3D composing over an image


 

 

If you like this tutorial, please write a comment or a suggestion in the "Guest Book" section. Thanks 

This tutorial is © COPYRIGHT by José María Andrés Martín (ALZHEM)
 The total or partial reproduction of this tutorial by any means is prohibited to third parties without prior written authorization from José María Andrés Martín (ALZHEM)


3D composing over an image


Approximated duration: 1-2 hours
Software: Softimage XSI 3.5, 3D Studio MAX 5.1 (VRay render engine) y Adobe Photoshop CS
 

How many times we have looked at a picture and don't know which parts are 3D objects and which real. This magic is mainly the merit of a good integration of the 3D object in the scene, but... how could we get this?

We must follow this basic steps:

  1. Modify our 3D model till we get the good one.
  2. Set up the camera with the better angle and perspective, as near as possible of the background.
  3. Set up the "matte" elements if they were needed.
  4. Analyze the illumination of the background where our character or 3D object is going to be composed.
  5. Select the render engine, set it up and render the final image.
  6. Make the needed retouches  in Photoshop.

In this composition of our character (Papo) on a frame of the film "Shrek" by PDI/DreamWorks, we use SoftImage XSI 3.5 for the modeling, texturing and set up of it, 3D Studio MAX 5.1 with VRay for the illumination and rendering, and Photoshop CS for the cleaning of the background and final retouches.

First steps

The first: to think what we want to do. In this case the main idea is replace "Shrek" of the original frame with "Papo". This is a not easy step, because "Papo" is too much thinner than "Shrek". So... select the clone brush and let's go!


After 15 - 20 minutes we could get something like this. Now we must look for a pose similar to Shrek's original one.


Looking for the best pose

We must modify our setup looking for just the same (or similar) pose of Shrek. Is good to look at the original picture one and more time.


If we put the original picture as the background, it could help us so much.


Once got the pose, let's export the mesh to MAX. That way we can use the VRay.

Setting the camera up

After merged the mesh in MAX and created all the materials, we need to set up the camera: position, angle and perspective. This is one os the most important steps of the project because here is the 50% of the final realism, or higher in an infoarchitectural project.
The best way to do this in MAX is move the perspective instead of the camera, because with perspective we have a lot of freedom, and after this, convert it to a camera (selecting  the viewport, and with the camera selected too, pressing Ctrl+C)
We set up as background the original frame  ("Shrek" and "Donkey") and orbiting, zooming and trucking the camera, we will get the best perspective with just the angle we wanted.


Lighting

The hardest thing to do in this composition is the lighting simulation because of the leaves the light pass through.
We must see how the light affect the objects (specially to "Shrek" and "Donkey") and set up the light.
After setting up the light, we must look for an image which "filtered" the light (black and white image of a tree's leaves), and create a plane, as a floor, with a "Matte/Shadow" material. The global illumination will do the rest.


Render Engine

Once set the camera and light up, we only need to choose the render engine. Here we use VRay due to its great render time and its quality.
After the rendering process, we save the image in TGA of TIFF, storing the alpha channel. In Photoshop, the alpha channel will help us to select Papo.

Final retouches in Photoshop

The best advise for a composition is to assign the original image (with "Shrek" erased) as the backgroung, and the Papo one as a layer. This way, we can adjust the colors and every value of each element separately..


The retouches made were a "Hue/Saturation" adjust to desaturate Papo a little, and modify its green color to a yellower one, a "Curves" adjustment layer to get a more contrasted colours, and another layer to add some grass over his feet.

                         ...


After many months, I've seen this composition again and... Oh! What is this!?!? The picture was not too marvelous as I remembered. What happened? Basicaly I discovered an error in the background. This is different in the right and in the left of Papo. 15 minutes later the background is right, and  with a more professional look.

Click on the image to see the final release

 

If you like this tutorial, please write a comment or a suggestion in the "Guest Book" section. Thanks