[vtk-developers] Re: [vtkusers] Modifications to vtkOpenGLPolyDataMapper.cxx vtkOpenGLProperty.cxx

[John Biddiscombe]

Goodwin, + CC'ing this to the developers list.

Yes. The pixel differences for wireframe are clear if you have a look at the
test images and the new ones just generated with your code. In the
regression test images, backface wireframes appear black, in yours, you can
see the shading coming through from behind. I think it looks better. The
graphics/capcow.tcl shows it most clearly and is one of the worse
(wireframe) failures.

>>
Just thought... the pixel differences would be caused
by point no. 1 below; possibly the facet problem too- backfaces in surface
representation aren't rendered black by default too.
<<

This would be my first guess, but I can't see why (for example in
graphics/eleState.tcl which fails the test too) some of the polygons would
be back facing. The polygons are read from a file of state outlines and then
extruded and capped and rendered. For some to be black on the edges would
require a very odd polygon at source. (Mind you I do remember some failing
when I was playing with seidel triangulation code because they were
clockwise oriented instead of anticlockwise - I'll check)

If anyone has any thoughts...

JB
[I'll try to post link to dashboard results tomorrow.]


----- Original Message -----
From: "Goodwin Lawlor" <goodwin.lawlor at ucd.ie>
To: "John Biddiscombe" <jbiddiscombe at skippingmouse.co.uk>
Sent: Wednesday, October 30, 2002 12:41 AM
Subject: Re: [vtkusers] Modifications to vtkOpenGLPolyDataMapper.cxx
vtkOpenGLProperty.cxx


> Hi John,
>
> Thanks for having a look at this stuff. I didn't thoroughly go through
> vtkOpenGLPolyDataMapper.cxx- I just got enough of an idea to get what I
> wanted working. I thought it might be of interest though. I wrote a
> vtkEPSWriter class to get some vector graphics for a thesis I should be
> writing up. I wanted wireframe backface culling enabled for the writer...
> thesis-writing procrastination in the extreme!
>
> I could be wrong but the pixel differences with wireframes could be
because
> vtk renders wireframes flat, where as the new mapper shades wireframes.
>
> I've no idea why some facets would appear black.
>
> The only differences I've noticed is that
>
> 1. wireframe backfaces default to the same property as the frontfaces
unless
> you explicitly set the BackfaceProperty. At the moment, the wireframe
> backface color defaults to black.
>
> 2. you can pick a point on a wireframe "surface"- at the moment, you have
to
> pick an edge to pick a point (with vtkCellPicker).
>
> I've read the present mapper's method of drawing the primitives (rather
than
> using glPolygonMode) could produce faster rendering- I didn't see this
using
> Sebastien's bench-sphere.tcl script. Rendering strips was (very)
marginally
> slower and everything else was faster. Probably should be tested on
> different graphics boards though.
>
> Thanks!
>
> Goodwin
> ----- Original Message -----
> From: "John Biddiscombe" <jbiddiscombe at skippingmouse.co.uk>
> To: "KenLists" <kenlists at nycap.rr.com>; "'Goodwin Lawlor'"
> <goodwin.lawlor at ucd.ie>; "'VTK List'" <vtkusers at public.kitware.com>
> Sent: Tuesday, October 29, 2002 11:08 PM
> Subject: Re: [vtkusers] Modifications to vtkOpenGLPolyDataMapper.cxx
> vtkOpenGLProperty.cxx
>
>
> > Ken, Goodwin, etc
> >
> > >Please make sure the tests all pass
> >
> > I'm getting 12 new fails due to the changes. I submitted an experimental
> > dashboard, but it doesn't seem to be showing up, so I'll have another go
> > tomorrow. The wireframe changes cause some pixel differences in several
> > images, but more seriously there seems to be a disruption of some facets
> in
> > a handful of others. Some facets appear dark or black, can't quite tell
> > what's going on.
> >
> > Will stave off a commit for now.
> >
> > JB
> >
> > _______________________________________________
> > This is the private VTK discussion list.
> > Please keep messages on-topic. Check the FAQ at:
> <http://public.kitware.com/cgi-bin/vtkfaq>
> > Follow this link to subscribe/unsubscribe:
> > http://public.kitware.com/mailman/listinfo/vtkusers
>