View Full Version : Baking problem, some wont bake
lucasg 06-01-2008, 05:26 PM So for some reason when baking with mental ray there are a few objects in the scene that refuse to be baked out. REFUSE to be baked.
The objects have: No overlapping uvs, deleted history, normals pointing the right direction, had blinn material (problem persists even after changing it to an material mia x).
Render globals: Final gather on (200 sample, 1.3 point density, 10 point interpolation), global illum on.
Bake sets: bakes shadows, no overides, 3 sample, 3 fg, 2048,1024,512's sets.
Its symptoms are: gets up to the point where the output window says .1 percent fg points prebaked and doesnt go any further. 3 textures refuse to be baked, if a few of the ones that do work are batched render it will only bake a few then lock down again.
Tried with network pcs and without and same symptoms. This is very painful since i have a deadline at work with these and it seems theres no LOGICAL reason as to why this is happening. The mesh is clean and simple. CAN it be memory related???
Its currently set to auto memory calc. For both render and batch render.
Also, one chrome like object bakes out uber artifacted. Even after: under raytracing is set to 20, refraction 5, and trace depth 25.....The SAME thing happens with the pcs at work and at home, so its file related setting related.
Please Im in dire need of help...
| |
lucasg
06-01-2008, 06:20 PM
Update on the matter: I found some of the textures had transparency set to one. Fixed that since it shouldnt be see through. ANd it bakes past .1 percent fg point prebaked when left to sit for a long while. Its deathly slowly on a relativly fast computer. Lowered AA and bake sample to 2 and Fg samples under bake options to 2 as well. After bout 20 to 40 mins it nudged up to 4.1 percent fg prebaked. =(
Its budging but its EVILY SLOOOOOOOOOOOOOOOOOOOOOOOWW (2048 x 2048)
OH Buddha!
lucasg
06-01-2008, 09:09 PM
SO when they say new updates are geared towards game and baking speeds HOW DA **** does that account for mental ray bake still being deathly slooooooww?????????
WTF seriously, u dont even know how many "**** you Autodesks" I wrote in the error reports. Countless.
ghostlake114
06-02-2008, 01:42 AM
I never doubt about some issue of Maya/MR because I enter it continuously, BUT, the way you complain will offense many around here.
Your bake is SLOW, not the F*** MR error, you are setting a terribly setting for bake and get the result you want. IT IS NOT MR THING, BUT THE FACT YOUR ERROR THING.
And get in your mind, BAKE MEAN RENDER, so have you ever think a render 2048x2048 high quality with GI will give you 1-2 minute? It is really nonsense!
lucasg
06-02-2008, 11:52 AM
How would that offend? Im not attacking you in anyway or anyone on this forum, quite frankly I respect this forum greatly. Im attacking the process, or not even. BUt I apologize if I did offend.
To me it just seems weird even at lower settings. It seems strange that some objects bake artifacted regardless of resolution....?
My suggestion would be: try to lower settings to bare minimum to pinpoint problem: to see if bake is slow or there is problem creating texture (for example globals: 10 FG samples, density 0,1; bake set: 1 sample, 1 FG, resol. 128x128). Then step by step raise settings ans see what is crippling your bake.
Hope it helps.
lucasg
06-02-2008, 09:43 PM
Thanks for checking it out. Heres what i discovered along the way (mental ray baking):
1. Smaller resolutions, samples..etc do render out much faster.
2. A setting was misused, the second diffuse pass didnt reflect the raytrace settings under FG.
3. Sometimes objects like to remain in there single uncombined state for some reason. (this applied to previous problems similar to this but didnt work here)
4. One mesh=one exclusive material. No mesh sharing same material shader.
So even after this there are a few shiny chrome objects that bake out very strange, very artifacted almost none sensical. Strange banding reflections that are not at all representing the final viewport render (cam is in a natural position within the scene, since its using it as a reference point for baking reflections). SO im completly dumbfounded as to why this is happening.
THey refuse to be baked without any logical reason why. I even seperated the objects and rendered them in a single fashion and with seperate materials--no dice.
They dont have overlapping uvs either. So yea, id be surprised if there was any form of logical explanation for this. Happy but surprised. Thanks to all. Been trying to figure why this happens with reflective items even with high reflection settings for awhile now. Theres just new problems with each new scene.
its amazing
Yes, i have problem with reflections too. Seems like this happens only with objects that have self-reflection (concave shape). Dont know if this is error on my part or with renderer. All normals are OK btw.
Here is just bake sample of spiral shape with reflecting material.
http://shrani.si/t/2I/iZ/4Rr9zVFN/untitled.jpg (http://shrani.si/?2I/iZ/4Rr9zVFN/untitled.jpg)
ghostlake114
06-03-2008, 11:47 AM
Reflection/specular IS view base effect, so it come error as the normal thing.
To prevent it, check orthogonal reflection in bake set to get a correct reflection.
lucasg
06-03-2008, 03:57 PM
Thank you all for your time first off.
Ok so i remember trying orthographic projection at times but that didnt work (i may be wrong).
Im making an observation here about chrome like objects----> They do not bake well if some of the light sources are objects with their ambiance cranked up with FG. Meaning the objects act as ambiance light objects. (not sure)
Still kinda foggy about whats causing the problems, the Ortho projection that Ghostlake mentioned or the ambiance objects? These are the last of the solutions for this strange problem. Ill test this and find out soon, its just the fact i have to wait awhile for it to finish baking to see results.
lucasg
06-03-2008, 07:01 PM
yea check this out, some objects CANT be baked out, hows that for logic?
Ortho projection Nien! Fg and GI turned off Niet!
New material? Nope. Duplicate of the object? None.
Solution that works: Set it on fire :)
CGTalk Moderation
06-03-2008, 07:01 PM
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.
vBulletin v3.0.5, Copyright ©2000-2009, Jelsoft Enterprises Ltd.