Paul Brook wrote:
The difference is that while many systems have two independent image sources
and arrange for only one of them to be seen by each eye (either via fixed
polarisation, or shuttered time multiplexing), the iZ3D system takes a single
base image and uses a per-pixel filter to control how much of that pixel
intensity goes to each eye. You get full resolution, no grainyness or gaps,
full color. The only limitation is that you loose some of the peak intensity.
However in my experience modern LCDs have more than enough brightness to
spare, so that's easily compensated for under normal conditions.
You lose half the intensity range. The iZ3D has to be able to display
white at P = 0 (right only), P = 0.5, and P = 1. The middle value
divides between the left and right eyes, so you can't display more
than that to a single eye. Effectively 7 bit per component instead
of 8.
After working through Viktors nifty compositing function I realize
the iZ3D is better than I thought at first. I can't understand why
they aren't open sourcing/publishing the driver code though - don't
they want people writing software for it?
Transforming from separate components (as required by conventional stereo
systems and generated from the 3D rasterisation) to an intensity+filter angle
(required by the iZ3D screen) isn't conceptually hard. It's basically the
same as converting a complex number from cartesian to polar form.
Conceptually simple, yes. Suitable for implementation in fixed function
logic within the open graphics card, no. (Although maybe the composite
function could be included, as that's just byte adds and shifts.)
--
Hugh Fisher
DCS, ANU
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
