#ct5anigp

Here’s a quick breakdown reel I threw together.

Finally, I have always been very critical of my own work, and I will be and am this time too. Given the chance to improve on this work I would, looking at it now I still find areas which need major improvement. My main focus and concern in this project was creating a Pixar/Disney worthy emotional performance. I feel like I fell short of my aim, but I also feel like the skills, experience and knowledge acquired in the process of creating this short dialogue are more than the vital building block for where I want to be with my animations.

The animations still feels off in certain areas, her hand movements are not cemented and feel unsure, because they are so – I lacked appropriate reference for the hand gestures, so I created much of it as I went along with little performances as I worked. Yet these were incredibly variable and from session to session I explored different routes the hands could take – and this resulted in an unsure performance.I am quite happy with the facial expression and lip sync, I feel that was an area I was relatively successful in, although given the option I would push even further some of the movements and expressions as looking back they seem quite restrained, also the timing of it all could be perfected and overall the animation could use with tweaks to reduce the floatiness of some of the poses.

Initially completely against finding the original clip where the audio was taken from, nearing the end of the animation I realized it could only help drive the emotions home, comparing to the performance I created to the original would have helped – and it did, even if minimally, because what I realized is that the facial movement in the original performance was minimal.

This made me realize another thing – realistically, facial movement IS minimal, eyebrows, cheek muscles, nose and lips during speech and dialogue and expressions all move incredibly subtly into poses we can however easily recognise and instantly read in the human face.

Leading to another realization – that therefore my animation could not be a REALISTIC identical to life dialogue mirroring – because in a 3d model as stylized as Mery, that would not be read well at all – and would look minimal and empty of emotions.

I instantly thought of Disney's portrayal of emotion in it’s 3d feature films, and given the striking resemblance between Mery and some of the characters in the feature film Frozen, I looked toward it as a point of reference for how speech was synced and emotion clearly expressed and driven to the audience.

I found that many of the emotions and facial expressions were pushed and exaggerated in order to be readable, and that majorly the eyebrows and eyes were what did most of the talking, and that the lips did not move to match every word and mouthshape (something I had previously learned, but was reminded).

Not particularly skilled with Arnold or rendering in general – this was a major learning curve in the project- Initially I was set on creating a high quality playblast , but near the end of the project (perhaps after a hard hit on the head) I though it was possible for me to created the cinemating background I wanted in the scene initially. 

Short on time I knew I had to create render layers and that Mery was not renderable. The main problem with her being that she was created with and for Maya 2015. Thus making the transition to Arnold very arduous, and even though I attempted to – I was unable to make her ultimately work with the rendered and she had numerous display errors I could not find a workaround. 

With the aid of tutorials and AiPhysicalSky I created the sky and clouds I needed for the background – the lamps were already modelled with a meshlight and the materials were all adjusted to resemble wet metal, and had I more time I would have worked on creating puddles that mirrored the final concept.

I decided to render a single frame and use it as a background and have a high quality playblast for mery’s performance – I however had to match the lighting of the scene – so I set up some lights that emulated the lighting in the rendered scene and playblasted her alone with a blue background ready for compositing and keying in aftereffects.

The three major props in my scene – they are responsible for the background scenery.All the metals and material are AiShaders tinkered to resemble wet metal – the lamp contains an MeshLight

Had I more time, the scenery and background is what I would have worked more on – however I spent innumerable hours on the animation alone and trying my best to get everything the best I could.

For the choosing of the model I wanted something as resembling as possible to the girl I concepted with (perhaps after getting visually attached).But having no intention to spend time on modelling and rigging my own, finding something that even came close was complicated.

I stumbled upon the MeryProject by José Manuel García Alvarez and Antonio Méndez Lora – an incredibly professionally designed rig which is offered to animators.Not incredibly customizable besides simple color and some clothes – she however is incredibly well rigged and complex to use.

It was my first time using anything similar to this rig. And it was a learning progress, but by the end I fell in love with everything from the controllers to the numerous features she has – and will in fact be using her again in future animations.