Children’s Society Project

In February we received a brief from the children’s society. This brief called upon us to make a short one minute animated film to accompany a clip of dialogue from children at the society.

The audio given to me was of a child speaking about social media, in particular its pros and cons. The clip I was given was in the context of cyber bullying. The effects of cyber bullying can be particularly damaging and so for me this was an important video to make.

This project was particularly challenging for a number of reasons, primarily because I had never created a 3D animated film before and my experience and knowledge of 3D was rather limited at the beginning.

Thankfully over the process of creating this film I have learned a great deal about animating, modelling and rigging with this software, purely through making many mistakes and being lucky enough to have tutors, Lynda.com and Google to guide me through and help me when mistakes or errors did occur.

I don’t think I was quite prepared for the amount of things that would go wrong during this project – primarily it was with rigging. I now have a great deal of respect for those working as riggers in the industry! The worst thing for me was that because it took so long to rig I could not animate until much later than I would have liked.

Here I want to share with you some of my mistakes and what I learned.

Blend shapes

The modelling process was probably the smoothest part of the process. We had modelled a few characters in class and there are plenty of modelling tutorials that are very helpful. The first pitiful I came across was applying blend shapes. We had covered this in class and it is generally a very simply process. For those that do not know, to create a blend shape you need to duplicate your original (base) model and from there you can manipulate the duplicated version into any pose you like. Once this is done you click on the duplicated model, shift select the base model go to deform > blend shape and in your channel box you will find the blend shape, and can essentially turn it off and on by using 0 and 1. You can also type in other values like 0.5, -3 etc.. to achieve a whole variety of different expressions from one blend shape. You can also create multiple blend shapes too.

I did make one mistake during the blend shape process that was quite costly to my time. I made a whole variety of different facial expressions: blink, smile, angry, open mouth, close mouth etc… totalling at about 10 different expressions, I then preformed the process that I outlined earlier and received a topology error. Meaning that maya could not perform the blend shape process. I realised that I had actually changed the topology of my duplicated models by inserting edge loops and multi-cutting, this meant that because the topology of the duplicated characters was different from the original base model it could not map the blend shape onto it. I had to restart the whole process of creating the blend shapes but now limiting myself to only manipulating the vertices in the duplicated model. This time it worked.

I did not know when I started that you could not insert edge loops into a duplicated model, and create a blend shape, this of course was a beginners mistake. The greatest mistake I made here however, was that I did not check as I went, I simply made all the blend shapes before testing one to see if it worked. As this mistake happened early on in the process for the remainder of the project I always checked things first before moving on to the next step. This can seem tedious but it saves a lot of time in the long run and is well worth doing

Rigging, binding skin and paint weights 

This part of the process took up roughly 70% of my time! It seemed as though at every step of this process something went wrong, whether it be IK handles set up incorrectly or skin weights causing all sorts of weird distortions on the model.
Once I had finally managed to successfully rig the character and bind the skeleton to the mesh I then had to start painting the skin weights. It is important to weight the skin accurately to make for a smoother time animating. Skin weights essentially determine how much influence a joint in the skeleton has on a part of the mesh. Certain parts of the mesh became particularly difficult to weight properly, because even though it appeared that the mesh was weighted correctly there were still some vertices that were being influenced by joints that shouldn’t have been. This problem was tricky to troubleshoot – one way I learnt was to go into ‘prune small weights’ this prevented joints that held a small amount of influence over a vertex to stop having an influence altogether.
If you want to have an accurate model to work with whilst animating it is important to take time on this step and to ensure everything is weighted correctly.

One note on skinning is that once your model is skinned you cannot really change the geometry of a model. This is why it is so important to make sure the set up for animating is correct; modelling, blend shapes, rigging, skinning, and applying paint weights are crucial parts of the whole animating process. It takes a lot of time but this time is well spent as it makes the animation process very speedy if done so correctly.

Animating

This was by far the most enjoyable part of the whole project – it was so great to be able to manipulate the character movements to bring the character alive. The best tools I found on maya that really made the animation process so smooth was the dope sheet. This was so helpful for getting timings right and was very easy and intuitive to use. It was while I was animating that I was really pleased I had spent time on getting the blend shapes right, it meant that facial acting could easily be animated by the flick of a switch. I knew I would have a lot of facial acting in this animation and that is why the blend shape stage of the process was so critical for me.

Editing and rendering 

Rendering in maya takes a long time! One frame takes roughly 17 seconds to render meaning a short 2 second animation (50 frames) will take about 14 minutes to render. I had 53 seconds to render! It also takes up a lot of storage so it’s definitely worth investing in an external drive! Some really useful things to do before rendering multiple frames is to firstly check in the rendering settings that you have selected the viewport you want to use, for example if you have set up a scene with a camera in, then you want to ensure that you are rendering through the view of that camera – Maya’s default rendering view is perspective. Another great thing is to do a batch render of just one shot as the IPR does not always show you exactly how the final image will look and in my experience can be very different to the actual render. It’s also useful to do this because you can easily notice compositional flaws from one shot and you can change it before you render a lot of frames. When I render I use targa (tga) I have only ever used this one as this one was the file extension we were recommended to use in class. It has worked very well and exports your batch render as a sequence of images that can be easily imported into after effects.

Overall this has been a very enjoyable, stressful, intense and educational experience and I am so happy to have been able to be a part of it.

Thank you very much to the Children’s society and to my tutors for all their help and feedback!