I’m going to need to do a minute or so animation for my school project, and I wanted to keep it high-quality but a feasible render-time. I’ll be rendering the whole thing on my own PC, a 2.4gHz quad core. Things to keep in mind: I want to incorporate sounds, I’ll probably render it at about 800x600 px, and I would like to use GI/Photon mapping if possible.
Should I render it in BI or Yafa? Will either/both require the sounds added in post-pro? How difficult is it to fake GI in BI?
-Tic
Lot of questions there, none of them with simple answers.
I like Blender Internal. Much faster results, full use of all features Blender has to offer.
Once you learn BI you can do anything. Many people disagree with me on this, it is just my opinion.
You need to render out still images and combine them later in a video editor. (which can be blender though I highly recommend using an industry standard NLE)
You will add your audio in post.
Faking GI in blender is easy using AO or spotlight-spheres or both. Lighting to me is much more complex than just “faking GI”.
GI, AO, Raytracing, Indirect Lighting, Photon Mapping all of these can be faked or baked using one method or another, None of these things will make or break your animation.
If you want to use Blender Internal, and you want fast results with decent quality, use the light rig and settings in the file I am attaching here. No raytracing, Buffer Shadows = Quick renders
For 2.49b
P.S. and for everyones sake render in 1280 x 720 or 1920 x 1080. I wish we could all just decide to use one ratio and be done with it. Ban 4:3 aspect ratio!!! 
lights-camera-action.blend (159 KB)
You need to render out still images and combine them later in a video editor.
Blender/Yafa don’t do this internally? 
Faking GI in blender is easy using AO or spotlight-spheres or both.
The only problem I have with using AO is that I’m hoping to get both dark and bright areas. (think, hotel hallways -> elevator shafts for the difference in brightness) I suppose that I’ll just have to play this by ear, or change some settings in-between areas.
How steep is the learning curve in BI? I’ve been spoiling myself with the simplicity of Luxrender up until now, but for obvious reasons that’s infeasible for an animation 
I’ve seen so many BI renders end up poorly that I’m hesitant to try it
For AO use the Approximate method and set the error to around 0.6-0.7. Turn on pixel cache. Set energy to 1.3. Use Plain and Add (only). When it is time for final render, lower the error value to 0.25 or 0.1.
I try to do most of my lighting with AO and maps (reflection, angular on world etc…). I only add in lights, typically a Sun (but not that useful for indoor scenes) after applying maps and AO.
Lux is a much higher learning curve than the internal render system. So if you already know that, you should be in good shape.
Don’t “overlight” by placing a bunch of lights in your scene.
Also, my tip is to split the difference on materials between diffuse and emit. By default, when you add a material to an object the diffuse setting is set to 0.5. What I do to brighten things up a bit is to take the 0.5 light contribution and split it between diffuse and emit. So I might try 0.2 diffuse and 0.3 emit. This will keep the total light contribution at the same level, but you get brightness boost without adding lights or render time.
Ok, I get what you’re saying. I’ll have to experiment to find something that works right