48c1017b@news.povray.org...
> And another one, this time I tried a high-pass filter on the ambient
> occlusion image before using it in the image to localise the placement
> of the corrosion/dirt/mold. I think the effect is a bit better for it.
>
> What do you think?
>
> Cheers,
> Edouard.
>
I agree it's better (maybe due to textures as well as process on the 
occlusion image.
I see more moss than corrosion though :-)

Marc

Post a reply to this message

m_a_r_c wrote:

> I agree it's better (maybe due to textures as well as process on the 
> occlusion image.
> I see more moss than corrosion though :-)

Heh - that's why I changed the statue to cheesy marble - it looked more 
like moss or something else organic to me too.

> Marc 

Cheers,
Edouard.

Post a reply to this message

This one looks good indeed!

The metal in the previous image seems too shiny to me, in combination with 
the corrosion. You could try to tarnish it some more. Maybe the corrosion 
should be a paler greenish tint?

Thomas

Post a reply to this message

Edouard Poor <pov### [at] edouardinfo> wrote:
> What do you think?

I think you are on the right track! This last image looks good, but not like
bronze, IMO. More like turquoise. This could be due to my monitor's gamma
setting, though. How are you applying the image? Rune's illusion.inc?

Dirt-mapping is something I always go back to. Lately I've found that Blender
can produce good UV maps that you can burn ambient occlusion onto.

Keep up the good work. I look forward to seeing more interesting images from you
:)

Sam

Post a reply to this message

Samuel Benge wrote:
> Edouard Poor <pov### [at] edouardinfo> wrote:
>> What do you think?
> 
> I think you are on the right track! This last image looks good, but not like
> bronze, IMO. More like turquoise. This could be due to my monitor's gamma
> setting, though. How are you applying the image? Rune's illusion.inc?

Since the first one looked so little like bronze with corrosion, I tried 
  more of a stone with mold look for the second one.

I'm applying the image with Rune's illusion.inc - I only started using 
that a week or so ago, and there's so much that you can do with it! I 
tried to do the mapping myself, but failed badly - luckily there are 
plenty of people out there who are much better with maths than me...

Hmmm, just had a thought - I wonder if you can do this trick in a single 
pass using the camera_view pigment in Megapov? Divide the universe into 
the same scene twice, one with AO and one with final lighting, then 
camera_view + illusion.inc's maths to pull the AO data into a 
pigment_pattern? Sounds both doable, and far trickier than I could 
possibly pull off. :-)

> Dirt-mapping is something I always go back to. Lately I've found that Blender
> can produce good UV maps that you can burn ambient occlusion onto.

I downloaded Blender the other day. I'm frankly a little lost in it - I 
need to find some good tutorials or similar...

> Keep up the good work. I look forward to seeing more interesting images from you
> :)

Thank-you very much!

I'm still trying to get my head around your proximity and related SSS 
macros. Applying the AO image works nicely for some scenes, but your 
proximity macro is a global solution, and not just a camera specific one.

> Sam

Cheers,
Edouard.

Post a reply to this message

Edouard Poor <pov### [at] edouardinfo> wrote:
> I downloaded Blender the other day. I'm frankly a little lost in it - I
> need to find some good tutorials or similar...

I sympathise. If you just want ambient occlusion baking then maybe it's worth
looking at xnormal (http://www.xnormal.net/2.aspx). Either way, baking the
occlusion effect into texture maps is likely to be the most efficient solution
for all-mesh scenes (if you're lucky enough to have them).

Tom

Post a reply to this message

Edouard Poor wrote:
> I'm applying the image with Rune's illusion.inc - I only started using 
> that a week or so ago, and there's so much that you can do with it! I 
> tried to do the mapping myself, but failed badly - luckily there are 
> plenty of people out there who are much better with maths than me...

Rune's illusion.inc is *very* useful! I wonder if he knew how it would 
be used when he made it? I too tried to map an image to the viewing 
area, but failed miserably. Thank goodness he figured it out for us ;) 
He should get some sort of award for that file!

> Hmmm, just had a thought - I wonder if you can do this trick in a single 
> pass using the camera_view pigment in Megapov? Divide the universe into 
> the same scene twice, one with AO and one with final lighting, then 
> camera_view + illusion.inc's maths to pull the AO data into a 
> pigment_pattern? Sounds both doable, and far trickier than I could 
> possibly pull off. :-)

The camera pigment takes render options from the whole scene file. I 
suppose you could place a copy of the mesh in an ambient cube with 
radiosity applied, but you would have to use radiosity for the final 
image. But with MegaPOV, can't you turn radiosity off for certain 
objects? If so, you can probably do it without having to use radiosity 
for the final output.

