I think the most challenging parts of this project were the ideation and conceptualisation phases. It was tough trying to design characters and settings that would support the comedy, look believable and be appealing.
Looking back I think I spent a lot of time unwrapping UVs that then didn’t need to be used so perhaps I should have unwrapped for example one bottle/container and experimented on that one item. A lot of time was also spent re-unwrapping and re-texturing the second design of Snappy (also true for Blayne and Katie who had to re-rig and re-model) so perhaps we should have looked for more feedback at the design phase in order to avoid doubling the work.
I was glad I took the chance to try out Mudbox even though I wasn’t sure it would work for me. I still think that the lighting and texturing could have been taken further – it’s missing a quality which I really like in works such as One More Beer!, No More Wine and Moments on the Cat Planet, but this may have also been due to the design and creating a believable fantasy world. Some simulation/dynamics may have added to the short also. For instance, I like how in One More Beer! the Viking interacts with the environment by knocking over a tankard of beer and also how he sends up a cloud of dust and shakes the table when he sits down to show his weight more. We also didn’t get to look into creating frothy tooth paste like we wanted.
It was still fun and I enjoyed working with my team:) They made the hard work worthwhile.
Blayne suggested that we should render the three layers – foreground, character and back wall of our scene separately. This saved hugely on the render time (especially because the background only needed to be rendered twice) and allowed for easier colour and contrast editing at the end. I rendered two different light set-ups for before and after Snappy turns on the light and then Blayne was able to tweak the background layer in Sony Vegas to make it look like sunlight coming through the back window. I tried some different versions of warm/cold between the foreground plus Snappy and the back wall in Photoshop for a quick visualisation of options.
As usual, rendering resulted in a few problems. For one I had forgotten, while setting up the lighting, that you can preview the light in the viewport. Because of this, and despite doing a few spot renders with the animation in place, I missed a few problems e.g. Snappy’s head became too bright when he leaned too close to the light and his teeth fell into darkness for too long when he leaned backwards. One of his eye textures also failed to load. I also had to make sure that all the lights were linked correctly throughout the short e.g. the geometry of the toothbrush changed when it snapped and needed to be re-linked. Luckily these problems weren’t too difficult to fix. For the eye and teeth problem we were able to reload the texture/set up new lighting and cover everything, except what needed to be fixed, with a black surface shader. Blayne was then able to layer these new renders in Sony Vegas. For the lighting on the head, I just keyframed the position and shape of the area light so as to keep the intensity and spread of light even throughout and re-rendered the affected frames.
Blayne had also rendered an ambient occlusion pass but we didn’t understand the correct way to apply it at the time and put it over the direct light instead of a diffuse pass first. Alec showed us this tutorial which I intend to go through:
I started to experiment with creating volumetric lighting in Maya but I found it difficult to get the look I wanted. At first I was using a spotlight and creating a light fog node but when I tried to shine it through a window shape the spotlight would shine through the geometry of the wall. I realise now that this was probably due to the samples being too low.
I was also finding it hard to make the light rays appear through the window.
Alec reminded me of this tutorial so I went back and tried it out as an alternative to using the spotlight fog node:
Maya: Creating Volumetric Lights in Mental Ray (Long Version)
This method uses volume primitives to add a container of fog within the scene. I liked that this method affects area lights as well as spot lights and I also liked how soft it can be. However, the fog rays were invisible from the angle I wanted to render at so I stopped experimenting at this point so as to work on other parts of the short.
We wanted to create an atmosphere that would be suitable for a scary monster but not something too dark and grungy. I experimented with what I had read about the lighting in American Cinematographer about Penny Dreadful and how they were encouraged to push the black. I found it hard to get a balance between dark shadows and still getting a clear read of the scene. I also didn’t want the shadows to be too deep and without detail. I experimented a bit with using image based lighting to lighten the shadows as Alec suggested but didn’t spend enough time playing with this. Final Gather also helped to lighten the shadows but added a lot to the render time.
For the wood texture I painted over a photograph of wood and then used my painted layer alone as a bump map. I tried roughing up the wood of the mirror with a painted bump map also but the team preferred the mirror being smooth.
I used mia_material_x presets for the other assets. I had unwrapped everything with the intention of adding more details such as worried surfaces and labels. I also shared the unwrapped and lit project files with Andrew to experiment with. Katie gave her eyeball model to me to put in the bottle like we originally intended but then we though it looked quite distracting (like you can see above). I would have liked to see a Brian Miller inspired graphic look on this also like I mentioned in a previous post – link
I hadn’t unwrapped UVs in a while so at least it was some practice. I also tweaked some of Blayne’s models to have more detail that would catch the light such as on the taps.
I brought Snappy into Mudbox to experiment with creating detail using the sculpt and paint tools there. I then exported my layers back into Maya as texture maps and tried applying them onto materials with different attributes. I had tried making a specular map on the scales but it didn’t work as I had expected. Also at this stage Snappy had not being rebound to his animated rig so I had trouble trying to add detail to the inside of the mouth. I tried using the pose tools in Mudbox for a quick solution but that didn’t work so well for me either. The displacement map was also giving me bother at first so I read through some Mudbox help pages:
In the end I found the problem was because I hadn’t connected a missing edge between Snappy’s feet. I then had to make sure that this was also fixed on any models that I was transferring UVs to as this had prevented the transfer in the previous model/design of Snappy. My .exr files then decided to stop working which I found out was due to another plugin not loading:
I tried using my paint layer as a bump map which created more pleasing results and caught the light nicer. Katie and Blayne liked both so I also tried a version where some of the bump map was softly erased around the mouth.
I experimented a little with how the eyes could look. I tried plain humanoid eyes and black eyes with an incandescent white center. Reptilian eyes may also have been an option but we liked the black/white ones enough to stick with them.
Gerard unfortunately didn’t get the same texture treatment which was a lot due to time. Katie made a colour scheme that would stand out against Snappy. I then tried a simple dot texture by tweaking the Maya default textures. We were afraid it looked too much like camouflage/dappled light which would make Gerard more difficult to see so we kept him as plain colour solids in the end.
We got further feedback from our classmates, the final years and Greg. The class consensus seemed to be that ore original idea was funnier than our proposed revised version of a monster with one tooth being the main character. Greg gave us further feedback about our monster designs for Snappy and Gerard. He said that in our later animatics we were loosing the contrast in shape between Snappy and Gerard, Snappy being broad and hulking while Gerard being small, tall and thin. Greg also suggested that Snappy’s horns were distracting and not needed for the story. I personally liked the graphic image of the horns against the backlight and thought as a design element they could be an icon of monster-ship but I also see how their absence of height emphasizes the hulking shape in Snappy’s new design. More teeth were also called for:D Blayne re-designed, re-modeled and re-rigged Snappy with the feedback in mind!:) My job was to re-light and re-texture.
Goodbye Aku-inspired horns.