good at.

 

I don't believe that anyone's A:M project will look better by being rendered in Renderman without a very serious reworking to use the additional stuff that RM offers and most A:M users are just scratching the surface of what A:M does already.

 

Using RM well is something that very devoted lighting and shading professionals make careers out of. By design, A:M is built to be not as demanding to get a result out of.

 

If the A:M rendering environment is too hard for a user, Renderman is not the answer for them.

....

 

 

If RM is somehow dramatically faster at the same tasks as A:M is, that would also be a reason to investigate it even if it were just to directly transfer A:M projects to have their math done in RM. But I don't think there is any claim that it is a dramatically fast renderer.

 

....

 

All of the stuff I've said above applies to other 3rd party renderers: How do you get it to work with A:M without requiring the user to redo all the lights and cameras and textures and materials in that other rendering environment?

 

Getting it to work with A:M's settings etc wouldn't be the worst thing ever. Current A:M materials lights etc would ideally be converted on the fly to Prman shaders.

 

The main advantage that PRman has is its a VERY extensible renderer. Shaders can be written to incorporate any new tech or custom effects that a TD and an army of coders can imagine.

 

However, there are also other advantages:

-Motion blur. Prman motion blur is outstanding. (Not always physically accurate, but usually "looks right")

-Complex scenes. The renderer can power through pretty much anything you throw at it, given enough time and computing power.

-Hair rendering. Again, PRMan excels at complex scenes. Hair can be rendered directly as shaded curves or geometry. There's no reason A:M hair couldn't be exported at render time.

-Network rendering optimized. Because it's built for large scale production, distributed rendering was built in early on.

-Displacement mapping. PRman displacements are relatively fast and very robust.

-Constantly Updated. This is HUGE. The renderer is CONSTANTLY being improved. New lighting models, shading methods, features, speed improvements, stability.

-Industry standard. (Supporting PRman in some way definitely improves the prestige of a 3D software. That seems pretty minor, but A:M has a pretty bad reputation of being a closed system. Adding Renderman support would be a great answer to that.)

 

I'm not sure that's it's necessary, but with this new noncommercial license, many hobbyists are going to want their software to support their ability to play with this new toy.