On Thu, 2 Dec 1999, Stephen J Baker wrote:
> On Thu, 2 Dec 1999, Doug Rabson wrote:
>
> > > Per-pixel lighting ("per fragment lighting" in OpenGL parlance) is the
> > > solution - but hardware rendering engines don't do that (yet) because it's
> > > a LOT more math. Phong shading and per-pixel lighting are really just the
> > > same thing.
> >
> > Phong shading can still have mach banding. To get rid of this, just enable
> > dithering.
>
> Correctly implemented Phong shading will only mach band where there would be
> a mach band in the real world.
>
> Dithering doesn't help - this is nothing to do with the number of bits per
> pixel or precision or anything like that. Even with an infinite number
> of bits per pixel and double precision floats throughout the math, that
> sphere would STILL exhibit those exact same mach bands.
>
> It's an artifact of your eye/brain and not of the graphics card.
>
> > > If you didn't see this with your old engine, it was either doing something
> > > fancier than gouraud - or you simply didn't look at test cases that were
> > > identical to this one.
> > >
> > > ALL Gouraud shaders are doomed to producing mach bands.
> >
> > Using dithering reduces the problem significantly as does higher color
> > resolution (e.g. 24bpp).
>
> That's simply not true. This is a well documented perceptual effect.
> The things you describe can also cause visual artifacts like contouring,
> but have nothing to do with the mach bands in that sphere image.
>
> Your brain uses shading to determine the shapes of things. When the
> rate of change of shading changes, that's something that in the
> real world would imply a sudden change in surface geometry. Your
> brain needs to be aware of that change - and as a result, it's
> 'amplified' in your awareness.
>
> Your eyes are NOT video cameras...the more you get into 3D graphics,
> the more you come to realise that fact. Even if the display had
> infinite precision and infinite resolution, there would still be
> mach bands in a Gouraud shaded image.
>
> Check out this interactive toy from the David Anson at Cornell:
>
> http://www.tc.cornell.edu/Visualization/contrib/cs490-96to97/anson/MachBandingApplet/
Sorry, I was thinking of a completely different effect.
--
Doug Rabson Mail: [EMAIL PROTECTED]
Technical Director, Qube Software Ltd. Phone: +44 171 289 4201
_______________________________________________
Mesa-dev maillist - [EMAIL PROTECTED]
http://lists.mesa3d.org/mailman/listinfo/mesa-dev
