Hi,

I've been experimenting with different ways of rendering non-intersecting
grains. One way to approximate this is to make each grain a blob, and then to
subtract negative blob components from it using neighboring points. The method
is similar to how I made the underwater air bubbles, except I now have access to
a list of each point's nearest neighbors, which means the parse time is
*significantly* reduced.

This scene took 11 seconds to parse. It took only 7 minutes 9 seconds to trace
with one 7x7 area illuminating area light. Blobs, when used in this manner, are
very efficient. Total particle count: 10,090.

The Voronoi calculations (using an external program) took about 1 minute to
solve for 32,768 points. However, the output from that program only needs to be
created once. The resulting data can then be used in different ways. For this
image the actual points filled a cubic volume, but (in POV) I kept only the
points which existed in a sphere and below a plane. That is why only 10,090 out
of 32,768 points made it into the final render.

The next challenge will be to vary the particle sizes without having them
produce incorrect results :/

~Sam

Post a reply to this message

Attachments:
Download '10090-grains.jpg' (102 KB)

Preview of image '10090-grains.jpg'

On 04/02/2011 06:24 PM, Samuel Benge wrote:
> Hi,
>
> I've been experimenting with different ways of rendering non-intersecting
> grains.

I see a collision between grain 8972 and 8973 ... just kidding ;-)
Some of them do look like lucky charms marsh mellow bits (the colors) 
You're such a wizard at this kind of stuff  ... fantastic!

Post a reply to this message

On 4/2/2011 2:45 PM, Jim Holsenback wrote:
> On 04/02/2011 06:24 PM, Samuel Benge wrote:
>> Hi,
>>
>> I've been experimenting with different ways of rendering non-intersecting
>> grains.
>
> I see a collision between grain 8972 and 8973 ...

what, how...

> just kidding ;-)

I would hope so, since I didn't number them :P

> Some of them do look like lucky charms marsh mellow bits (the colors)

Yeah, and the regularity of the grain sizes doesn't help...

> You're such a wizard at this kind of stuff ... fantastic!

Thanks, Jim!

Post a reply to this message

"Samuel Benge" <stb### [at] hotmailcom> wrote:
> The next challenge will be to vary the particle sizes without having them
> produce incorrect results :/

Sam, I think we need bad weather in your area more often - this is another
/really/ cool concept you've developed. I can see this being so useful for
gumball machines, or jars of beans, or piles of skulls, or... well, keep going
man; looking forward to seeing what you come up with next!

-------------------------------------------------
www.McGregorFineArt.com

Post a reply to this message

On 4/2/2011 4:33 PM, Robert McGregor wrote:
> "Samuel Benge"<stb### [at] hotmailcom>  wrote:
>> The next challenge will be to vary the particle sizes without having them
>> produce incorrect results :/
>
> Sam, I think we need bad weather in your area more often - this is another
> /really/ cool concept you've developed. I can see this being so useful for
> gumball machines, or jars of beans, or piles of skulls, or... well, keep going
> man; looking forward to seeing what you come up with next!

Thanks, Robert!

No intersection tests aside from pitting one blob component against 
another were used, so to place non-intersecting spherical objects would 
require a different paradigm altogether. The current setup assumes that 
the creation of oddly-shaped objects is expected.

There is the possibility of iteratively checking for intersections, but 
the C++ library I am using changes the order of the points, which means 
I will need to find a way to keep track of those changes before I can 
pack spherical objects.

The experimentation continues...

~Sam

Post a reply to this message

He he... in sedimentological terms, this would be a very well-sorted, 
moderately rounded sand sample :-)

You constantly amaze me.

Thomas

Post a reply to this message

On 4/3/2011 12:55 AM, Thomas de Groot wrote:
> He he... in sedimentological terms, this would be a very well-sorted,
> moderately rounded sand sample :-)

In breakfast cereal terms that's also true. "Contains 3 important 
minerals!" ;)

Post a reply to this message