Big Scene And many cameras

#1

Hi everyone!
I have a very important issue.
So I have a very big scene with many many displacement textures its 4k 2k and maybe 8k its maximum. Basicly this is a Indoor Architect scene.

What I need, Can I do baking and do render from other cameras faster?
I have a lot cameras more than 60 cameras.

I think first maybe I to do a lighting baking, In another render… Maybe its in Mental ray or Vray, any program can to do that than use Renderman for Maya for faster render Displacenent with baked texture?

Basicly I have scene in Maya.

Any propositions?
Very important for me your findings and suggestions

P.s For finally I need a render passes…

Thanks
Best Regards
Evgeny.

#2

up :bounce:

#3

If you are using Renderman then why don’t you try LOD control while rendering ,it would be efficient way of handling memory issue as well as faster way to render displacements.

Also if you bake light first,don’t you think that would affect your displacement render as on displaced surface light would scatter differently.I am confused.
My Lighting Reel

#4

Hey, thx for answer.
So, Generaly I have a question for, how to do that, I have 1 scene with all objects statics, and I have a lot camera animations. How I can to do render faster, because if I rendering from camera every frame, this is very many time spend for calculation FG. Maybe to do that if possible with Irradiance particle few frames then locking a Irradiace map. then render faster because all calculation using hard for displacement.
So I need Best idea for doing render, because I have a lot animations

I have idea to do with baking, but I think I cant get passes from that, need Diffuse specular zDepth passes for composition…
I use mental ray in Maya…I know displacement in Mr very hard… but usefull

I have a machine for this :

I7-930 3.7Ghz
24Gb Ram
Rampage II

p.s I try Rendering this scene in Renderman for maya 3.01, but I cant setup good Lighting…Hard for me, how rig light…I use only IBL for indirect light, without any light in scene.

#5

The best way to do a render is in passes.
Generate direct and indirect light passes along with specular and reflection etc.
It is always a good idea to generate a FG map for camera animation and use it for rest of the animation.

I don’t know how baking would help you in current situation though you should try baking Occlusion.

Mental Ray is not bad in handling displacements ,All you have to do is use Mental Ray approximation editor and try using different approximations for different objects for better control.

I have no further info about you scene so can’t suggest anything further
You have a pretty decent machine so I don’t find any issues related to rendering.

Yeah, if you have renderman you can also use point cloud system for faster rendering.
Though I have no access to renderman 3.01 but heard it is faster in raytracing too.

#6

hey, I thinking about not FG map its only for example, because FG maps need to do not once its need to do increase frames, maybe jump every 5 frames, its very hard, I thinking about maybe Irradiance map its not same, but same situation and different engine shuting from camera and get a IR map, I don`t know how merge this maps because its works only one Camera - one map, But I want to do that with multiple shuts with another sides… - maybe its importons if I can use that, I need ideas for this.
so for example I have 1000 frames:

  1. Rendering FG map only jump approximate 5-10 frames
  2. Rendering few Frames under difference angle, I have apportunity render 1 frame and have a 1 map.But I cant merge that((
  3. Render Point cloud - I think it best way…

I have one question about point cloud, I don`t understand how I can to do a point cloud from scene, and how I can render this… maybe Have any information or pipeline for this system - how to)

I see images from point cloud this is what I need I think, this is baking object vertex information.But how to?))

Best Regards
Evgeny

#7

As I told you I can’t suggest you anything without looking into your scene and also you yourself is the best judge, for the method you want to use for rendering by either generating FG or Irradiance map.I have not worked with IR maps but I think it also works in same fashion as FG ,you generate one map by combine different camera angles possible.

If you are interested in rendering through point cloud there are various posts related to it and the methodology behind it is also the same you need to generate point cloud for your GI ,OCC etc.
(http://kikipoh.com/vsfx755/PointCloud/PointCloud.html),(www.chrisnabholz.com/…/tutorials/Brickmap_Occlusion.pdf)

As you have renderman you can also take benefit of brickmaps.

#8

hey Again, Thanks for links its usefull.
I work in progress for fill scene because this is scene have many areas is empty. (wall,ceiling…etcetera)

there images What I have now



Test renders Displace only


11min

13min
Maybe if You see that You have ideas for this)

Thanks.
Best Regards
Evgeny.

#9

Nice modeling .
From the current image it seems you are more dependent on indirect light as I am unable to see any other light.
yes you have too many objects with displacements,I recommend the use of proxies MR as well as Vray has this support .
Also bake stuff on areas where camera is not so close like book shelf etc.
Here Normal maps are your friend.
Add other lights too , it might help you getting better result in a faster way.

#10

hey, Thanks.
Whats is Does means about MR proxy, You mean about clone objects for example big Columns small columns?..Wich function do this(MR Proxy)?
Yes it is at this situation I think Normal Map good idea for optimizing scene only for Far distance objects.

Thanks
Best Regards
Evgeny.

#11

http://www.youtube.com/watch?v=SgiUqrZU6Cw

#12

Thank You so much!
:wavey:

#13

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.