Hi,

I was modelling a pop can using a couple bezier spline lathes and there is a
reflection that does not look right to me.  I am confident the segment
boundaries are straight but the reflection does not seem to match the curve.

Am I just seeing things?  If not, where is a good place for me to start looking
in the source (or is this a known issue?)

Thanks,
Mike

Post a reply to this message

Attachments:
Download 'bezier lathe problem.png' (490 KB)

Preview of image 'bezier lathe problem.png'

Am 18.08.2011 07:21, schrieb Solar Mike:

> I was modelling a pop can using a couple bezier spline lathes and there is a
> reflection that does not look right to me.  I am confident the segment
> boundaries are straight but the reflection does not seem to match the curve.
>
> Am I just seeing things?  If not, where is a good place for me to start looking
> in the source (or is this a known issue?)

At boundaries between flat and highly curved surfaces, reflections can 
play some pretty surprising tricks with our eyes, so it may well be 
accurate after all.

Try to put some long narrow cylinder(s) at that particular place on the 
ground to see things more clearly; I somehow suspect that what you think 
is a single contiguous line may actually be two orthogonal lines.

Post a reply to this message

On 08/18/2011 08:21 AM, Solar Mike wrote:
> I was modelling a pop can using a couple bezier spline lathes and there is a
> reflection that does not look right to me.  I am confident the segment
> boundaries are straight but the reflection does not seem to match the curve.

   If using 'sturm' does not solve the problem, you could try the 
MeshLathe include at lib.povray.org to see if it helps (if for nothing, 
then to corroborate that it should look as you expect).

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
> On 08/18/2011 08:21 AM, Solar Mike wrote:
> > I was modelling a pop can using a couple bezier spline lathes and there is a
> > reflection that does not look right to me.  I am confident the segment
> > boundaries are straight but the reflection does not seem to match the curve.
>
>    If using 'sturm' does not solve the problem, you could try the
> MeshLathe include at lib.povray.org to see if it helps (if for nothing,
> then to corroborate that it should look as you expect).

I did try adding 'sturm' and created a matching mesh and got the same results.
I'm still have not wrapped my head around why the reflections are like that but
I'm satisfied that the render is done correctly and I've simply fallen for a
"surprising trick" the reflection is making.

Thanks :)

Post a reply to this message

"Solar Mike" <wem### [at] yahooca> wrote:
> I was modelling a pop can using a couple bezier spline lathes and there is a
> reflection that does not look right to me.  I am confident the segment
> boundaries are straight but the reflection does not seem to match the curve.

Have you tried clipping the lathe (or intersecting it) with plane { -z, 0 } to
see if it looks straight from that point of view?

Post a reply to this message

"Cousin Ricky" <rickysttATyahooDOTcom> wrote:
> "Solar Mike" <wem### [at] yahooca> wrote:
> > I was modelling a pop can using a couple bezier spline lathes and there is a
> > reflection that does not look right to me.  I am confident the segment
> > boundaries are straight but the reflection does not seem to match the curve.
>
> Have you tried clipping the lathe (or intersecting it) with plane { -z, 0 } to
> see if it looks straight from that point of view?

I have done that and it seems fine to me.  I've attached an orthographic view
straight on from the side.  It's not perfect by any stretch but it is pretty
close.

Post a reply to this message

Attachments:
Download 'popcan ortho detail.png' (83 KB)

Preview of image 'popcan ortho detail.png'

"Solar Mike" <wem### [at] yahooca> wrote:
> "Cousin Ricky" <rickysttATyahooDOTcom> wrote:
> > "Solar Mike" <wem### [at] yahooca> wrote:
> > > I was modelling a pop can using a couple bezier spline lathes and there is a
> > > reflection that does not look right to me.  I am confident the segment
> > > boundaries are straight but the reflection does not seem to match the curve.
> >
> > Have you tried clipping the lathe (or intersecting it) with plane { -z, 0 } to
> > see if it looks straight from that point of view?
>
> I have done that and it seems fine to me.  I've attached an orthographic view
> straight on from the side.  It's not perfect by any stretch but it is pretty
> close.

I noticed that at the top your can has a dark band at the bend (and a less
visible lighter one in the middle) although logically I would say there is no
reason for that color shift. Maybe that has the same background as your
reflection problem?

Joost

Post a reply to this message