Announcement

Collapse
No announcement yet.

Basic Comping of a Vray render

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Basic Comping of a Vray render

    OK... it seems a lot of you know that you can rebuild the renders that are split out of the vray g-buffer, but may not be clear on how to do it. Vlado did a very good breakdown on what the elements are here...

    http://www.spot3d.com/vray/help/VRay...buffer.htm#ex2

    But this may help you.

    The following is a very basic breakdown of how to reassemble a vray image in a compositing program. I am showing it to you in a very straightforward node breakdown, but this can be done in AE or any other package. First thing... you need to be in Linear Color Space the whole time since many of these elements such as Raw GI will be linear.

    Lets start by showing the different g-buffers that are needed:

    The core is this: the Diffuse. All it is a solid color of what comes out of diffuse, no shading. If you have a texture, it is a texture, a color is a color.



    The light is then broken down into direct and indirect lighting. Use the RAW versions.



    To get the effected diffuse of these, you multiply them by the diffuse

    SO if you take the RawGI and multiply it by the diffuse, and you take the RawLight and do the same, you get the Diffuse direct light and the Global Illumination light.

    They look like this:


    So now you have the two basic elements of light: the direct light and the GI light (in terms of diffuse light). You can even think of it as key and fill. Light is additive, so all you need to do combine them is the add them together in the comp.

    But there are still some pieces missing. And that is your specular light. While Vray really considers specular as a reflection, the cool thing is that it will separate out the reflection of a light as specular and reflection as reflection. Think of it in the same way as the direct light and the GI light. Those two passes look like this:



    These two passes can be added to as well. By the way, if you have refraction or self-illumination or things that get split out, you will have to add them too.

    So this is how it looks in a diagram:



    I am showing some color correct nodes in there since that is what you want to do with this.

    You combine all of this: (Diffuse * GI) + (Diffuse * Light) + Specular + Reflection, and you get the same image that you rendered, which is this:



    OK great, so you got the same thing you started with, what good is that? Well if you separate things, you have a lot of power to adjust it in post. First thing, I saved everything as full floating point unclamped color. So I already have the power of adjusting the exposure and the gamma of the image. Here are some examples of that:



    Ok, but let’s say you want to desaturate the GI and maybe increase the contrast on it a bit. You could do that in Vray GI dialogue and redo the render or you could do a color correction on the Raw GI pass, and you get this:



    Want to decrease the intensity of the sunlight? No problem, just adjust the RawLight pass:



    Want to increase the reflection and maybe darken everything else a bit? No problem. Simply, darken the diffuse pass, and lighten the reflection pass:



    Anyway, I hope this helps you guys.

  • #2
    Thanks for sharing this info, cpnichols!

    But, what about changing something in one particular object, let´s say, lowering the reflection of the floor without affecting the rest of the scene?
    I´ve found that splitting the scene this way, i lose the ability of retrieving the info in gbuffers, DOF, motion blur, object id and material selections, because vray can´t embed all this gbuffer information in the layers u are talking about. So for me, it´s a matter of choosing to have one type of flexibility or the other, not both at the same time...
    My Youtube VFX Channel - http://www.youtube.com/panthon
    Sonata in motion - My first VFX short film made with VRAY. http://vimeo.com/1645673
    Sunset Day - My upcoming VFX short: http://www.vimeo.com/2578420

    Comment


    • #3
      Well that is done with simple comping. Take the Material ID or Object ID and make it a matte to only effect part of the layer. The point of what I was showing was more about showing how to comp the lighting passes, not about comping in general.

      Comment


      • #4
        That´s ok, I think everybody will find this tutorial very useful!
        Keep up the good work cpnichols!
        My Youtube VFX Channel - http://www.youtube.com/panthon
        Sonata in motion - My first VFX short film made with VRAY. http://vimeo.com/1645673
        Sunset Day - My upcoming VFX short: http://www.vimeo.com/2578420

        Comment


        • #5
          Great tutorial,

          interesting to see the possibilities on a real project.

          Thanks for sharing this information
          You can contact StudioGijs for 3D visualization and 3D modeling related services and on-site training.

