Renderman now free

[John Bigboote]

Let's hook-up to this as an alternate renderer!

 

http://www.animationmagazine.net/vfx/pixar-releases-free-noncommercial-renderman/

[robcat2075]

It's interesting, it's potentially powerful but you'll need to be seriously coding-aware to get something out of it that is beyond what we do in A:M already. It's not an artist-friendly renderer like A:M has.

[Rodney]

Hmmm....

 

Let's

 

 

That's short for 'Let *us*' right?

 

 

When do we want to start the project Matt.

 

The journey of a thousand miles starts with the first step.

(Although I'd say the first step will be for us to begin a bit of research.)

[Rodney]

I'm curious if there is anyone that frequents the forum that is familiar with the command line use of Renderman.

It seems to me that might be the way forward with A:M and Renderman.

 

Added:

Here's a little information on using the standalone Renderman (which doesn't need Maya, Katana or any plugin):

 

https://community.renderman.pixar.com/article/293/using-renderman-without-maya.html

 

 

Aside: Way back in the day A:M use to have a RIB exporter...

 

 

While it's not a high priority to use Renderman I've started the process of installing it.

Let's see if there is anyone tenacious/stubborn enough to beat me to the first A:M/Renderman rendering.

(Using the Maya or Katana plugin as a conduit doesn't count!)

 

Update: Installed

(Rendering with the command line will likely require installation/launching of 'it'. For those taking notes:

 

"it" is a robust framebuffer/render view window, offering complete floating point support and a powerful and flexible catalog, as well as a fast and powerful imaging tool that is capable of production-qualilty image manipulation and compositing usually found only in high-end standalone products.

 

It can be run via plugin or command line.

[nemyax]

Way back in the day A:M use to have a RIB exporter...

Why was it abandoned?

[Rodney]

Why was it abandoned?

 

Mike Hough (the author of the AM2RIB and AM2RIB2 plugins) stated he was frustrated with (i.e. didn't like) the RIB format.

His effort was back in the 1999/2000 timeframe (before the current HXT plugin format).

 

Nicolas Yue (I think that's how you spell his name) also wrote a RIB exporter called MasterRIB.

I don't know anything about that plugin.

 

Added: I see that Nicolas is or at least fairly recently was involved in some RIB/github activity.

 

Also of possible interest to you... regarding the RIB format it has been said (by PIXAR):

the only files that support the full feature set of OpenSubdiv currently are Renderman's Interface Bytestream 3.0 spec. The RIB file specification is supported by many commercial renderers (PRman, Mantra...) but is not very suitable for DCC applications.

Alembic and FBX each support a subset of the full feature set, and they can be used to transfer information across various DCC packages (so long as your assets do not rely on some of the missing features). Time allowing, we will probably submit contributions to Alembic to make it fully compatible.

*emphasis added (Note that I assume the reference is to OpenSubdiv 3.0 (currently in Beta release)

[nemyax]

3.0 is the version of the original (1988) RISpec. In subdivision surfaces, it supports creases of varying hardness and some options for handling boundaries. (I made a subdiv-only RIB exporter for Wings some time ago.)

Of course, an exporter for A:M wouldn't generate subdivision surfaces. It would probably use Bezier patches. Hamapatch does that, for instance.

[robcat2075]

I notice that Pixar currently has four openings for Renderman workers, more openings than any other part of the company.

 

Even Pixar has trouble finding people who can make Pixar Renderman work!

[fae_alba]

So my first thought was..cool get Pixar quality renderings. But then don't we already? So my next thought is what does renderman get A:M users other than bragging rights to say something like "rendered to Pixar quality" or some such, which you of course can't really say without Pixar taking issue with it. Add to that if you have any desire to say create a short and monetize it in some way (we all do, admit it!), then you can't use renderman. So what does it buy us, other than an intellectual exercise of playing with Pixar tool sets?

 

I had looked at it, but without some understanding of its benefits I shied away from the learning curve.

[nemyax]

if you have any desire to say create a short and monetize it in some way (we all do, admit it!), then you can't use renderman

You can. For $500.

