I just remarqued a peculiar detail:
In the SEM omage, the lower edge of the indents are never visible, while 
on the POV version, they are visible starting with the second row.

My proposal:
Use an orthographic camera.



Alain

Post a reply to this message

"kmcpeak" <kmc### [at] ethzch> wrote:

>
> After some movement of the lights and adding bumps to the etch pits attached is
> my latest best effort. Might stop here...
>

That's looking almost perfect. Well done, really.

But don't stop yet ;-)  It's always the last 5% of the work (or maybe the last
1%!) that really 'sells' a computer graphics rendering. And that's usually the
most difficult and hair-pulling part, because the final tweaking is so subtle at
this stage.

The *only* thing I would say is that the bright parts of the indents don't have
quite enough 'tonality' or fading-to-darkness as they recede into the pits. Your
lights code looks right, so I'm thinking that the gradient y pigment might be
scaled too deeply into the surface (in other words, fading to black somewhere
*below* the bottom of the pits.) OR, maybe try reducing the diffuse value in the
finish. Which will naturally make the SURFACE darker too, of course ;-/  Yet
more tweaking...

But the result will be worth it!

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:
> "kmcpeak" <kmc### [at] ethzch> wrote:
>
> >
> > After some movement of the lights and adding bumps to the etch pits attached is
> > my latest best effort. Might stop here...
> >
>
> That's looking almost perfect. Well done, really.
>
> But don't stop yet ;-)  It's always the last 5% of the work (or maybe the last
> 1%!) that really 'sells' a computer graphics rendering. And that's usually the
> most difficult and hair-pulling part, because the final tweaking is so subtle at
> this stage.
>
> The *only* thing I would say is that the bright parts of the indents don't have
> quite enough 'tonality' or fading-to-darkness as they recede into the pits. Your
> lights code looks right, so I'm thinking that the gradient y pigment might be
> scaled too deeply into the surface (in other words, fading to black somewhere
> *below* the bottom of the pits.) OR, maybe try reducing the diffuse value in the
> finish. Which will naturally make the SURFACE darker too, of course ;-/  Yet
> more tweaking...
>
> But the result will be worth it!

Thanks for the encouragement Kenneth! I will keep plugging away. One issue
regarding the pigment gradient is that not all the pyramids have the same y
value for the bottom of their pit. So I struggled finding the *best* value since
it worked for some but not all. Do you see any way around that?

Alain, I will also give the orthographic camera a shot, thanks for that
suggestion.

Kevin

Post a reply to this message

"kmcpeak" <kmc### [at] ethzch> wrote:

>
> Thanks for the encouragement Kenneth! I will keep plugging away. One issue
> regarding the pigment gradient is that not all the pyramids have the same y
> value for the bottom of their pit. So I struggled finding the *best* value since
> it worked for some but not all. Do you see any way around that?

Ah, I didn't think of that. My own little test had the indents bottoming-out at
almost the same depth (give or take a tiny amount.)

There actually *is* a way to give each indent it's own special gradient-y
treatment(!) But it's crazily complicated, so I won't go into great detail; here
are the basics, though:

POV-Ray has a pattern called BOXED. In essence, it can be used to 'encapsulate'
any of POV-Ray's other patterns (like gradient y.) In this use, it's kind of
like a 'once' statement for patterns--the pattern on the inside of the 'box',
and a transparent rgbt 1 added for everywhere outside.

Anyway, first your entire SEM model would get its own texture/pigment (NOT a
gradient-y this time but just whatever makes the surface look good--bumps,
averaged, whatever.) Then multiple 'boxed' textures are added with their
encapsulated gradient-y's--so each indent get its own boxed pattern to override
the 'main' texture just in those places (with the boxes appropriately scaled of
course, which is the tricky part--and placed right under the surface.)

There is a method in POV-Ray for almost anything!

Post a reply to this message