          Comment


          • #6
            I saved everything as full floating point unclamped color.
            eh? you did what? is that a particular file format?

            cheers for the run down tho, very helpful. hey, anyone got some spare time i can buy?

            Comment


            • #7
              I guess he means that he has rendered with the clamp output in vray unticked. It´s located under gbuffer/colormapping section. And that he has used 32 bits per channel in the file format he has used, openexr, i suppose.
              My Youtube VFX Channel - http://www.youtube.com/panthon
              Sonata in motion - My first VFX short film made with VRAY. http://vimeo.com/1645673
              Sunset Day - My upcoming VFX short: http://www.vimeo.com/2578420

              Comment


              • #8
                what file formats can AE accept when it comes to floating point?

                Now this method allows for flexibility, i guess since you have to do a full GI render to get these passes that your not saving on render time at all.

                ---------------------------------------------------
                MSN addresses are not for newbies or warez users to contact the pros and bug them with
                stupid questions the forum can answer.

                Comment


                • #9
                  Render time is the same, but it allows changes that would lead u to rerender the whole thing again, which, of course would increase the total rendertime
                  As far as i know, AE accepts rpf but not in 32 bit mode... and no openexr I guess...For this floating point workflow u should give combustion a try, or if u are into the hardcore side of compositing, Digital fusion, NUke or Shake...
                  If u are using photoshop cs2, it will accept your openexr renders!
                  My Youtube VFX Channel - http://www.youtube.com/panthon
                  Sonata in motion - My first VFX short film made with VRAY. http://vimeo.com/1645673
                  Sunset Day - My upcoming VFX short: http://www.vimeo.com/2578420

                  Comment


                  • #10
                    AE will accepts OpenEXR now with a free plugin from The Orphanage.

                    http://www.theorphanage.com/tech/OrphEXR/

                    And yes, I used EXR. But you may be able to use HDR as well.

                    Comment


                    • #11
                      Originally posted by Da_elf
                      what file formats can AE accept when it comes to floating point?

                      Now this method allows for flexibility, i guess since you have to do a full GI render to get these passes that your not saving on render time at all.
                      Not sure what you mean by this... It is the same as doing the regular render but it saves out all the extra passes of the G-Buffer.

                      Comment


                      • #12
                        Originally posted by cpnichols

                        And yes, I used EXR. But you may be able to use HDR as well.
                        Is there an advantage to one or the other. Or is it just another way of storing the same data?

                        Thanks.
                        www.dpict3d.com - "That's a very nice rendering, Dave. I think you've improved a great deal." - HAL9000... At least I have one fan.

                        Comment


                        • #13
                          Originally posted by dlparisi
                          Originally posted by cpnichols

                          And yes, I used EXR. But you may be able to use HDR as well.
                          Is there an advantage to one or the other. Or is it just another way of storing the same data?

                          Thanks.
                          In this case no, because all I really wanted was the full float data. But in production yes. Open EXR is amazingly efficient at saving full floating point data in a lossless format using very high compression. HDR files are huge. This is a huge deal when you are rendering things like 20+ passes for 800+ vfx shots on a movie with jets for example, and you only have about 50 terabytes of diskspace. EXR is also open source so you can alter it to fit what you are doing. It can save layers, yada yada. Basically it is the format of the future.

                          Comment


                          • #14
                            [quote="nvanherpt"]
                            hey, anyone got some spare time i can buy?
                            what did you have in mind mate?

                            chris.....nice thread mate

                            mods....shouldn't this thread be made a sticky?

                            Comment


                            • #15
                              what i meant about saving time was that if you rendered a simple diffuse pass (which could easily be done in scanline as well) it doesnt take long per frame. then if you render a direct light pass without GI then that renders fast then fill light passes etc. those dont take as long to render each. however since this is GI everything needs to be rendered in 1 render which is seperated into your different passes for you to composite. as mentioned earlier it saves on having to rerender just for the sake of light tweeks. So saving in passes is not a method to speed up render times, its benifits are in its flexibility when comping.

                              ---------------------------------------------------
                              MSN addresses are not for newbies or warez users to contact the pros and bug them with
                              stupid questions the forum can answer.

                              Comment

                              Working...
                              X