Home > Training > Some basic postproduction and texturing concepts
Etérea Workshops Banner

Algunos conceptos basicos de texturizado y postproduccion
Please: don’t make a copy-paste of this material to your page, weblog or forum without my permission. You could include a link here, better.

WARNING: this it is a small tutorial where I will explain the basic techniques of post-production and texturing that I have come to use in my work over the years.

The workflow provided here is not of a quick fix nature, but one that demands —on the part of the artist— an attitude of hard work and dedication. But what I can assure you is that the effort you put into using these techniques is more then compensated for by what you will get out of using them (quality of the final image and render times).

I have invented practically nothing that I explain here, these being the very basic techniques used by the best CGI artists, especially when it comes to the texturing concepts and multi-layer rendering.

Finally: If you are looking for a way to obtain a pretty result with the push of a button or a tutorial that uses predefined materials, then this tutorial is perhaps not suited for you ;-)

Cristóbal Vila, October 2006

English translation by Ian Waters, Lead Illustrator WAMM, from Manchester, United Kingdom [ Thanks Ian ]

On the right we have a model of the popular and entertaining game known as “Peg Solitaire” that consists of a wooden base with 33 holes and 32 small stone balls, arranged in such a way that the central hole is free. This it is a very simple model that I made for an old work —“Bereber Still Life”— and we are going to use it as a test range to explain some concepts, techniques and tricks of texturing and postproduction.

I always like to say that the best way to tackle a big problem is to turn it into several smaller problems, this way it's much easier to approach them and find solutions. What we have on the right is neither a great problem nor is it a complicated project, on the contrary, is a very simple piece. I have chosen it because I believe that some basic and important concepts can be understood more completely from a simple project, and that from this starting point, you will soon be able to apply these concepts to more complicated projects.

The first detail that wanted to show is that the wood piece that makes up the board has been divided —modeled— in two parts, thus we have a total of 3 objects:
Modelo basico solitario
Parte central Anillo exterior Bolitas
Often we all go mad trying to create the finished material of an object with only one texture —resorting sometimes to careful methods of UV projection— when the simplest solution is to divide our object up or to model it from scratch in several parts (although it would still like it was made up of a single large piece, as in this case). Now we can texture each part with its own projection according to what we think fits best.
Texturizado planar Texturizado esférico
This example shows planar projection being used to texture the center body of our disc. And here we use a type of spherical projection to apply the texture to the curved edge.
Modelo con las texturas “limpias”  By doing it this way we will be able to avoid the vertical projection of veins in the wood on the curved edge of the ring. Anyway, experiment, because sometimes that type of projection can indeed be more convincing, depending on how our texture behaves volumetrically (we have to analyse each particular situation and to decide which method we like more).

If we chose to texture the disc like it was a section of tree trunk, meaning that in the cut we could see tree rings and again in the outside the crust, this would still be the appropriate way to go about it (though applying a cylindrical projection for the crust). Though here we have used spherical projection in the outer piece because it seemed to adapt better to the double curvature of the edge.

Here is a pretty cleanly texturised wooden disc.

Perhaps too clean? Does it look to you like it is just left the box?

Reality is imperfect

In the real world the things usually have slight flaws, subtle changes in different superficial characteristics (ruggedness, reliefs, changes in brightness and even in saturation, etc). We are going to see how we can give more “life” to this piece by working in the different material channels. As a rule, almost all objects show small disturbances in their specular, reflection, bump… so it is our task to introduce those subtle changes to each of those channels.

Let's see the different textures that I have used to create the material of this model. I'll only show reduced versions of those images. I like to use textures with a sufficient resolution so as not to compromise final render. The original images were 2048 píxels each way, but here they are shown at 400 píxels for layout reasons.
Textura original Máscara agujeros
1. This is the original texture for the wood used in the disc. We do not have to worry ourselves about it being “seamless” (repetitive) because it is in fact going to to cover the whole of the disc. You can find good textures on the internet, many of them free (,, etc).   2. Having the model as a base I created a mask for the spaces occupied by the holes in the wood. This will be useful to obtain the surface finish as wood texture changes where it is milled: in the zone of the holes there is no brightness and the area is rougher.
Textura con agujeros Bump
3. Using the mask above we can make the saturation and the tone in the zones where the wood is milled change. At the same time I have made the bottom of each hole a little darker, using simple feathered circle.   4. Using the red channel from the original texture as a starting point, we obtain what will serve as our bump (relief) map. It is necessary to observe that not all the surface has variations of grays —we don't want a totally rough wood—, for that reason we see that there are pure white zones: in these the wood will be smooth.
Mapa de suciedad Mapa de suciedad
5. This it is a dirt map from the “Surface Toolkit” by DVGarage. A fabulous collection: I believe that they are the best 99 dollars I have ever spent ;-)

