Conclusion and Final Rendered Piece

Over the course of the last 10 weeks I have learned so many skills. I initially evaluated my ideas using brainstorming and testing. I re-evaluated my ideas listening to advice given to me on the formative assessment. I then came up with a very interesting idea based on one suggested by my tutor (that the inside of the box was of a gallery).

After deciding to have a transforming art gallery, my eyes were set on creating an effect that I initially had no idea how to do. Taking a risk, I learned and developed some existing skills in MEL/Python scripting to create the logical pseudo code that would develop the logical grounds on which to base my script on.

Through trail and error and independent work, I manage to solve the the problem to create the wave effect only to be face with a new problem- randomizing the effect mildly so to create an edge of realism. Having solved this I had to begin the next part of my project- i.e modeling and texturing the room. Before this project I had no real idea on how this should be done. However, using my fellow peers as a great source for learning, i found that I was able to learn the basics of texturing simple objects, becoming more confident on Photoshop, Illustrator and the Hypershade within Maya.

Finally, with some experience from my environments project, I began Lighting my room and now completely confident in using Global Illumination in conjunction with Final Gather to create some of the best lighting I have done so far.

Finally I optimized my render and added the final colour touches within After Effects.

My overall impression of this unit has been very positive and I am glad to have developed a wide range of knowlege with many softwares which revolve around Maya as well as deepening my understanding and confidence in Maya.

With out any further delay, I present to you the final piece of ten weeks work.

Thanks to all that assisted me :-)

My Printed Model

Part of this project was to use some of the latest facilities in our college: i.e the 3D printers.

Since my project did not have so much application in modeled geometry I decided to print out the letter of one letter of my Sky Arts Logo.

Unfortunately the letters in my animation comprises grouped nurbs. In order to print out a successful model, I would need to convert my nurbs into a polygon.

However, I did run in to some trouble- after modeling it in the appropriate way then eventually opened it in netfabb I ran into a very strange error.

After troubleshooting all of the possible answers indicated that something was not quite right with the model.

I guessed that the only difference between my model and other peoples models was that mine had been converted from a polygon and that i may not have payed attention to the face normals.

Surely enough, when I analyzed the model in Maya again I found that the model and face normals varying in direction.

Once I fixed the root of the problem, I re-imported the necessary file to be printed with no problems at all.

The next day my model was successfully. Here is a picture of my printed model...and me!

The Last Stage: Compositing

After rendering out all of my scenes, I needed to bring all of my rendered tiff files and composite them as a sequence with the appropriate music.

But one more thing was required. From my lighting in Maya, I achieved an 80-85% effect that I wanted. Now I wanted to finish the rest off in After Effects. As you can recall, I wanted the room to seem desolate at the beginning of the animation, dark, and colourless. By the end of the animation I wanted it seem bright and very well lit. Instead of trying to do this directly in Maya- where the exact lighting can be very difficult to achieve, I could get an equivalent result in a much quicker time in a compositing package.

I found felt that the lighting for for the ending was very close to what I wanted, where as the beginning seemed far too bright- the animation needed more contrast. I did this by creating and keyframing a colour grading effect which gradually allowed the colours to saturate over time, giving the overall contrast to the animation.

Here is a screen shot to demonstrate the difference between the initial frame with a relatively simple colour grading effect that would be very difficult to achieve an Maya alone.


The bottom right shows the colour from the original maya render. The top left shows the colour graded version which is desaturated with colour corrections in After Effects.

After finishing the final touches to the composition, the final video was finally placed with the music and then ready to be rendered as a final piece.

Rendering: An Optimization Process

For my animation I was required to render out around 250 frames. Each frame was varying in render time from 5-10 minutes/frame.

Having two computers to render from, I had to optimize my processing power to get my renders back as soon as possible.

I developed a nice way of going about this.

Imagine you have 2 computers to render from and you don't know which one renders the fastest.

If you gave the first half to computer A, and the second half to computer B, this would mean that you are limited by the speed of your slowest computer (but you don't know which one that is). Ideally you would want to give the faster computer more frames to do and the slower computer less frames such that both computers render and FINISH at the same time- this would mean that you have optimized time with respect to your processing power.

The technique I made involved giving 50% to computer A, and 50% to computer B.

If Computer A was the faster renderer, it would finish it parts first.

Look at computer B and give the second half of the REMAINING frames back to computer A.

Assuming that A is still faster, it will render its new set of frames quicker than computer B. 

If computer A finishes much earlier than computer B, you can again, take the second half of computer B's frames and render them on computer A while computer B renders the first half.

You can repeat this process until all of your individual tiffs/targa files are made.

The really clever thing about this process is that computer B represents "the slower processor" at any given time. Imagine half way through, computer A became slower. All you would need to do is give the second half of the slowest computer, to the faster computer, and repeat this until you finished your render.

Using this technique I was really pleased to find that I got both of my computers to render constantly over time, and eventually got both computers rendering their last frame simultaneously- thus optimizing the render to within minutes of each other!

Having done this, my scene had fully rendered in efficient time, now ready for the final stage- compositing.

Animation Tests, Finalizing the video

Up until this point I was lighting the scene in such a way that fit the ending. However, I now faced another challenge. I have never lit a scene in which the lighting changes within the scene- i.e animated. I had to be very careful that I would animate the lighting of the room correctly so that it was suited for all the frames.

The way I intended on doing this was to turn off global illumination and test whether the direct lighting illuminated correctly.

One great skill that I learned to use over this project was to maximize efficiency in renders. When dealing with complex computational calculations inherent in Maya such as global illumination and final gather, the render of each frame increases by a huge factor. If one was to keep all of their render settings high- especially whilst lighting and rendering, it may take a very long time. In industry, however, deadlines have to be met and there is more often than not to get results as fast as one can.


In the following clip I felt that had matured somewhat in terms of my pipelining. Having taking renders of the start and end frames, I now needed to interpret the animated lights and match them up to the animation of the scene as a whole.

I achieved this by reducing the quality of my render (min sampling to -1) and only rendering the direct lighting. This was an excellent way to isolate the specific animation and nearly conclude the animation before taking it into compositing.






From here it was very simple for me to analyze my scene and then make the final adjustments.

Lighting, Texturing and Rendering Tests 06

Here is the procedural steps which I made whilst adjusting global illumination settings.

The following was an example of using max distance to even out the photon emission.

Increasing the photon radius size

A visual representation of the global illumination photons in my scene:

Gradually increasing the radius size of each photon

With advice from certain peers, notice that I incorporated a shadow in the last frame. This was suggested by Ollie Kane. I was very appreciative of all of the feedback and support I got from team members in my other project. I felt I developed the quality of taking feedback in a positive way to improve my work.

lighting, Texturing and Rendering Tests 05

Now I have the basic layout and objects modeled, and UV'd, it was a perfect time to start re-assessing the lighting situation.

Here are some pictures which indicate that state of my scene as it progressed.

One thing which I have learned over the course of this project (and my environments project that has been running along side this one), is that when you set up lighting, you should always have your direct light done, before you play with indirect global illumination and final gather. This is my initial setting up of the direct lighting. The main light sources come from the overhead light fittings.

Lighting, Texturing and Rendering Tests 04

Here are some screenshots which show the number of layers required to create a texture for Maya.

For my paintings, I wanted them to have both a colour and a bump map associated to them. I.e, I often created an unsaturated canvas texture to act as the bump. Coupled with layers to add depth and a papery feel touch, they were then used along side blending nodes in photoshop to give the desired look.