credited to Clint Rodrigues for the model and texture
step1: modified and added some extra models and textures.
step2: setting up lights(LWF) and atmosphere in maya ( mental ray) look development, rework all the shaders, texture refinement.
step3: arranging render layers and passes for nuke composite.
step4: composite in nuke added in starglow effects and color adjustment, matte painted sky in ps.
step5: DONE enjoy
1st time using nuke for compositing, i am AE user ;p.
very smooth composite in nuke for my very old laptop.
by the way this piece idea inspired by Hayao Miyazaki film spirited away. ~
Hey Fedja. Nice render. How did you convert the supplied Maya software shaders to Arnold. Did you manage to keep the normal maps and displacement when converting? I found it was a pain to convert all those textures one by one.
Yes very pain to convert all those textures one by one
I found the script on the http://www.creativecrash.com/ script is free.
when converted into ai standard.
All textures are associated exactly.
only does not apply to displacment
When I prepared the scene for the lighting challenge, I hand-converted each shader from VRay to Maya Phong to allow those folks without VRay an opportunity to light the scene.
I figured ambitious participants would convert for their desired renderer (and shaders). Here’s some links to the creative crash website; reviewed these when I was preparing to hand-convert the shaders.
Ah man this was fun! First lighting challenge here. Thanks to everyone organizing it, Clint Rodrigues for providing us with this wonderful scene and to Emir Bojorquez for his free Halloween asset you can find in my entry.
Used 3ds max, Vray and photoshop for this and I really wanted to give a completely different feel to the place, so I went for a bit late Halloween topic :) Hope to get some constructive criticism
Ahqi, stunning image (y) I absolutely love the way you sutiated the arch and clocktower in the far background and the rim light on the wires on top, I think it is a small nice detail, but adds so much. A question though, why did you decide to have such strong flare on the lights further away from the camera?
Well done to everyone else, it feels really inspiring to be part of this challenge
I am not able to render even a preview render of the scene with all the textures. I have seen a lot of you are able to render. How did you guys manage to do it? I need some tips. I am able to do a grey render but the moment I turn on textures, it takes forever and I am not able to render a single frame to preview my lighting even at low res. So far I have tried the following:
delete all the birds
delete all the flies
try and assign shaders to fire hydrants (it complains some of them do not have shaders assigned)
i have only one directional light in my scene with no FG
Maybe it’s my machine which could be the bottleneck:
Quad core AMD Opteron
4 GB Ram
Maya 2015 SP6 + Mental Ray
Any suggestions? How does one handle such heavy scenes?
Images in scene use here about 2.5 Gb. Objects in scene use about 1.25 Gb. If you are on Windows, it will use about 2 Gb more. So you would need about 6 Gb of physical memory to run things more smoothly in this case.
You could resize the images at about the half and you would use only about a fourth (625 Mb) of RAM for images, which will reduce the whole RAM usage at about 3.8 Gb (including the OS). Other solution (or additionally) is to convert grey scale images (and perhaps normal maps) to indexed color (palette files) by reducing the RAM usage of these images to a third without even changing their size. Perhaps some diffuse albedo maps might be also converted to lower color depth. A better solution would be of course increase physical memory.
@gerardo Thank you for your response and analysis. That was very detailed. Just curious, how did you come up with the memory consumption figures for textures? I analysed the sourceimages folder and the textures themselves are just 128 MB. Did you calculate these based on image resolution?
@gerardo I have one question, what do you mean by “indexed color”? Can you please provide more details? I know “map” files which mental ray can produce, optimizes memory usage it seems, is it the same?
It's reported by the 3D package. Images in the disk are compressed (images folder here is about 270 Mb), but in the 3D package they are loaded into memory decompressed (about 2.5 Gb).
It’s different. than .map files which load from the disk only the rendered bitmap portion. Guess they are more useful for very big images, but not so useful if the texture will be completely visible in the rendered image (like it’s the case in a long shot for example).
Indexed colors have a reduced the bit depth. original textures are 24-bit RGB (or 8-bpc / 16,777,216 RGB colors), their indexed version is just 8-bit RGB (or 256 colors for whole RGB values in the palette map), or even less if you want to. Since colors are reduced from more than 16 million to just 256, this compromise reduces also the memory usage. Do not expect great results for close-ups but it may be worth to try them in normal, bump, specular, reflection and albedo diffuse maps. Not good for displacement (scalar/vector) maps, or textures with lighting baked on them.
In Photoshop you can convert the textures from RGB color to indexed colors by going to Image=>Mode=>Indexed Color. In Gimp 2.9 (free) you can go to Image=>Mode=>Indexed Color Conversion. It’s very recommendable to use the dithering option for these operations.