The intention is to use it to simulate certain stresses caused by the board rubbing with other surfaces.
  6. And another dirt map, this time belonging to the “Water Damage” collection, also by DVGarage. With this map we are going to simulate the effect of some liquid that has been spilt on the board (that cup of spilled coffee, or that Coca-Cola tin that spreyed everytwhere when opened…)
Mapa de suciedad Mapa de especularidad y reflexión
7. This third texture is from the same collection as number 5. All these maps can be used in combination or separately to influence the specular behavior or variations in reflectivity of the surface.   8. Here we see a combination of maps 2, 5 and 7 to obtain an image that we will use to control the specular and reflectiviy levels of the disc. Logically, where the black circles relate to on the wood the map will be no effect and in the other areas there will be subtle changes.
Next we see how the behaviour of the surface changes as we are apply the different maps to it:
Superficie virgen Especular + Reflexión
Bump Bump + Especular + Reflexión
When I am adjusting a certain aspect of the image I like to isolate it: if I want to refine the bump I prefer not to be confused by the other aspects of the image (color, brightness, reflection, etc). For that reason often I turn off the rest of these characteristics and concentrate exclusively on the relief (bump) of the surface. I do the same with the levels of reflectivity, transparency… and generally with many other aspects of the project (lights, composition of the elements, motion…). What I said at the outset: to isolate, recognise and solve small problems individually rather then be confused by the complexity of the whole set.

Next we see the difference between “the virgin” wooden disc and the one that has been worked on. We could touch up and exaggerate some details a tad more: it could have deeper paint spots, cracks in the wood, etc. I'll leave this to you though (what it matters is that we stick with the concept ;-)
Superficie limpia Superficie final

The small balls: procedural texturing

When texturing the wooden disc we used a set of bitmap textures, but to obtain the surface finish of the small balls we are going to use some shaders. And we will be using what is by far the most useful: FractalNoise. I suppose that in almost any 3D software you will find a similar kind of procedural texture. The key of its power is in using it in layers; in each one we introduce a certain level of detail that will help enrich the whole set.

So let's see how we can obtain a relatively complex final texture exclusivly using this simple shader (in my old illustration “Escarabajo Joya” you can find a sample of this: the branch in which the insect leans is textured with a combination of many FractalNoise layers . It is also very useful to combine it with other textures — bitmaps or procedurals— to obtain very rich results, as in the grapes of “Canestra di Frutta”).
Primera capa de shadingIn our texture undercoat we use two colors: black and dark gray. As the application of this shader is volumetric and its nature fractal, the result obtained will not yeald any kind of repetitive pattern: this is the great advantage of using shaders like FractalNoise.
Segunda capa de shadingIn the seccond layer we apply an oxide colour. The second color is made transparent, this is so that we will continue to see what we had in our previous layer. It is also interesting also to play with the different opacity levels of each layer, as well as the layer modes, just like Photoshop (multiply, divide, clarify, etc).
Tercera capa de shadingIn the third layer we'll enhance the chromatic range of the objects with a “not so predictable”: blue hue. It is strange to see how many surfaces look superior when we introduce a colour that essentially doesn't seem like it should be there. The impressionist painters knew this very well ;-)
Cuarta capa de shadingAnd to finish add depth to the material by adding a colour that is much clearer then the others. What we have done to these simple small balls is to paint by layers, as did the 15th and 16th century Early Netherlandish painters: instead of looking for a final effect with just one colour, we are applying different shades in several layers, with subtley varying transparency so that the result acquires greater colour depth.

