[Soc-2007-dev] It's Friday again!

[Joe Eagar]

This week I finished a (very rough, buggy) implementation of DSM.  I had 
a test render to show off, but photobucket.com is down (and I hate 
imageshack).

Basically, blender's abuffer implementation is used to create a DSM.  
Global material alpha per vlak (render face) is used for transparency.  
The DSM is fully tiled; however, at the moment it only works if rendered 
in one tile.  There's also a problem that (for some reason I can't 
imagine) the DSM cannot be setup at the same place in convertblender.c 
as normal shadow buffers.  At the moment their created/destroyed in the 
same place as irregular buffers, which is very bad (since they then get 
remade each render tile -- in fact, at the moment they only work in one 
render tile).

Next week (well and this weekend, in fact right after I finish this 
email) I'm going to do some debugging.  Work out the major bugs.  
Eventually, I probably should render DSM maps in solid/transp layers 
like the rest of the renderer does, to make blender's abufferer's life 
easier memory wise.  I'm not sure if this is a priority though.

I also had an idea for blurred shadows I plan on implementing.  I 
noticed a while ago, that if you duplivert a bunch of spots to a small 
sphere, all pointing in the same direction, you get much better blurred 
shadows (especially if each of them individually has a small amount of 
blur going on).

Joe

Chris Want wrote:
> Hi students,
>
> Please take about 10-15 minutes today to answer the
> following questions in your weekly report:
>
> 1) What did you do this week?
> 2) What will you do next week?
> 3) Do you have any problems that need particular attention?
>
> Also, do you have any questions or comments?
>
> Regards,
> Chris
> _______________________________________________
> Soc-2007-dev mailing list
> Soc-2007-dev at blender.org
> http://lists.blender.org/mailman/listinfo/soc-2007-dev
>
>