Hi,

A few months ago I modeled a protein structure in DS Visualizer and exported
a POV Ray file. The render was beautiful. Now I have a publisher that wants
to buy the figure, but is asking for a different color scheme. I reopened
my Visualizer file, set up the new color scheme and exported a POV Ray
file. Now the render is all black. Not only that, but I can see that bits
of the structure here and there are missing. I have no idea what could be
wrong with this file. All I did was change the colors of the materials
applied to the protein.

Any suggestions on where I should begin trying to fix this?

Thanks

Greg Williams

Post a reply to this message

Greg Williams wrote:
> Any suggestions on where I should begin trying to fix this?

Can you take the original POV export and just change the textures used 
there?

...Chambers

Post a reply to this message

I have used some of those programs to construct molecules (and proteins) and
they export very simple textures. Only using pigment (without patterns) and
simple finish blocks. So it should be very easy to open the pov file and
check if everything is ok. If you dont find something wrong post the file
or part of it.

Post a reply to this message

I don't think so. I have,literally, several thousand spheres and cylinders
in the file.


Chambers <ben### [at] pacificwebguycom> wrote:
> Greg Williams wrote:
> > Any suggestions on where I should begin trying to fix this?
>
> Can you take the original POV export and just change the textures used
> there?
>
> ...Chambers

Post a reply to this message

Unfortunately, POV Ray is still pretty mysterious to me. I can render, and
not much else. I did find this message on the discussion board for the
modeling software:

"There have been changes to the representation of alpha values for
transparency in DS 1.7. The pov ray exporter was not updated to reflect
this change. The result is that now, the wrong transmit value (t instead of
1-t) is written in rgbt color specifications, so objects that should be
opaque are treated as transparent."

To be honest, I don't know how to implement this information.




"kike" <dry### [at] hotmailcom> wrote:
> I have used some of those programs to construct molecules (and proteins) and
> they export very simple textures. Only using pigment (without patterns) and
> simple finish blocks. So it should be very easy to open the pov file and
> check if everything is ok. If you dont find something wrong post the file
> or part of it.

Post a reply to this message

Here's the lead information from my file. The rest of it is specification
for several thousand spheres and triangles.


//
//  Accelrys ViewerPro POV-Ray file
//  Created:  Sat Jun 17 12:10:44 Pacific Standard Time 2006
//

camera {
orthographic
location < 0.0, 0.0, 1000 >
right -4/3*x
angle 30.000000
look_at < 0, 0, 0 >
}

light_source { <0,0,275> rgb<1.000000,1.000000,1.000000> parallel
point_at<0.0, 0.0, 0.0> }
light_source { <-275, 0,0> rgb<0.3,0.3,0.3> shadowless}
light_source { <275, 0,0> rgb<0.3,0.3,0.3> shadowless}

background { color rgb < 1.000000, 1.000000, 1.000000 > }

object {
#declare _WLVP_line_width = 0.02;
#declare _WLVP_dot_size   = 0.03;
#declare _WLVP_text_depth = 0.0;
#declare _WLVP_text_scale = 0.3;
#declare _WLVP_finish = finish { diffuse 0.45 ambient .35 specular .2
roughness .1 }
#macro triangle_texture (P1, C1, P2, C2, P3, C3)
   texture {
      #ifdef (triangle_base_texture)
         triangle_base_texture
      #end
      pigment {
         #if (C1.red=C2.red & C1.red=C3.red &
              C1.green=C2.green & C1.green=C3.green &
              C1.blue=C2.blue & C1.blue=C3.blue &
              C1.filter=C2.filter & C1.filter=C3.filter &
              C1.transmit=C2.transmit & C1.transmit=C3.transmit)
            C1
         #else
            #local SX = P2 - P1;
            #local SZ = P3 - P1;
            #local SY = vcross(SX, SZ); average
            pigment_map {
            [1 onion pigment_map {[0 C1] [.85 radial pigment_map {[0 C3] [1
C2]} frequency 6]}]
            [1 onion pigment_map {[0 C2] [.85 radial pigment_map {[0 C1] [1
C3]} frequency 6]} translate x]
            [1 onion pigment_map {[0 C3] [.85 radial pigment_map {[0 C2] [1
C1]} frequency 6]} translate <.5, 0, .8660254>]}
            matrix <1.01, 0, 0, 0, 1, 0, -.57735, 0, 1.1647, -.005, 0,
-.002>
            matrix <SX.x, SX.y, SX.z, SY.x, SY.y, SY.z, SZ.x, SZ.y, SZ.z,
P1.x, P1.y, P1.z>
         #end
      }
   }
#end
#macro colored_triangle (P1, C1, P2, C2, P3, C3)
   triangle {P1, P2, P3
   triangle_texture (P1, C1, P2, C2, P3, C3) finish {_WLVP_finish}}
#end
#macro colored_smooth_triangle (P1, N1, C1, P2, N2, C2, P3, N3, C3)
   smooth_triangle {P1, N1, P2, N2, P3, N3  triangle_texture (P1, C1, P2,
C2, P3, C3) finish {_WLVP_finish}}
#end