 

 

get Pixar quality renderings. But then don't we already?

Do you?

[robcat2075]

Martin once told me that A:M could render anything you see in the first 6 Pixar movies. I'd say that's a valid claim and that is pitting A:M's renderer out-of-the-box against Pixar renders that had to be spruced up in post with compositing to get them to what you see on screen. If you were willing to do compositing and post-processing with A:M renders you could get even further into the Pixar line of movies.

 

The overwhelming reason people are disappointed with the images they get out of A:M (or any 3D app) is that they haven't lit and textured the scene with the same human expertise as the people at Pixar do... and they don't recognize that that is the reason.

 

I presume (I don't know for a fact) that Renderman has some fine nuances like better SSS or a more comprehensive water simulation solution but better SSS or water is not the reason Pixar movies are successful.

 

 

[Rodney]

My first attempts at using Renderman...

 

I suppose if one starts one has to start somewhere so it might as well be with Rabbit (or a reasonable .obj facsimile of him).

[robcat2075]

Is this like environmental lighting?

[Rodney]

Is this like environmental lighting?

 

 

More like pasting an environment onto the model.

I tried to get the surface to reflect the environment.... failed.

[robcat2075]

I salute you for at least getting an image. I would not have know where to start.

[Roger]

Speaking as an average AM user, Renderman is probably over my head (at least for now).

 

And I'd have to agree that Robcat is most likely correct for the reason people get frustrated with a given renderer.

Lots of people go with whatever the default setup is for lighting or a very basic setup and then wonder why they have poor results ( lighting is something I struggle with and lighting a scene properly is a discipline in and of itself).

[John Bigboote]

I like the concept of alternate renderers for A:M... not saying that A:M has a bad renderer, just that alternatives are great, even the Maya and C4D folk are always interested in optional renderers, so we should keep topics like this going- I hope without rubbing noses the wrong way. I look forward to seeing your 1st test, Rodney. Speaking of samples... here is a render I made using Element 3D V2, which is a GPU renderer which works right within After Effects. I animate in A:M using all Hash's belles-n-whistles, export a obj sequence into E3DV2 and assign materials, lights and camera in AE/E3DV2. I had troubles figuring-out the magilla of getting the characrters decals into a UV world (3 decals, the Tireman on his hat, the white highlites in his eyes and the tread on the bottoms of his shoes...) so I ended-up just modelling them in 3D instead. I am no good at UV mapping in any app.

 

At full HD 1920 X 1080 resolution the the frames took 13 seconds each. The kicked-up dust used Trapcode Particular in AE.

Tireman JumpIn 1small.mov

[Rodney]

I look forward to seeing your 1st test, Rodney.

 

Too late for that... you've already seen my first test (and second, third and fourth).

 

At present I'd say the direct route to Renderman is above my patience level. This coming from someone that loved typing in DOS commands and creating batch files back in the good 'ol day. For those taking notes, in the short term I'd say a shorter and more successful route to Renderman would be to go through the non-commercial version of Nuke, which I've only experimented with enough to know it works at a basic level. Export your assets out of A:M as OBJ... import into Nuke... texture and light... render out via Renderman (there's a Renderman node for that).

 

The problem on my side is that I have little need (outside of satisfying curiosity) for an external renderer.

This isn't to say I have no interest though. I've even been using Netrender these days.

A:M has a great renderer.

 

I should have added: The most difficult part of setting up Renderman to render A:M models is converting your models/scenes to .rib format.

[robcat2075]

Speaking of samples... here is a render I made using Element 3D V2...

 

Looks great, but is there something there we can't do in A:M?

[John Bigboote]

 

 

Looks great, but is there something there we can't do in A:M?

 

 

Not really- other than the swift render time and the 'live' after effects environment... other than that- it's just a little...different. Different reflections, different look to the AO... easy-peasy HDR lighting with realtime feedback... other than the HDR, just 1 klieg light in AE...

[robcat2075]

I would say that the reflection in that shot is "wrong". It's brighter than what it is reflecting.

[robcat2075]

Hey, Matt, what ever happened in the fight over who would get that character?