One of the elements of production I had not expected to find myself dealing with was the blend shapes for the boy character, as this had been a job that was expected to be completed alongside the rigging task that Scott had offered to complete. Indeed, there had been some lovely facial movement occurring on the rig that gave the boy some intense expressions that would work lovely within the animatic short.
However, mere days before the deadline the rig was subject to a revision to correct some controller, geometry and weight issues that had been identified during the animation stage. I am still unsure what may have happened during the revision stage, but when the boy was returned to the team those blend shapes that existed previously had disappeared entirely.
I was in contact with Scott immediately once this was discovered, but he was unable to shine some light on the situation. I enquired if it would be possible to simply import existing blend shape bases into the scene in order to recreate them, but he was unable to guarantee if that would be successful or not. Knowing that Scott was under his own immense pressure as the deadline approached, I made the decision to try to identify the problem and do what I could to correct the issue.
Provided that I have never done any facial animation before, creating facial expressions in order for that to be completed was an interesting experience.I referred to a lot of tutorials (located here) as part of my research before beginning the task, so that I could have a comprehensive understanding of the technical process that was required in order to produce blend shapes.
(E.1 – Shown above, the blend shape head manipulation sources)
The process was not as difficult as I had perceived it to be. Granted, in total it had taken me three attempts to manage to successfully have my original head stay in position and have the adjustment heads influence the vertices on that, rather than disconnecting from the boy’s body to transition towards the target head – that had made me want to tear my hair out in frustration. However, I was aware that every time that I was required to restart the process, the faster my workflow increased. The technical ability to create blend shapes had improved from zero to acceptable, enabling me to create expressions for the boy that I personally would deem okay. The feedback from the rest of the team was generally positive too, which was also reassuring.
Some of the areas in which I struggled during the blend shape production were around the eyes and eyebrows, due to the inability to create symmetrical changes to both sides of the face. The minor manipulations that had been created using that tool created issues for the animators, as the blend shape did not work if the head was tilted downwards – front facing, however, still continued to function as expected.
I really developed an understanding of how to use additional features within the Maya tools to help the expressions come to fruition – soft select, in particular, was very efficient in adjusting the placement of vertices in small sections. Whilst there were periods that I had to turn soft select off and manually move singular vertex at a time, soft select allowed me to create the general shape of the expression as a base to work towards. The sculpting tools – specifically the smoothing tool – corrected some of the deformation and shadows that had appeared as a result of drastic geometry manipulation.
The eyebrows were also challenging – creating the blend shapes for one side of the face had been a breeze – the difficulty was trying to match the same deformation on the other side also. Whilst there may have been a proficient way of doing so, I was unable to locate a tutorial that provided any insight into doing this task with a different approach. Therefore, I used my intuition to judge the deformation by eye, hoping that I could make them appear symmetrical enough that it would not be obvious to the casual viewer that they were not entirely matching.
(E.2 – Shown above, a render of the blend shapes that were created for the boy’s facial expressions)
The difficult situation with the blend shape creation was that there was very little hours that could be spared in order to create them. If the situation had of occurred a few weeks previously, I feel that the additional time that would have been available would have enabled me to research some character facial expression sheets and build quite a dynamic range of expressions and facial movements. However, despite the restrictions that were in place, I believe that I was able to create a sufficient amount of expressions to communicate the emotion and reactions of the boy.
My technical ability to create blend shapes developed from the requirement to fix a significant issue in a short period of time. Whilst I feel that there is still a significant amount of research to conduct in order to prepare blend shapes that would facilitate greater facial deformation and lip syncing, the first attempt at creating expressions for a character model appears to have been successful.