I was thinking about grime etc. and I had the following idea.

For any given spot on an object,
the amount of deposited grime that can be rubbed off,
is proportional to the amount of the same object that is visible,
from that point with a 180 degree FOV looking along the current normal,
with a bias controlled by the zmap of that view.

I guess the "dirt camera" resolution is equivalent to "number of samples", 
so you can speed up the calculations by lowering the accuracy.

For the ultimate grime you would not just test for "self" but for all 
objects within a given distance, that could influence the deposit, but this 
would be slower again.

Initial dirt deposits could be controlled with a dirt source that would 
work very much like a light source that is more or less diffuse. This would 
allow for everything from a wet spray look to a soft dust effect.
Obviously the concept of "up" would need to be factored in too.

I can see this in my head, clearly, I hope my words make as much sense to 
you.

:o)

Can anyone else see what I mean?

Post a reply to this message

CORRECTION: "is INVERSELY proportional to the amount of the same object 
that is visible" 

i.e in a hollow where things collect you can see the rim edge, but on a 
hill you can't, the deeper and more narrow you hollow the more of the sky 
is obscured by the surrounding walls. So on a hill top all the grime is 
removed as it has no shelter.

Daniel Matthews wrote:

> I was thinking about grime etc. and I had the following idea.
> 
> For any given spot on an object,
> the amount of deposited grime that can be rubbed off,
> is proportional to the amount of the same object that is visible,
> from that point with a 180 degree FOV looking along the current normal,
> with a bias controlled by the zmap of that view.
> 
> I guess the "dirt camera" resolution is equivalent to "number of samples",
> so you can speed up the calculations by lowering the accuracy.
> 
> For the ultimate grime you would not just test for "self" but for all
> objects within a given distance, that could influence the deposit, but
> this would be slower again.
> 
> Initial dirt deposits could be controlled with a dirt source that would
> work very much like a light source that is more or less diffuse. This
> would allow for everything from a wet spray look to a soft dust effect.
> Obviously the concept of "up" would need to be factored in too.
> 
> I can see this in my head, clearly, I hope my words make as much sense to
> you.
> 
> :o)
> 
> Can anyone else see what I mean?

Post a reply to this message

Daniel Matthews wrote:
>
> i.e in a hollow where things collect you can see the rim edge, but on a
> hill you can't, the deeper and more narrow you hollow the more of the sky
> is obscured by the surrounding walls. So on a hill top all the grime is
> removed as it has no shelter.

This is exactly what I did with the radiosity patina. When
recursion_level 1 is used the illumination on the surface
tells directly how much "outside" is visible from that point.

news://news.povray.org/3B8DD549.D0A9588E%40engineer.com
news://news.povray.org/3B8E0F26.6DDC17B3%40engineer.com


_____________
Kari Kivisalo

Post a reply to this message

I guess it is similar, but lighting dependent, you need a 2 pass render?.
How would you account for the initial deposition type e.g. fallout, using 
your methodology?

In another way it is different as it does not have a zmap bias to the 
effect as I have described it. Diffuse reflection is related yet again not 
exactly the same.

I don't think my idea can actually be done in POV yet, but perhaps the 
radiosity code is a starting point for the required functionality?

There a 4 parts,
the deposition value ( from the "dirt projectors" ),
the point's view,
and the weighting of the view values based on distance from the point.
There needs to be a directional bias value to simulate the direction of any 
"cleaning". This vector and it's calculation offers all sorts of new 
possibilities if it is not constant for the entire object.
i.e you could have dirt fall onto the object and then be rubbed of with a 
swirling like pattern or a back and forth motion.

This allows the calculation for the final grime mask,
dirt minus cleaning
With zbias being vital for filling small details on larger flat areas.
I guess all the calculations need to be post displacement etc. and should 
be compatible with a simple bump too.

This cleaning value could also be used separately to control such things as 
dent and scratch representations.

If implemented in an open enough form it could have many uses.

Kari Kivisalo wrote:

> Daniel Matthews wrote:
>>
>> i.e in a hollow where things collect you can see the rim edge, but on a
>> hill you can't, the deeper and more narrow you hollow the more of the sky
>> is obscured by the surrounding walls. So on a hill top all the grime is
>> removed as it has no shelter.
> 
> This is exactly what I did with the radiosity patina. When
> recursion_level 1 is used the illumination on the surface
> tells directly how much "outside" is visible from that point.
> 
> news://news.povray.org/3B8DD549.D0A9588E%40engineer.com
> news://news.povray.org/3B8E0F26.6DDC17B3%40engineer.com
> 
> 
> _____________
> Kari Kivisalo

Post a reply to this message

"Kari Kivisalo" <ray### [at] engineercom> wrote in message
news:3B9C4393.13C7BC7E@engineer.com...
> Daniel Matthews wrote:
> >
> > i.e in a hollow where things collect you can see the rim edge, but on a
> > hill you can't, the deeper and more narrow you hollow the more of the sky
> > is obscured by the surrounding walls. So on a hill top all the grime is
> > removed as it has no shelter.
>
> This is exactly what I did with the radiosity patina. When
> recursion_level 1 is used the illumination on the surface
> tells directly how much "outside" is visible from that point.
>
> news://news.povray.org/3B8DD549.D0A9588E%40engineer.com
> news://news.povray.org/3B8E0F26.6DDC17B3%40engineer.com
>

When you first posted these, I meant to ask something. Could you get more
localization by using regular reflection and setting max_trace level to just 1
or 2?

Batronyx ^"^

Post a reply to this message

Batronyx wrote:
>
> When you first posted these, I meant to ask something. Could you get more
> localization by using regular reflection and setting max_trace level to just 1
> or 2?

With blurred reflection maybe.

_____________
Kari Kivisalo

Post a reply to this message