This tutorial will demonstrate the difference between bump maps, normal maps and displacement maps.Color maps
Color maps allow us to represent some of the diffuse detail that is present in real life materials, however it lacks any light information. The two main qualities reflected in materials are shadowing and reflectivity, so let's begin by looking at how we can simulate shadows in materials.
Bump maps
Bump maps are often described as maps that allow you add intricate details easily to textures, as its name suggests - adding all those little 'bumps' to objects that would otherwise take hours of painstaking modelling.However, this isn't exactly true. Bump maps actually control the way light is reflected on objects. It simulates detail on objects by artificially redirecting light rays as they would if the geometric deformation existed.
It tweaks or perturbs the normal based on the information in a bump map - usually a greyscale map where lighter colors signify the bump intensity.
Normal maps
One failing of bump maps is the fact they simply alter the existing normal and so are limited in the size and depth of surface detail that can be represented by them. Normal maps (or dot3 bump mapping) is similar to bump mapping, however whereas bump maps alter the existing geometry normals, normal maps replace them completely.As such normal maps normally three color channels one reflecting each coordinate (xyz) for that texel.
In particular you can see that the image has significantly more realistic shadow information present and as such you'll see the most dramatic difference between bump and normal maps when you have moving light sources in your scenes. This is part of the reason why normal maps have so quickly become essential in modern computer games.
Displacement maps
Also know as height mapping, displacement maps literally displace the mesh itself based on texture information from a map - usually a greyscale map where white is the highest displacement and black is the lowest.
This allows for highly realistic texture maps as your objects are literally being displaced by the map. This means that any light source casting on teh object will cast true shadows and silhouettes.
In our example look how the shape of the sphere is actually deformed now by the planks which are displacing it.
The downside of displacement mapping is that as it is creating actual geometry at render time, it will have a significant impact on the render time. Whilst this used to make it prohibitive in use, increasing CPU power has meant displacement mapping is becoming a more and more everyday technique in professional rendering environments. The next generation of GPUs are also integrating displacement mapping, which will ultimately allow it to become common place in realtime environments like computer games.
Conclusion
Bump, normal and displacement maps all have their distinct advantages and disadvantages. The key is to be aware of these strengths and weaknesses and use them in complement. Often you'll see scenes where the objects closer to the camera are rendered with displacement maps and the background is using normal or bump mapping where less detail can be observed anyway. You can also use the maps together in complement - with a distinct displacement map being used for larger objects needing displaced and a bump/normal map for the smaller, more intricate detail. So for our example you can see the planks themselves have been displaced, whereas the wood grain is relying on bump/normal mapping for realism. This will then allow you to lower the subdivisions required for your displacement and speed up the render time.