Improving the result: postproduction

Once we have all the texture elements and have added another pretty texture for the ground, we can render the image. The result of which, we have on the right. Evidently the look of the final image will also depend on other factors, like the lighting setup or the render engine. The final render is good, but let's take a look at the weak points and see what we can do to improve it substantially:

— There are no contact shadows under the wooden base.

— The small balls don't seem to be resting on anything, again this is due to the lack of a contact shadow.

— The colour range of the image is poor viz. It's a little dull.

— There is no feeling of depth.

— We know we're looking at a 3D image, it lacks credibility, this in spite of the work we did with texturing. 
  Render virgen
In order to solve the contact shadow problem we can use global illumination (or we could also use ambient oclussion). I like to render this seperatly, by doing it this way I can control the effect much more precisely.

Next we can place this over our render and play with the layer modes in Photoshop (or AfterEffects, Combustion, Shake… if it's an animation). If we use “Multiply” the dark parts of our GI layer darken the render.

Using “Overlay” will cause the dark parts to darken the render and the light parts brighten it. This is what I like to use, because we can “kill two birds with one stone” (remembering that in addition we can also control the look of this GI pass with levels or curves adjustment layers).
  Iluminación Global
In order to be able to create the depth of field effect and to de-focus the image based on distance we can use a depth map (Z-map). These can be created automatically in many programs, but to my I like to create them using depth fog, as this gives more control. Here is the method, in small steps:

1. Duplicate the scene (save as), give it a new name.
2. Apply over every object a plain white material.
3. Increase the general luminance up to 100%.
4. Create a black fog and adjust the In and Out radii.

Another one of the advantages this has over Z-maps is that in many cases Z-maps are not anti-aliased. In the “black fog” method they are ;-)
  Mapa de profundidad
Something I don't like about 3D renders is that the edges of objects are nearly always excessivly sharp.

When we see a photo of object, although perfectly focused, when we look closely at the object edges we see that they appear —to some extent— smoothed. Some render engines create very good and smooth antialiasing, but others excessively define objects making them appear hard.

In order to try to correct this problem we can use an edge map. We apply it in such a way that it serves as a mask to introduce a slight blurring in the silhouettes of the objects.
  Mapa de contornos
In order to introduce the defocusing characteristic that is depth of field we can use the Lens Blur filter that comes with Photoshop CS. This kind of defocusing is far more realistic — “more photographic”— then the kind that we can obtain with a simple guassian blur, it allows to controll the extent of blurring based on a depth map that you place in the document as a channel or mask. There are 3rd party plugins with which you can get a much more realistic result (Lenscare, DOF PRO,…) but for most situations we will be able to get out of trouble with Photoshop's built in filter:
Lens Blur Lens Blur
Next we will examine how to organise different layers in Photoshop and see how to postproduce in a non destructive way. This way of working is very effective because it allows us to activate and deactivate our adjustments, correcting and continuing with the postproduction in a more controlled manner.
layers layers
Take a look at the difference between the original render and the image obtained after postproduction.
Render virgen, sin postproducción Imagen final, con postproducción
Imagen final
Some of these refinements could have been introduced directly into render: we can render with the depth of field effect already applied; we can create a render with global illumination built in; but doing it this way has major advantages:

- Time: Depth of field when calculated in the original render usually takes much more time (depending on the render engine and the system used for the rendering) then applying depth of field in postproduction; and the same goes for global illumination as it also increases the render time (especially if we have objects that are transparent or reflective).

- Control: when having everything in layers we can control the exact outcome of the image to great effect and more quickly, this allows us to try out alternative settings in a very fast and agile way: we varied the opacity of the GI pass, we controlled and adjusted the levels settings, we can use a depth map to blur everything except the GI, we can vary the center of attention so that the point of interest in the scene changes, etc, etc.

Cristóbal Vila, Zaragoza - Spain, October 2006.

I hope that some of these techniques can serve to help you in your renders. But you must think of the tricks exposed here as simply the departure point and continue investigating, working and putting your imagination into practice ;-)

NOTE: I have tried to be as precise and exact as possible, but it is probable that some errors exist in the text . If you find an error please let me know and I'll correct it.

Etérea Workshops Banner
Please: don’t make a copy-paste of this material to your page, weblog or forum without my permission. You could include a link here, better.
All images copyright Cristóbal Vila

you are the visitor | eres el visitante