Hi Kristie,
I'm myself producing fulldome video with Maya for the 2 greek planetaria (Eugenidio/Athens and Tmth/Thessaloniki) so I think I can explain the different workflows and caveats ..
The main issue is how to render a fisheye image. For that two different ways :
1.
use mentalRay and a fisheye lens shader "plugin" (i.e domeAFL, MR_aperture). The camera has to placed in the +Y axis with tilt corresponding to your dome config. (a bit like you'll do will a real fisheye lens on a camera pointing upwards). The rendered image doesn't need any more tweaking and can be used straight forward. In this case, you can render straight to TGA (24 bits) and send to your splitting application.
Problems :
-No way to preview in the viwport what the fisheye camera sees (need an additional rectangular camera linked to it)
-filtering on texture images can create problems (bugs with quadratic, only mipmap seems to work ..)
-No lens flare (glow is now supported in MR), no hardware particles or sprites.
-Can result in extremely long render time by frame with the risk of having memory problems (especially if rendered in 4k).
- In case of renderfarm, licenses for MentalRay need to be purchased for each node.
- Maya render with a 5 (or 6) camera setup. Each view in rendered separately and recombined in a fisheye image. Here you have all the possibilities of Maya, except for event happening as post-process (i.e glow) which can create problems when the image is recombined. For stitching, we're using wether Glom or the Fulldome plugin for Adobe After Effect from Fundation Navegar. (check www.fulldome.org for more infos).
In this case I recommend rendering in a lossless format (i.e maya iff) depending from your stitching application, or you'll have to deal with enormous amount of data.
Problems :
huge amount of files and render queues.
Again no real way to preview what is really going on. I usually open the 5 cameras views in separate windows placed side by side to have a better understanding on the animation. This way you can also detect what is animated or not on each camera and decide to render only specific frames for each specific camera ...
So that's the two techniques you can use for rendering.
For the rest (modeling, animation ..) of course the main difference is that all objects need to be in high resolution, and you need to have something to show all around you, which means very heavy scenes to work with, render and (worst of all) compose or correct.
Hope it helps, don't hesitate to ask if you have more specific questions
Guyb.
