Artefact Two - Lighting (Part 3)

A study of light

A camera was set up in a corner of the room to view how light changes throughout the day. Pictures were taken every half an hour for a period of 5 1/2 hours until it got too dark. Depending where the source of the light coming from the outside depends on how much illumination the room gets. Towards the end the light is still quite bright outside however the room is dark.

With the scene that was created before the mr spot light that is outside can be altered with the multiplier to change the strength of light that peers into the room.

Changing the multiplier at different intervals increasing, the scene gradually illuminated more. Below are the resulted renders.

Below is the final fullsize render. The multiplier light setting was at 8.0 and the final gather preset was High (500) rays and a diffuse bounce of 4.

Rendering time 35.31 Min

One final poll was created to see if the final rendered image with the lighting has given the scene real world realism that i have been seeking. Eighteen people were asked and the results were very positive all the people voted that the lighting has made the scene more realistic. The main comments were the reflection on the floors surface (left of the picture), the dark areas at the bottom/underneath the couch and the rug turned out very well.
So overall a success! Below is the results.

Evaluation

After the scene was created the textures that were assigned to the objects were looking flat and this made the scene uninteresting. The bump map and specular map gave these objects the realistic depth they needed as if they were in the real world. These changes can be very subtle especially when a light is optimising these irregular details such as the stone look on the plant pot. Whilst the combination of specular and reflective maps gave the objects in the scene gave the feel of light, the best example is the edges of the plant pot where the light is emphasising the boundaries of the objects. After getting the scene to look more realistic the file sizing part of this artefact was to optimise the dimension of the textures and compromising between the physical size of the texture and the eventual output render of the texture by changing its dimensions. As textures have to be a power of 2 for graphics cards to cope easily - the process began at 1024 and worked down to 128. From the results table the best dimensions were obviously the higher 1024x512 and this file size was large especially if multiple textures were to be involved and the worst were 256x128. The later saw bad results in terms of a realistic texture to an object, the wood looked bare and too specular for the scene. The best result that had the most excellent compromise were the dimensions of 512x256. This is justified by the small physical file size 383KB (from 2.1MB) and a respectable amount of detail lost in the final render that did not compromise the scene.

Conclusion

Another in depth experiment that could have been produced to make this artefact would have been to test the different dimensions of mapping textures such as the bump and specular maps that are applied to the texture and notice the difference in realism and file sizes. If optimizing these file sizes and compromising the final render, the scene can be created for very little physical memory with excellent results.