Forum Groups
  All forums
    Help & Feedback
      Work in progress
      Finished Art
      Non-Max related

Maxunderground news unavailable

Mental Ray time...
show user profile  LionDebt
What's up you magical, wonderful people... so in my (nearly) two years of using 3DS Max I've been blissfully ignorant of mental ray (and by expensive proxy, vRay). My ignorance was caused by a desire to only focus on 'game art', as such I've extensively played with zBrush, topogun, UDK, etc etc.

The caveat to this - of course - has left my knowledge of rendering; somewhat lacking, to say the least.

My current work involves a pipeline of modelling an airport to CAE/ICAO specifications, unwrapping and texturing (mostly photoreal textures) using A&D materials, flatiron (scene) render to texture, standard lights in conjunction with a daylight system, converting to osg, adding appropriate nodes and sending it off to the client for review and money.

A few things I would like clarified:

1. Inverse-square decay. As far as my (limited) understanding goes, all 'light' in the real-world follows the inverse-square rule (aka. uses a quadratic falloff). I've had very little luck in implementing lights with this method and have always fallen back on using straight up linear decay for results that looked 'better' but not... as real as they should be. So then I delved into messy and mathematical world of Linear Workspace Vs. Gamma correction. In Gamma & LUT settings, if I specify the gamma to be .455 (because 1/2.2 = roughly 0.455) then at long last my inverse-square lights give the results I've been after. As an artist by trade but inquisitive by nature kinda guy, this all feels a little bit too over my head.

If you can sum up this entire linear workflow thing in under 20 (short) words I'll give my right kidney to a charity or underground criminal organisation of your choosing. Otherwise, I'll be happy to know that I don't need to change any other (gamma) settings while performing a render to texture (for my inverse-square lights) to correctly (physically/realistically) affect my scene and output textures (lightmap and diffuse primarily).

2. HDRI environment maps. Say hypothetically I've been sent to some location a client wants built armed with my trusty camera to gather source material. The nature of our scenes are often in a scale of kmĀ². Would it be feasable to use a chroma ball, somewhere in the middle of the location, snap snap, stitch stitch (simplifying for sake of I've typed enough words already) - then plug the resultant map into the environment slot for all reflective objects in the scene? Or would I need to take multiple location shots and use each one logically for surrounding buildings (windows etc). Baring in mind that the end product is an .osg export.

3. There were other things, but it is late and my brain is melting.

read 701 times
5/19/2013 11:05:49 PM (last edit: 5/19/2013 11:05:49 PM)
show user profile  herfst1
Inverse square: gives better shadows. You just move the light closer or further away to control them, else increase the multiplier. Yeah, that's the same as non-inverse square, it's just easier to control.

Linear (as told to me by Donvella, but I might have forgotten some things), set your Gamma & LUT input to 2.2 and output to 1.0. For exposure (for Vray, but I'm sure it's the same for mental ray) set it to linear (1.0). For saving set gamma to 1.0 (or was it 2.2? Hmm, try both). Think that's about it.
read 674 times
5/20/2013 2:22:26 AM (last edit: 5/20/2013 2:22:26 AM)
show user profile  mike_renouf
I struggled with the linear workflow thing for a while, but as far as the implementation goes, its really very straightforward. I'm pretty sure that what I've listed below is correct - it works for me pretty well.

set this up:

Turn on mentalray's photographic exposure control. Use only mentalray lights from the photometric drop down (this is important). Ensure that diffuse colours are never full black or full white. Turn on the 32 bit frame buffer. Now your output format should be a 32bit file format, and you'll have a full 32bit image which is ready for compositing.

If you prefer to have a 16bit image output (e.g. jpg format) as your output. Simply override the default gamma when you save the file - make it 2.2.

read 654 times
5/20/2013 9:12:36 AM (last edit: 5/20/2013 9:13:24 AM)
show user profile  Mr_Stabby
Display gamma: only affects how you see stuff in max, matching it with the output gamma will ensure the closest match between what you see in viewport to what gets saved in the render. Unfortunately max's viewport energy distribution isnt worth shit so at correct values it tends to wash out the viewport while incorrect(lower) value will affect the render window too so either way its shit.. but whatever.
recommend: set it to same as final output(either max's output or if you plan to change the gamma in another program then that output)

Input gamma: needs to be the same display gamma the bitmaps (or you can also specify per bitmap) was created in. This ensures that their energy gets used linearly, which is the way the render engine works.
recommend: if for an example you make a texture with a screen that has gamma of 2.2, you input that bitmap with 2.2

Output gamma: needs to be at the same value that the target medium will work best with.
recommend: whatever the target medium displays, if its a regular computer screen then 2.2 will ensure the most relevant data gets stored.

Basically: the default values are 2.2, 2.2, 2.2, they are right, dont touch them.

as for the linear vs quadric decay: theres no actual decay of energy, its just spread out more. If you arent getting enough energy in the scene theres either something wrong with the materials or your choice of a GI solution doesn't get around enough.

read 630 times
5/20/2013 1:01:13 PM (last edit: 5/20/2013 1:07:04 PM)
show user profile  herfst1
@ Stabbs, but I was told (and found it to be true) that if you set your input AND output to 2.2 you're doubling the effect, hence blowing out the image.
read 626 times
5/20/2013 1:05:51 PM (last edit: 5/20/2013 1:05:51 PM)
show user profile  Mr_Stabby
the input value signifies the value it was saved with for easyness but is actually treated as inverse (1/x) value when applying to the bitmap bringing it to linear space

this topic has caused so much confusion and misinformation that in 2014 all gamma values got locked to display gamma under same value which is 99% of the time what they should be, lol

read 620 times
5/20/2013 1:09:39 PM (last edit: 5/20/2013 1:12:57 PM)
show user profile  donvella
if you dont plan on doing any 16bit+ work then 2.2 output will be fine.

If your not finished with your render ie. its going to be output after compositing then 1.0 output will be fine (exr does this for you) - thus giving you the option to save it out as a gamma 2.0 or 1.8 image if you want if its 'too washed out'.

By default these days my renders are coming out not too bright, not too dark, all 32bit images with curves and exposure for extended ranges in post, more control really to your white/black areas.

Max 2014 gamma control has removed the problem as stabby says, if you set your gamma 2.2 then you have the ability to save out your renders however you like (vray gives you this option aswell) - considering most input images should be converted to 2.2, and your final render will be close to 2.2 - thus the RGB display option in vray.

read 612 times
5/20/2013 1:33:20 PM (last edit: 5/20/2013 1:35:52 PM)
show user profile  Bolteon
Thank you ^

Display and input gammas are rather important with exr based output senior stabby...

-Marko Mandaric

read 604 times
5/20/2013 1:41:24 PM (last edit: 5/20/2013 1:41:24 PM)
#Maxforums IRC
Open chat window