There was always a plan B in the event of things going irretrievably wrong. And, unfortunately, things have gone irretrievably wrong.
I will be unable to use my Harry figure in Maya. I imagine the same will hold true for the Sarah figure. I spent some time with Paul Lucking, a VFX tutor. We went through my workflow and tried various things to extract a usable displacement map. In the course of attempting to extract a map several flaws in my methodology became apparent.
The most fundamental error was in my handling of the UV map for the figure. I created the UV map in Maya in an attempt to fix an unfixable problem. What I should have done is create the UV map in Maya first and imported that into Mudbox. Mudbox automatically updates the UV map at higher mesh resolutions but one needs to work on the UV map as one works on the sculpt. This dual process takes place from the bash mesh level (0) up to, and including, level 3.
Another issue was that I had been extracting the maps in 32 bit and 16 bit TIFF format. Apparently this particular format is not very reliable with Maya and it is far better to use 16 bit EXR format for maps. Yet another error was the scale of the model. The figure was about 10 feet tall. Although I had not changed the scale of the figure when transporting between Mudbox and Maya I did scale the base mesh up in Mudbox in order to provide a larger sculpting area. I do not know if either pieces of software had altered the scale of the model during the exporting/importing process. Obviously, it’s important to check that all of the settings for things like size are congruent between the applications.
Despite the fact that the models are unusable and will not ever be able to be displaced properly during the course of Paul’s investigation I got to see how effective manipulating node attributes can be in improving the behaviour of displacement maps. The aplha gain attribute found in the node’s color balance section is most effective at bringing out the peaks and troughs of the maps.
Originally, my plan B was to use the Autodesk Character Creation website to create two models. The down side to that is that the models would be adult. And not very appealing (to me). Paul suggested MakeHuman, a piece of open source software that allows you to quickly sculpt human figures. It also creates UV maps and has a host of other useful functionality. You can also make child figures which is perfect for my needs.
He also suggested SubstancePainter. I have registered with Allegorithmic and downloaded the suite of software available for students.
The third thing he suggested, for clothing, is MarvellousDesigner. It is prohibitively expensive for me to buy but I may try and persuade the University to get it in.
For this module, my intention is to try and quickly prototype two figures in MakeHuman and rig them ready for animation. As far as the set goes I am going to simplify. Many of the buildings will be replicated. There will be no rubbish, bins, fences to indicate a post-apocalyptic urban environment. I may scale back to two vehicles as I kind of have them made.
I do not know if I can produce a satisfactory 60 second animation. There are only a few weeks left and I still have an animatic/soundscape and November’s 11 Second Club. Fuck. I may have to chalk this one up as a loss.
If I am to continue with my work in 3D I am going to have to get to grips with the pipeline for sculpting and animation. Although Mudbox is enjoyable to sculpt with the talk amongst my peers is that Mudbox has been eclipsed with the arrival of newer, more powerful software. The fact that there is little to no online community of people working in Mudbox does not bode well for its future. One has to wonder whether the folding of Mudbox’s features into Maya is a prelude to Mudbox’s discontinuation.
Paul mentioned a ZBrush module next semester. If it’s possible I would like to join that module. As it currently stands my knowledge of the pipeline from sculpt to animation is so lacking as to make the task impossible. In all honesty, I do not think I want to tell my stories in 2D. I do not know if I can. As I have stopped doing stop motion I have to learn how implement a 3D pipeline. Otherwise I’m a bit screwed.