AccuRender nXt

Just messing around with the newly improved decals. This is a single block of 2 3D Faces inserted 726 times. I set the refractive bounce to 20 (probably higher than it needed) to make sure you never see the Alpha Channel of the decal.

"dkh007" wrote in message news:4804@nxt.accurender.com...

Just messing around with the newly improved decals. This is a single block of 2 3D Faces inserted 726 times. I set the refractive bounce to 20 (probably higher than it needed) to make sure you never see the Alpha Channel of the decal.

---

Daniel Hargreaves
aka: Lord Darth Daniel, Master of Anti-Light

---

http://nxt.accurender.com/forums/p/1101/4804.aspx#4804

wrote in message news:4804@nxt.accurender.com...
> Just messing around with the newly improved decals. This is a single block of
> 2 3D Faces inserted 726 times. I set the refractive bounce to 20 (probably
> higher than it needed) to make sure you never see the Alpha Channel of the
> decal.

Wow. Did you do some kind of array, or actually
insert 726 instances manually?

Gene

The first time I did an array and was not pleased with the results. So I did a copy rotate with groups with much better results. My next step is to start to work on a random insertion routine.

Just a thought, it might be more efficient in this case to use actual grass geometry.  The decal is forcing you to up your refraction samples.  This could really slow renderings down under certain conditions.  If you use actual blade geometry and instance it, that will be efficient and allow you to keep refractions at a reasonable level.

I've attached a lisp program that creates squares of grass using 18
different simple blocks (grass blades) that are randomly distributed and
scaled along a grid within the square. The actual insertion on the grid
line is also randomly adjusted so that they don't all fall on the grid.
It uses a "Random Numbers" lisp program that was not generated by me.
Right now it puts them pretty close together so it's best to keep the
squares relatively small for AutoCAD's sake. The lisp program creates
the block definitions.

Steve M

Splash wrote:
> Just a thought, it might be more efficient in this case to use actual
> grass geometry. The decal is forcing you to up your refraction
> samples. This could really slow renderings down under certain
> conditions. If you use actual blade geometry and instance it, that will
> be efficient and allow you to keep refractions at a reasonable level.
>
> ------------------------------------------------------------------------
> Scott Davidson
> Robert McNeel & Associates
> ------------------------------------------------------------------------
> http://nxt.accurender.com/forums/p/1101/4828.aspx#4828
>

Sorry,

Here are the attachments

Steve M

Steve,

Thanks for the routine. That is pretty slick... it created 18 different blades (blocks) instered 17,136 times. A total of 60 faces instanced 57,202 times. I let this run for over night (7 hours) with over 1,700 passes.

Just a thought, perhaps a little random scaling of the block might help it look more natural.

Splash:

Just a thought, perhaps a little random scaling of the block might help it look more natural.

and maybe a patch of dry/dead grass where my dog "visited"

"KSL" wrote in message news:4848@nxt.accurender.com...

Splash:

Just a thought, perhaps a little random scaling of the block might help it look more natural.

and maybe a patch of dry/dead grass where my dog "visited"

---

Kevin Lockwood
Saving the Worlds Bacon

---

http://nxt.accurender.com/forums/p/1101/4848.aspx#4848

Hmmmm... or a crop circle. now that could be fun ;)