help with specific requirement for v-ray render element z-depth type mask or similar

[ctasker]

Hi

I'm trying to output a very specific z-depth mask element for my compositor.

I have previously used the standard camera clipping plane set just in front of and just behind the first and last pieces of geometry to generate a focused gradient over the field of geometry which has worked for smaller scale work.

 

The current job has a very large site area and I was asked if I can produce a similar mask with the following specification:

 

For camera to target range 0 to 1Km I need all geometry to output a black mask

 

For range of 1km to 15km I need a gradient transition from black to white (similar to what I have been doing in the past.

 

For range 15km to infinity I need a completely uniform white mask output.

 

Basically he has configured his photoshop workflow such that he wants to output 0 blurring/dimming for the first 1km, then a transition of blur/dimming effect in the 1km to 15km range and finally a uniform level of blurring/dimming from 15km onwards.

 

The process will be repeated for hundreds of views so for practical reasons I need to try to output this as one pass per camera viewpoint.

 

Any ideas if the z-depth can be modified or scripted to get the above results or if there is an alternative method that will produce the same mask?

 

Hope someone has tried to do something similar before.

Any suggestions welcomed

 

Thanks

C

[John Dollus]

sounds like you just need the 1km-15km zdepth which is pretty straight forward.

anything closer than 1km will already be 0,0,0 and anything further away than 15km will be 255,255,255. just make sure to not clamp the output and save to a floating point format

Maybe I'm not understanding your problem correctly? It's Monday and I've only had one cup of coffee :/

[Francisco Penaloza]

Low caffeine John is right, a simple Zdepth covering the whole area will do, plus in post you can clamp the areas that you need intensive masking, using level or gamma adjustment.

also you can output different Z passes, usually this is recommended with very large areas so you get more detail in small objects, later in post you can mix all this passes in to one to create a more detailed Z pass.

[ctasker]

Hi John

I had been using the camera clipping boundaries to control the start and end of the z-depth gradation as it was easy to visually set up the pair of boundary distances within the viewport without having to measure distances each time I set up another camera viewpoint.

 

If I understand you correctly then this is the solution i'm looking for if I simply un-tick the "zdepth use camera clip boundaries"

Set my new distance parameters beneath, 1km in zdepth min, 15km in zdepth max and either side of those boundaries will be processed exactly as I need them to be by default. I wasn't sure how vray handled the geometry either side of the min and max values, I thought it might disregard them as it does with my initial method.

 

Lastly I should un-tick "clamp zdepth"

 

I couldn't get my head around exactly what the clamp zdepth parameter does.

I read some posts about how it affects the way results appear in the frame buffer but I

didn't really know what implications that would have for the compositor's end of the workflow.

 

I'm not ready with the modelling side yet to create this output but will try these settings as soon as I'm ready to render.

Thanks for that.

Edited December 2, 2013 by ctasker

[ctasker]

Francisco

I'll run your comment by my post/compositor to see if it makes sense to him.

 

Really don't want to have to break it up into multiple passes if I don't have to.

There are so many panoramic viewpoints, if I have to render each one 3 times I will loose it.

Time frame is tight also so if what John is describing is the case then I think that is all I need.

Thanks for the suggestions