union {
sphere {
<0,0,0>, 0.500000
translate <193.328003,-13.325000,-130.389999>
pigment { rgbt <0.745098,0.921569,0.411765,1.000000> }
finish { _WLVP_finish }
}
sphere {
<0,0,0>, 0.500000
translate <191.098007,-12.048000,-127.431999>
pigment { rgbt <0.745098,0.921569,0.411765,1.000000> }
finish { _WLVP_finish }
}
sphere {







"kike" <dry### [at] hotmailcom> wrote:
> I have used some of those programs to construct molecules (and proteins) and
> they export very simple textures. Only using pigment (without patterns) and
> simple finish blocks. So it should be very easy to open the pov file and
> check if everything is ok. If you dont find something wrong post the file
> or part of it.

Post a reply to this message

Greg Williams wrote:
> union {
> sphere {
> <0,0,0>, 0.500000
> translate <193.328003,-13.325000,-130.389999>
> pigment { rgbt <0.745098,0.921569,0.411765,1.000000> }
> finish { _WLVP_finish }
> }
> sphere {
> <0,0,0>, 0.500000
> translate <191.098007,-12.048000,-127.431999>
> pigment { rgbt <0.745098,0.921569,0.411765,1.000000> }
> finish { _WLVP_finish }
> }
> sphere {

It looks like this is the part you're interested in.  Change the final 1 
to a 0, and it should be OK.  You could even do a search and replace to 
change every instance of "1.000000>" (sans quotes) to "0>", and things 
should be fine.

...Chambers

Post a reply to this message

> I don't think so. I have,literally, several thousand spheres and cylinders
> in the file.

You could record a macro(cntrl-shift-r) that searches (cntrl f) for "rgbt" 
and replaces it
with "fix_trans(" then searches for the next ">" and replaces it with
">)".... end macro, assign a hot-key... then you can hit that
hot-key a few thousand times (at least in the windows version)

And then add a macro near the beginning something like...
#macro fix_trans(A) rgbt <A.red,A.green,A.blue,1-A.t> #end

Post a reply to this message

This entire exchange reveals my gross inexperience with POV Ray. It's the
only command line software I use, and I'm not very good at it!

Here's what I figured out with all the help above: Select all the "objects"
in the file. Do a Replace All 1.000000 with 0 within the selection. The
only catch is to make sure none of the RGB coordinates in the modeling
program I use are set to 255. These become 1.000000 in POV Ray and get
reset to 0, too.

Many thanks to the folks who helped with this problem. It was a simple
solution, really. (I hate looking dumb!)

Greg





"Tim Attwood" <tim### [at] comcastnet> wrote:
> > I don't think so. I have,literally, several thousand spheres and cylinders
> > in the file.
>
> You could record a macro(cntrl-shift-r) that searches (cntrl f) for "rgbt"
> and replaces it
> with "fix_trans(" then searches for the next ">" and replaces it with
> ">)".... end macro, assign a hot-key... then you can hit that
> hot-key a few thousand times (at least in the windows version)
>
> And then add a macro near the beginning something like...
> #macro fix_trans(A) rgbt <A.red,A.green,A.blue,1-A.t> #end

Post a reply to this message

Greg Williams wrote:
> Here's what I figured out with all the help above: Select all the "objects"
> in the file. Do a Replace All 1.000000 with 0 within the selection. The
> only catch is to make sure none of the RGB coordinates in the modeling
> program I use are set to 255. These become 1.000000 in POV Ray and get
> reset to 0, too.

That's why I recommended doing a search & replace including the closing 
 >, so that you'll only get the trailing value (the transmit value).

As in, replace 1.000000>. with 0.000000>

(Note: if you use partially transparent objects anywhere, then you'll 
need to use macros to invert the transmit value instead of replacing it 
with 0, as 0 means no transparency).

Would you mind posting a sample image in p.b.i so we could all get a 
look at what you're working on?

...Chambers

Post a reply to this message