"Anders K." <and### [at] f2scom> wrote in message
news:3b92f511@news.povray.org...
>
> There appears to be a major bug in POV-Ray's SOR implementation. Here's
some
> example code.
>
> camera { location <10, 15, 15> look_at <0, 1, 0> angle 15 }
> light_source { <2, 15, 5> color rgb <1, 1, 1> }
> plane { +y, 0 pigment { rgb <1, 1, 1> } }
> sor { 4, <2, -1> <2, 0> <2, 2> <2, 3> pigment { checker } }
>
> If the camera is at <10, 10, 15> instead of <10, 15, 15>, the scene
renders
> fine (sorbug1.png, attached). But when the camera is at <10, 15, 15>, the
> SOR has a hole in it (sorbug2.png)! And the SOR is certainly not open,
since
> adding "open" to the SOR results in a totally different picture
> (sorbug3.png).

Hi Anders,

Well, I believe what you are experiencing there is that the first and last
points are only controls.  From the Doc:

"The first and last points are only used to determine the slopes of the
function at the start and end point. They do not actually lie on the curve.
The function used for the SOR object is similar to the splines used for the
lathe object."

I rendered your example and if the first and last are duplicated (putting
first before first and last after last, using 6 points) it closes the sor,
although looks a different size (or shape).

So I'd have to say this is not a bug, seen this before too.

Bob H.

Post a reply to this message

Please don't post binaries in this group. Thanks.

Andy Cocker

(I always wanted to say that! :-))

"Anders K." <and### [at] f2scom> wrote in message
news:3b92f511@news.povray.org...
> [I tried to post this in povray.bugreports but it didn't show up; then I
saw
> a note about posting to .general before .bugreports, so I'm doing that
now.
> I apologize for not reading directions before. Since then, I've corrected
a
> minor error in my sample code that was unrelated to the bug. I also found
a
> workaround.]
>
> There appears to be a major bug in POV-Ray's SOR implementation. Here's
some
> example code.
>
> camera { location <10, 15, 15> look_at <0, 1, 0> angle 15 }
> light_source { <2, 15, 5> color rgb <1, 1, 1> }
> plane { +y, 0 pigment { rgb <1, 1, 1> } }
> sor { 4, <2, -1> <2, 0> <2, 2> <2, 3> pigment { checker } }
>
> If the camera is at <10, 10, 15> instead of <10, 15, 15>, the scene
renders
> fine (sorbug1.png, attached). But when the camera is at <10, 15, 15>, the
> SOR has a hole in it (sorbug2.png)! And the SOR is certainly not open,
since
> adding "open" to the SOR results in a totally different picture
> (sorbug3.png).
>
> I found a workaround that seems to render fine (sorbug4.png):
>
> camera { location <10, 15, 15> look_at <0, 1, 0> angle 15 }
> light_source { <2, 15, 5> color rgb <1, 1, 1> }
> plane { +y, 0 pigment { rgb <1, 1, 1> } }
> difference {
>     sor { 4, <2, -1> <2, 0 - .000001> <2, 2 + .000001> <2, 3> }
>     plane { -y, -2 }
>     plane { +y, 0 }
>     pigment { checker }
> }
>
> Still, this is very strange. Is this a known problem, and is there an
easier
> way to fix it?
>
> Anders K.
> and### [at] f2scom
>
>
>
>

Post a reply to this message

> Well, I believe what you are experiencing there is that the first and last
> points are only controls.

No, I know that the first and last points are controls. What I'm creating is
a cylinder of radius 2 and height 2 resting on the plane. Of course I could
do that with a cylinder, but I'm trying to demonstrate a bug in SOR. I'm
having the same problem trying to model a set of chess pieces; certain
camera angles are giving me the same problem.

The trouble is that the endcaps look wrong with the camera angle I showed.
The documentation says the SOR is closed by default, and it looks closed
when the camera is at <10, 10, 15>. But when the camera is at <10, 15, 15>
as in my code, it looks like the SOR is colored as if it were closed, but
rays are tested for intersection as if the SOR was open. (Look at
sorbug2.png.)

Anders

Post a reply to this message

> Please don't post binaries in this group. Thanks.

Sorry; as you can see, I'm new here.

Anders

Post a reply to this message

"Anders K." <and### [at] f2scom> wrote in message
news:3b9396e8@news.povray.org...
> > Well, I believe what you are experiencing there is that the first and
last
> > points are only controls.
>
> No, I know that the first and last points are controls. What I'm creating
is
> a cylinder of radius 2 and height 2 resting on the plane. Of course I
could
> do that with a cylinder, but I'm trying to demonstrate a bug in SOR. I'm
> having the same problem trying to model a set of chess pieces; certain
> camera angles are giving me the same problem.
>
> The trouble is that the endcaps look wrong with the camera angle I showed.
> The documentation says the SOR is closed by default, and it looks closed
> when the camera is at <10, 10, 15>. But when the camera is at <10, 15, 15>
> as in my code, it looks like the SOR is colored as if it were closed, but
> rays are tested for intersection as if the SOR was open. (Look at
> sorbug2.png.)

I'll try and see what the experts and v3.5 pre-beta testers think about
this.  It still occurs in v3.5 so nothing has changed from v3.1g.
The following does the same thing and I expected it might close the sor yet
did not.

camera { location <10, 15, 15> look_at <0, 1, 0> angle 15 }
light_source { <2, 15, 5> color rgb 1 }
plane { +y, 0 pigment { rgb 1 } }

sor
 {
         4,
         <2, 0>, <2, 0>, <2, 2>, <2, 2>
 // open
         pigment { checker }
}

Bob H.

Post a reply to this message