> I downloaded Blender the other day. I'm frankly a little lost in it - I 
> need to find some good tutorials or similar...

I recommend that you don't give up on learning Blender. While looking at 
tutorials, keep your eye on the keyboard shortcuts, as they make the 
program much easier to use.

> I'm still trying to get my head around your proximity and related SSS 
> macros. Applying the AO image works nicely for some scenes, but your 
> proximity macro is a global solution, and not just a camera specific one.

But you can't use the proximity macro for mesh objects :( If POV 
supported calling inside() for mesh objects, it would work. I wish 
somebody would figure out a way to get inside() working with meshes!

Sam

Post a reply to this message

stbenge wrote:
> But you can't use the proximity macro for mesh objects :( If POV 
> supported calling inside() for mesh objects, it would work. I wish 
> somebody would figure out a way to get inside() working with meshes!

You can make solid meshes now, which *should* work with inside().

Define a mesh using the inside_vector modifier, and see if that works.

-- 
William Tracy
afi### [at] gmailcom -- wtr### [at] calpolyedu

"Given a signal with a 2-Hz bandwagon, the Nyquil frequency must be 1-Hz 
to avoid aliening of the furrier transformer."
     -- phdcomics.com

Post a reply to this message

William Tracy wrote:
> stbenge wrote:
>> But you can't use the proximity macro for mesh objects :( If POV 
>> supported calling inside() for mesh objects, it would work. I wish 
>> somebody would figure out a way to get inside() working with meshes!
> 
> You can make solid meshes now, which *should* work with inside().
> 
> Define a mesh using the inside_vector modifier, and see if that works.

I've tried it, and it did not work :( I used the simplest mesh I could 
think of--a tetrahedron--yet still I had no success. If you get it 
working, by all means, post the code! I'd love to bypass the whole 
AO-burning process altogether :)

Sam

Post a reply to this message

stbenge wrote:
> Edouard Poor wrote:
>> I'm applying the image with Rune's illusion.inc - I only started using 
>> that a week or so ago, and there's so much that you can do with it! I 
>> tried to do the mapping myself, but failed badly - luckily there are 
>> plenty of people out there who are much better with maths than me...
> 
> Rune's illusion.inc is *very* useful! I wonder if he knew how it would 
> be used when he made it? I too tried to map an image to the viewing 
> area, but failed miserably. Thank goodness he figured it out for us ;) 
> He should get some sort of award for that file!

It's so useful it should be built into POV... "map_type rune" :-)

>> Hmmm, just had a thought - I wonder if you can do this trick in a 
>> single pass using the camera_view pigment in Megapov? Divide the 
>> universe into the same scene twice, one with AO and one with final 
>> lighting, then camera_view + illusion.inc's maths to pull the AO data 
>> into a pigment_pattern? Sounds both doable, and far trickier than I 
>> could possibly pull off. :-)
> 
> The camera pigment takes render options from the whole scene file. I 
> suppose you could place a copy of the mesh in an ambient cube with 
> radiosity applied, but you would have to use radiosity for the final 
> image. But with MegaPOV, can't you turn radiosity off for certain 
> objects? If so, you can probably do it without having to use radiosity 
> for the final output.

My AOs are done with a simple light dome. So you'd make a copy of the 
object off in it's own universe (some random place a long way away). 
Stick it in a giant sphere if nessessary. Then create a light_group of 
the AO lightdome (so as to not have those lights interfere, or slow, the 
real scene). Then use a camera_view to generate the pigment for the 
pigment_map on the real object. Probably be mostly achievable in a 
clever macro? But it almost sound like more work that it's worth 
considering how multiple renders are easy... Would get points for "extra 
degree of difficulty" though :-)

>> I downloaded Blender the other day. I'm frankly a little lost in it - 
>> I need to find some good tutorials or similar...
> 
> I recommend that you don't give up on learning Blender. While looking at 
> tutorials, keep your eye on the keyboard shortcuts, as they make the 
> program much easier to use.

I'll stick with it :-)

>> I'm still trying to get my head around your proximity and related SSS 
>> macros. Applying the AO image works nicely for some scenes, but your 
>> proximity macro is a global solution, and not just a camera specific one.
> 
> But you can't use the proximity macro for mesh objects :( If POV 
> supported calling inside() for mesh objects, it would work. I wish 
> somebody would figure out a way to get inside() working with meshes!

Can't you just do multiple traces to some point outside the object's 
bounding box? If you had to shoot an odd number of rays, then the 
location is inside, otherwise it's outside. Exactly the same as POV does 
  for meshes internally?

> Sam

Cheers,
Edouard.

Post a reply to this message