Podcast #019: Rendering
Podcast #019: Rending is online and shows some tricks for optimizing your renders and dealing with expansive environment shots.
The Pigeon: Impossible Blog
A blog covering the production of the 3D animated short film: Pigeon: Impossible
July 22nd, 2009 at 11:35 am
Great episode as always. Are you planning to putt some 3dsmax stuff in your podcast or will it be XSI only?
July 22nd, 2009 at 11:55 am
I haven’t used 3dsmax since v2, but the idea is that just about everything in the podcast should be general 3D techniques that should be applicable to any package.
July 22nd, 2009 at 12:57 pm
Yes agreed. Thats why i watch it. I formulated the question badly. But i got my answer from your reply
July 22nd, 2009 at 4:23 pm
Wow, man, this one is the most necessary bunch of tips that I ever heard about render. I have a question: This “Blue Channel” is a render option selectable at final render or is kind a filter and a light that you use to made it?
If you don’t remerber me, I’m the guy who tries to do a 3D movie, but money is still missing somewhere out there. Jojojojojojo
July 22nd, 2009 at 4:45 pm
The “blue channel” is just the blue channel of a motion vectors pass. Mental ray can create this pass automatically, or you can use a custom shader. We used lm2dmv because at the time it wasn’t built into Mental Ray.
July 23rd, 2009 at 12:08 am
Great stuff lucas. I took lots of notes from that one… more research for the to do list! I hope I can make some of this stuff work in Maya mental ray. I’d not heard of .map, or the motion vectors or the depth trick for blur…. I bow before the rendering and compositing master!
July 23rd, 2009 at 3:07 am
Hi Lucas,
I use 3dsmax with the Vray plugin in this way:
1- I ussually render a “.rpf” or “.exr” with z-depth pass inside
2- Then I render the motion vectors using the default scanline renderer (using a common shader for all objects)
3- In After Effects for DOF: I use the rpf files rendered with Frischluft lenscare plugin to apply DOF. Awesome!
4- Then, in AE, I use RE Reel Smart Motion Blur but with the motion vectors rendered in step 2. Properly configured it’s very similar to 3d motion blur and it reacts perfect to 32 bits motion blur (cool highlights! yeah!)
Very Nice Video! Good luck!
July 23rd, 2009 at 3:15 am
Hi Lucas,
Here’s a job I have done using this DOF plugin in AE:
http://www.animatik.com.ar/opium/rtl2_microcity_web.mov
I don’t know if I can post links. Delete if if necessary
July 23rd, 2009 at 5:46 am
Hi Javier Bravo,
I also use 3dsmax and vray pipeline. But i am curious. How do you do motion vector pass in scanline? AFAIK mental ray has a motion vector shader and scanline and vray do not.
So how do you do it? Can you point me to a location where I can read about it or could you post a quick explanation?
Thanks!
July 23rd, 2009 at 10:41 am
Wow, Javier, awesome work. Nice, clean and very funny. Cheers for you.
July 23rd, 2009 at 12:28 pm
Looks great. Yeah the difference between the Frischluft plugin and just a standard blur is subtle without the big discs, but it still makes a nice difference. I wish I had used it on PI.
July 23rd, 2009 at 12:59 pm
Hi Darknemos,
First of all I put a standard material in all objects, then I activate the “Velocity” pass in render elements with a “2048″ value in the Max settings (in that render elements panel).
So it will render a velocity pass besides the simple standar material render which I won’t use. I only would use the velocity pass render which I point to another directory.
Then in AE using the Reel Smart Moton Blur plugin in an adjustmen layer on top of everything (using the motion blur vector plugin) I point to that sequence of files and I put in the Max value the same number I had put inside the render panel in 3dsmax (2048 max value).
That is for make the plugin understands the value of motion displacement used inside 3dsmax. You can use another value. But I found that 2048 makes a nice motion blur!
Regards, Hope it’s understandable.
July 23rd, 2009 at 4:58 pm
Hey Javier Bravo. Thanks for the reply.
Actually i use the same technique for rendering zdepth pass-es but i never thought that i can use scanline for velocity pass also. Thanks a lot man.
@Lucas. Sorry if we crumpled your post with our discussion.
You should open up an discussion forum.
July 23rd, 2009 at 5:02 pm
No worries, I’m glad to see the comments so active. We had a forum that we used during production but they are a huge pain to set up and manage.
July 28th, 2009 at 7:47 am
Some very nice input here. Very valuable for people who haven’t really worked on productions before because if you don’t know better, you try and do everything “in camera” and in one go. I know I did when I started doing 3D at school, but after working in production you start to pick up techniques like the swapping in and out of proxies, splitting scenese in layers, rendering passes etc.
All this helps to minimise possibly the longest part of your project: rendering. I don’t think a lot of people, even on “proper” productions realise just how long rendering takes and the more you can do to split it up or ensure that you don’t have to redo everything, the more sane you will be!
Matt
August 15th, 2009 at 2:21 am
Although not (yet?) a portable renderer like Renderman or MR, LightWave is a viable alternative as well; it has a beautiful renderer – its certain render ‘personality’ reminds me of MR – and it handles very well: motion blur is painless, third-party custom lights, render nodes are unlimited (e.g. 999 nodes) making it a nice choice for those concerned with render node costs. Data transfer is through FBX or third-party solutions (like PointOven).
It does suffer from the lack of inbuilt pass management (available only as third-party plugins at the moment), and the fact that you need to orient yourself to the specifics of the app. In 32-bit it doesn’t handle large scenes to my liking, but others have reported way better performance in the 64-bit version. Another hindrance is that you have to get your models and animation into the app; LightWave won’t read RIB files, for example. That means that the setup process will be doubled: setting up in your main app(s), and then in LightWave.
Nevertheless we used it as a renderer very successfully in heavy effects productions (including feature films), and I believe it’s definitely worth a look, too.
November 18th, 2009 at 5:36 am
Very good!
Congratulations!!!
January 27th, 2010 at 5:56 pm
Hello. I am having difficulty understanding how to do multi-pass lighting in mental ray. In my experience, I have only been able to achieve good images in MR in a still image. There seems to be a flickering (or shadow artifacts) effect when I produce an animation. I can eliminate the effect by storing a photon and final gather map, but when it comes to an animated object in the scene, I get that problem. My main question is, how can my animated objects/characters be rendered with a flicker-free final gather map even with multi-pass lighting? I have been stuck on this same question for about 3 years now and nobody has been able to properly explain it to me. Any help would be great.
January 28th, 2010 at 11:55 am
The quick answer is most likely not. We didn’t use any GI/FG because the mental ray implementation isn’t quite up to production-level, and its incredibly slow. We did everything with occlusion.