I mentioned in p.o.t (Re: Need to put a picture here) that I was making 
an animation to reproduce something I read in a book many years ago. A 
woman was wearing a dress to a party. The material became transparent in 
random patches.

The latest WIP is here. (Partial nudity, if your eyes are quick enough)

http://youtu.be/_PpJHBg44rA

I am not happy with the pattern of transparency and after lots of 
experimentation this is the best I can come up with. I would appreciate 
it if anyone could point to a better method.

even incrementing it by 1/5000000000 per frame makes the output too 
jerky. I feel that I am taking the wrong approach.
The code follows.

#declare Dress_Map_Pos1 = 0.3  ;
#declare Dress_Map_Pos2 = 0.7  ;

#declare Turb_x = 2.5 ;
#declare Turb_y = 1 ;
#declare Turb_z = 2.5 ;

#declare Turb = < Turb_x,Turb_y,Turb_z>      ;

#declare Lambda_Var = 3 ;        ;

#declare Phase_Var =  0.000   + (frame_number/(5000000000)) ;


#declare JessiDress_1_=
material{
   texture {
     pigment {
       gradient < 0.000, 1.000, 0.000 >
       pigment_map {
   [Dress_Map_Pos1,
     uv_mapping
     image_map{
       jpeg "JessiSet2.jpg"
       interpolate 2
       map_type 0
       filter all 0.000
       transmit all 0.000
     }
   ]
   [Dress_Map_Pos1 +.01,
     colour rgbft <1.000,0.000,0.000,0.000,1.000>
   ]
   [Dress_Map_Pos2 -0.01,
     colour rgbft <1.000,0.000,0.000,0.000,1.000>
   ]
   [Dress_Map_Pos2,
     uv_mapping
     image_map{
       jpeg "JessiSet2.jpg"
       interpolate 2
       map_type 0
       filter all 0.000
       transmit all 0.000
     }
   ]
       }

       frequency 3.000
       phase Phase_Var
       scallop_wave
       noise_generator 3

       warp {
         turbulence <Turb_x,Turb_y,Turb_z>
         octaves 4
         lambda Lambda_Var
         omega 0.3
       }
       warp {
         turbulence <Turb_x,0,Turb_z>
         octaves 4
         lambda Lambda_Var
         omega 0.3
       }
     }
   }
}



-- 
Regards
     Stephen

Post a reply to this message

On 08/06/2012 12:38 PM, Stephen wrote:
> I am not happy with the pattern of transparency and after lots of
> experimentation this is the best I can come up with. I would appreciate
> it if anyone could point to a better method.

> even incrementing it by 1/5000000000 per frame makes the output too
> jerky. I feel that I am taking the wrong approach.

I changed the blend map pattern to marble, which gives a slightly better 
pattern to the transparency, I think.
The biggest difference to the animation was adding uv_mapping to the 


material{
   texture {
   uv_mapping
     pigment {

Comments?

  http://youtu.be/jvZfbKGbUDs


-- 
Regards
     Stephen

Post a reply to this message

> On 08/06/2012 12:38 PM, Stephen wrote:
>> I am not happy with the pattern of transparency and after lots of
>> experimentation this is the best I can come up with. I would appreciate
>> it if anyone could point to a better method.

>> even incrementing it by 1/5000000000 per frame makes the output too
>> jerky. I feel that I am taking the wrong approach.
>
> I changed the blend map pattern to marble, which gives a slightly better
> pattern to the transparency, I think.
> The biggest difference to the animation was adding uv_mapping to the

>
> material{
> texture {
> uv_mapping
> pigment {
>
> Comments?
>
> http://youtu.be/jvZfbKGbUDs
>
>

Looks beter, but somewhat lacking in realism.

Normaly, those transparency are at least somewhat dependent on the 
vewing angle. The light's incidence angle also have an effect.

You should probably explore something using the aoi (angle of incidence) 
pattern.
For that, you'll need to use version 3.7.


Alain

Post a reply to this message

Alain wrote:

> Normaly, those transparency are at least somewhat dependent on the 
> vewing angle. The light's incidence angle also have an effect.

Other things that could have an effect

- curvature of fabric
- stretching / acceleration of fabrik
- distance from skin
- ...

all not so easy to get as pattern, but if the pattern were not
random but related to the movement the effect would probably
look more natural.

Post a reply to this message

On 10/06/2012 5:15 PM, Alain wrote:
> Looks beter, but somewhat lacking in realism.
>

Thanks, but I don't know about realism. It is suposed to be a SF fantasy 
material.

> Normaly, those transparency are at least somewhat dependent on the
> vewing angle. The light's incidence angle also have an effect.
>
> You should probably explore something using the aoi (angle of incidence)
> pattern.

Definitely something to think about.

> For that, you'll need to use version 3.7.

And that can be arranged. :-)

-- 
Regards
     Stephen

Post a reply to this message

On 11/06/2012 2:30 AM, Christian Froeschlin wrote:
> Alain wrote:
>
>> Normaly, those transparency are at least somewhat dependent on the
>> vewing angle. The light's incidence angle also have an effect.
>
> Other things that could have an effect
>
> - curvature of fabric
> - stretching / acceleration of fabrik
> - distance from skin
> - ...
>

I agree with your points.
But I am just playing around just now and for simplicity will only 
consider the curvature of the fabric.

> all not so easy to get as pattern, but if the pattern were not
> random but related to the movement the effect would probably
> look more natural.

That would be changing the criteria. :-)
I will keep the randomness but try to make the transparency somewhat 
dependant on the viewing angle. My first thoughts are nested texture 
maps with the transparent lower ones using the aoi feature.
My tests with just aoi gave a too course pattern.

Thanks to both you and Alain.

-- 
Regards
     Stephen

Post a reply to this message

On 08/06/2012 12:38 PM, Stephen wrote:
> The latest WIP is here.

Probably the final version with an aoi section between the transparent 
and semi- transparent parts. I cannot tell if it is any better having 
watched it and other versions too much.

http://youtu.be/xSoFrFC51Lg



-- 
Regards
     Stephen

Post a reply to this message

Stephen wrote:

> I cannot tell if it is any better having 
> watched it and other versions too much.

Seems better to me ;)

Post a reply to this message

On 17/06/2012 11:36 PM, Christian Froeschlin wrote:
> Stephen wrote:
>
>> I cannot tell if it is any better having watched it and other versions
>> too much.
>
> Seems better to me ;)

Thanks, now to make the material in RL. ;-)

-- 
Regards
     Stephen

Post a reply to this message