So, in the following, three images. The image ending on 1 is the Dropsnow
Macro released upon a mesh2-tree from Arbaro *without* the smallest
branches. The result is then placed onto the tree with all branches. It
looks nice, clean and crisp.

The image ending on 2 was Dropsnow released upon the *entire* tree, with all
the small parts. The small "noise" introduced by various very small
components (the smallest branches) would require much higher sampling
settings and, along with higher sampling settings already used to place
enough particles on the tree, the resulting parsing time would increase
exponentially. Of course, for a High-Res Final image, this might be what you
need, and I'm already thinking about how to refine the parsing process and
the resulting data so that small tweaks wouldn't result in an entire
reparsing of the particles, rather than just picking a few different
parameters for the display.

I've got no timings on these, but the simpler one required roughly 5 minutes
or so total, whereas the more complicated one would have required somewhere
around 12 minutes or so, both with roughly 8000 final particles out of a
20.000 initial bunch of which some were discarded due to steepness of
surface, surface too thin, inside mesh, etc.

Overall, I'm very pleased and will have to conduct some more tests on
various objects to see if I need to include a few more parameters for more
control on the User-Side.

And the third and final image is just the macro on a CSG object to show off
the main effects I wanted to achieve: stacking along walls and removing
particles underneath roofs, two of the important issues about snow.

Regards,
Tim

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>

Post a reply to this message

Attachments:
Download 'tree_snow_1.jpg' (49 KB) Download 'tree_snow_2.jpg' (53 KB) Download 'csg_snow.jpg' (15 KB)

Preview of image 'tree_snow_1.jpg'

Preview of image 'tree_snow_2.jpg'

Preview of image 'csg_snow.jpg'

The word to describe this: sweetastic. Have you tried a more massive render,
with a rediculous number of tiny particles? Does your method allow for
stacking, or will this just leave you with a tiny dusting?

Still, a very impressive result, especially for only a few minutes parse
time.




-S
5TF!

Post a reply to this message

> The word to describe this: sweetastic. Have you tried a more massive
render,
> with a rediculous number of tiny particles?

Doing one right now, as I'm off to courses now.

> Does your method allow for
> stacking, or will this just leave you with a tiny dusting?

If you look closely at where the big branches meet the trunk, the snow is
stacked there. Stacking in general is allowed, you'd just have to make
several passes, that's all.

> Still, a very impressive result, especially for only a few minutes parse
> time.

Agreed. I'm very fond of how well this turned out after such short scripting
time. Now I'm tweaking the algorithms to near-perfection. :-)

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>

Post a reply to this message

Tim Nikias wrote:
> So, in the following, three images. The image ending on 1 is the
> Dropsnow Macro released upon a mesh2-tree from Arbaro *without* the
> smallest branches. The result is then placed onto the tree with all
> branches. It looks nice, clean and crisp.

This looks great Tim, along with the animation too.  It looks like a few
more particles are needed for the 3rd image (the sides look a bit "blobby")
but apart from that looks really great.

Would look good in my sunset scene :-)

Post a reply to this message

So, here's a version which required half an hour parsing, 46.000 cylinders
and 1000 spheres in a blob. Note that of the half hour, some times was
needed to first generate the particles and remove some due to steepness of
surface etc. The results of those presteps can (and were) saved to disk and
can be easily accessed for tweaking just the final step.

The result wasn't as good as I expected, there are still stalagmites
found... I'll reprocess the particles and tweak the parameters to see if I
can get a better version of this. But generally, such things are to be
expected with miniscule detail and discrete sampling. I imagine combining
image 1 (where the smallest branches have been left out) with a sampling of
positions without the stacking-algorithm might give much better results.

Regards,
Tim

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>

Post a reply to this message

Attachments:
Download 'tree_snow_3.jpg' (56 KB)

Preview of image 'tree_snow_3.jpg'

> This looks great Tim, along with the animation too.  It looks like a few
> more particles are needed for the 3rd image (the sides look a bit
"blobby")
> but apart from that looks really great.

Thanks. Yeah, the 3rd image is in rough-shape to show proof-of-concept. Some
more particles in some areas might be useful, I'm thinking about how to
implement a manual-particle-addition-macro or something alike for such
cases.

> Would look good in my sunset scene :-)

Unfortunately, the macros will only be released once the image I'm scripting
them for is finished. But then again, it might be fortunate, because I've
got a longer testing period. :-)

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>

Post a reply to this message

Tim,

This macro is excellent. I can see it being used for more than just snow. I
think an interesting use for it would be moss growing on the side of a
tree.

Tim

Post a reply to this message

So, the final version, of which I won't do another excessive test. Here we
have 1700 spheres and 49600 cylinders in a blob, and there are still a few
places where a cylinder pops up when it should be a sphere. This one already
required 50 minutes. The point is that even though I could do long and
excruciating parses with ridiculously high settings, the same effect can be
achieved by combining a simpler, pure sphere-version with a cylinder-version
on a less detailed structure.

OTOH, if you have an admirable piece of hardware as a CPU, the parsing times
could drop to a level where it might be feasible doing them. But these "ifs"
pop up everywhere: *if* the PC where much faster, I'd do some Navier Stokes
Equations in Realtime... ;-)

Regards,
Tim

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>

Post a reply to this message

Attachments:
Download 'tree_snow_4.jpg' (53 KB)

Preview of image 'tree_snow_4.jpg'

> This macro is excellent.

Thank you! High praise... :-)

> I can see it being used for more than just snow. I
> think an interesting use for it would be moss growing on the side of a
> tree.

Yeah, great idea. I guess this surface covering can be used for lots of
different effects (e.g. dust). IIRC someone created a subsurface-scattering
simulation using blobs and image_patterns. The speed-up due to the
pretesting and dividing the area into active and inactive cells should be
worthwhile for lots of these kind of effects.

Regards,
Tim

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>

Post a reply to this message

Tim Nikias wrote:
> So, the final version, of which I won't do another excessive test. Here we
> have 1700 spheres and 49600 cylinders in a blob, and there are still a few
> places where a cylinder pops up when it should be a sphere. This one already
> required 50 minutes. The point is that even though I could do long and
> excruciating parses with ridiculously high settings, the same effect can be
> achieved by combining a simpler, pure sphere-version with a cylinder-version
> on a less detailed structure.

That one looks quite good.  How do you sample the surfaces - do you just 
trace the whole bounding box area or is it some adaptive process?  How 
well does it work with some wind (i.e. with the snow coming a bit from 
the side and sticking to the vertical parts)?

> OTOH, if you have an admirable piece of hardware as a CPU, the parsing times
> could drop to a level where it might be feasible doing them. But these "ifs"
> pop up everywhere: *if* the PC where much faster, I'd do some Navier Stokes
> Equations in Realtime... ;-)

Note parsing speed is not so much a question of the CPU but of general 
system performance.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 23 Sep. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message