Shouldn't I get identical results when applying gamma in post than when i'm not?

[andrewrpreza]

Ok I have this issue, I asked a few weeks ago and I didn't get a reply.

 

The thing is that some noise becomes more apparent when I gamma correct instead of baking it directly into the image.

 

 

Please look carefully at the attached image.

 

They both were rendered with the exact same settings, except for "don't affect colors".

 

The one on the left was rendered with adaptation only, with the sRGB button ticked on the VFB. The image on the right was rendered without the "don't affect colors" option ticked, and obviously the sRGB button disabled on the VFB.

 

You can see noise becomes more apparent when applying gamma in the VFB. I get the same results when I bring them into photoshop.

 

Shouldn't I get in theory the exact same results? Why does it differ?

 

Thanks

[Francisco Penaloza]

That's a very small image I can't see !@#!&@#

now regarding your question, yes it should be different, because the famous gamma, I am guessing that you are working with a gamma fixed to 2.2 if so and you do not check "don't affect colors" you'll see in your VRay buffer a SRGB image with gamma already applied (different than clicking the SRGB button) and this is wrong because there you a breaking the "lineal work flow" I mean not wrong, but just breaking it because if you save that image what ever gamma fix you do later will be a double gamma.

Now if you check "don't affect colors" how its name says, it will calculate the GI with a fixed gamma but it will keep your image in linear space, then your render will look dark until you click the SRGB button in the VRay buffer or adjust the gamma in photoshop After Effects or so.

 

so in resume, set gamma 2.2 linear color mapping, click "don't affect colors" check SRGB button in your VRay buffer, happy with your render?, save image as exr 32 or 16 gamma 1.0 do your magic in Photoshop, after effects,combustion, Shake, Fusion, Nuke or what ever heck you use.

[andrewrpreza]

That's a very small image I can't see !@#!&@#

now regarding your question, yes it should be different, because the famous gamma, I am guessing that you are working with a gamma fixed to 2.2 if so and you do not check "don't affect colors" you'll see in your VRay buffer a SRGB image with gamma already applied (different than clicking the SRGB button) and this is wrong because there you a breaking the "lineal work flow" I mean not wrong, but just breaking it because if you save that image what ever gamma fix you do later will be a double gamma.

Now if you check "don't affect colors" how its name says, it will calculate the GI with a fixed gamma but it will keep your image in linear space, then your render will look dark until you click the SRGB button in the VRay buffer or adjust the gamma in photoshop After Effects or so.

 

so in resume, set gamma 2.2 linear color mapping, click "don't affect colors" check SRGB button in your VRay buffer, happy with your render?, save image as exr 32 or 16 gamma 1.0 do your magic in Photoshop, after effects,combustion, Shake, Fusion, Nuke or what ever heck you use.

 

No.

 

I understand the LWF and know when double gamma has been applied. I'm sorry I used a small image as an example, but if you notice carefully...the gamma, or "brightness" is actually the same, the only thing that changes are the blueish pixels that have a higher intensity when the gamma is applied on the VFB (don't affect colors checked).

 

To summarize, the overall look of the rendered images is identical, except that stray pixels appear when "don't affect colors" is checked and gamma is applied in the VFB.

[Umesh Raut]

your points will be understood better if the image is visible to naked eyes

and vray's official help (spot3D) will give you more info on that issue.

[christofferthulin]

As far as I know V-ray calculates differently depending on if you burn the gamma correction in the render or not. Correct me if I'm wrong but when you have the don't affect colors ticked v-ray will find different areas of the image that needs more calculating than if it's not ticked. Do I make any sense?

[andrewrpreza]

As far as I know V-ray calculates differently depending on if you burn the gamma correction in the render or not. Correct me if I'm wrong but when you have the don't affect colors ticked v-ray will find different areas of the image that needs more calculating than if it's not ticked. Do I make any sense?

 

I think render speed does differ, but I read somewhere that this is some sort of bug or something like that.....other than that i'm clueless and have never read anything related to the issues i'm having.

 

as someone else said, I think i'll ask in the spot3d forums..

 

 

thanks for your replies though.

[christofferthulin]

Yea, read a little bit about it and found out I just hade some things mixed up.

[Francisco Penaloza]

what's the problem with posting a bigger image?

I still stand with my answer, read the Chaos group instructions regarding this and you'll have the same answer is not a bug it is the way the system work.

Now if you can see something different that we can't. well post a bigger image so we all can see it.

[andrewrpreza]

what's the problem with posting a bigger image?

I still stand with my answer, read the Chaos group instructions regarding this and you'll have the same answer is not a bug it is the way the system work.

Now if you can see something different that we can't. well post a bigger image so we all can see it.

 

 

The image was actually larger, I just cropped to the area of interest. Click the attachment and let the background gray out, If you still do not see the difference then you probably won't be able to help me regardless of image size anyways.

[Scott Schroeder]

You are seeing the differences in how Vray itself samples your image which is all done under the hood. The don't affect colors check being on essentially allows Vray to see a much wider range of noise which results in a cleaner looking render. So to answer your question, no, you should not see identical results and no it is not a bug. You should see close results, but not exact matches as your image clearly shows.

 

A better question is do you really need to be overly worried about a few pixels in a 100x100 square? The blur tool in Photoshop is your friend sometimes.

[andrewrpreza]

The don't affect colors check being on essentially allows Vray to see a much wider range of noise which results in a cleaner looking render.

 

In my case, the don't affect colors being on is what gives me a less clean render...? In my original example, baking gamma is what gives me an overall cleaner look...

[Scott Schroeder]

Are you using any sort of clamping or sub-pixel mapping? Since vray is working in a linear space you could be going above the normal ranges and getting the sparklies in your scene with dont affect on. Without it, you are working in a quasi clamped environment to begin with.

[andrewrpreza]

Are you using any sort of clamping or sub-pixel mapping? Since vray is working in a linear space you could be going above the normal ranges and getting the sparklies in your scene with dont affect on. Without it, you are working in a quasi clamped environment to begin with.

 

 

In both renders, sub pixel mapping is enabled and clamp is set to 1.