[Soc-2012-dev] Weekly progress report #11 - Smoke Simulator Improvements

[Miika Hämäläinen]

> What you did this week

* I got some promising results with my emission "noise" tests. The idea 
is to use Blender's procedural textures to add some animated variation 
to surface emission instead of emitting evenly from whole surface.
Based on my tests this often adds more natural look especially to fire 
simulations. Only issue I have is trying to make it easier to use. Right 
now lots of time is needed just tweaking the settings and textures to 
get convincing results, and I'd prefer that to be mostly automatic.
* I spent two evenings tracking down that mysterious high res smoke bug 
I have been referring to in couple of previous reports. While it's 
effect is fairly rare in trunk, it is much more common when adaptive 
domain is used. Now I finally managed to find the cause and fix it with 
Daniel. Fix was committed to trunk in r49485. :)
* There is one more serious issue left in smoke code that has been very 
troublesome for fire simulations: The vertical velocity of any 
simulation keeps increasing over time eventually making the smoke/fire 
shoot out of domain in couple of frames. This becomes very apparent for 
fire simulations due to increased heat. It's often visible in fire 
simulations of just 250 frames.
For the end of the week I have been trying to solve this, so far without 
breakthroughs. I can't really focus on fire improvements until some 
solution is found. :(

> What you plan to do next week
I'm planning to continue solving remaining smoke simulator issues and 
start preparing the code for trunk.