On Jul 31, 2011, at 9:14 PM, Guilherme Kunigami wrote:

> Indeed. The only thing is that we'll need a buffer to keep partial sums. As 
> far as I udnerstodd, in incremental mode, each pass is independent from the 
> other.

Good point, but then if done cleanly, that "working buffer" will be highly 
useful for other purposes too.

> Do you mean adding an option like the lightmodel? Something like -fb #

Something like that..   There's basically two operations going on:  1) in what 
pattern are primary rays being fired, and 2) how frequently is the framebuffer 
being updated.
 
> Where #
> 0 - unbuffered
> 1 - incremental
> 2 - line-buffered
> 3 - frame-buffered
> 
> Is the frame-buffered mode this one I just suggested?

The 0-3 you described basically are forms of #1, patterns for the primary rays. 
 Currently, there is also implicit framebuffer updating going on with each of 
those modes too:

A - per ray (unbuffered pixels)
B - per level of resolution (replicated values, full-frame updates)
C - per line (buffered lines)
D - per image (fully buffered frames)

> I didn't get this "random streaming" mode. How does it work?

Basically, I was suggesting a #4 - random, which would pair well with 'A' or 
'D' with hypersampling.

> Do you mean converting a non-OSL shader into a OSL one automatically?

Basically, yes.

Of course, many of our parameters are meaningless to OSL, but at a minimum it 
could recognize lights and treat everything else as a simple diffuse surface.  
Basically, a render mode (perhaps toggled by a command-line flag or even just 
an environment variable) that would bypass liboptical's normal handling and 
make everything go through OSL.  The goal would be to render any scene using 
OSL without having to modify the scene.

Cheers!
Sean




------------------------------------------------------------------------------
Got Input?   Slashdot Needs You.
Take our quick survey online.  Come on, we don't ask for help often.
Plus, you'll get a chance to win $100 to spend on ThinkGeek.
http://p.sf.net/sfu/slashdot-survey
_______________________________________________
BRL-CAD Developer mailing list
brlcad-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-devel

Reply via email to