From: poynton@poynton.com (Charles Poynton)
To: plishman@cs.tamu.edu (Philip L Lishman)
Subject: Re: Displays
References: <2d74k8$b49@TAMUTS.TAMU.EDU>

"Latched" pixels don't work very well when there's motion in the scene.

When an "object" moves across the screen, your eye tracks it according
to its average velocity -- say in degrees per second, or degrees per
frame, or pixels per frame. You can compute any of these according to
the scene content and the viewing arrangement.

Now, consider motion in the scene of ten pixels per frame, not an
unreasonable amount. Your eye's gaze point doesn't make discrete jumps
of ten pixels every frame time, it track smoothly. Your eye traverses
ten pixels during that time. If a pixel stays lit for a whole frame
time, that moving object will appear to blur ten pixels in the
direction of motion. [Of course the situation is much much worse if the
display has a pixel-on time of even longer than a frame, such as in a
passive-matrix LCD, but the situation I mention applies even to the
best case of active-matrix LCDs.]

In a CRT, each pixel is lit in a blinding flash of duration 50
microseconds or so. When there is motion in the scene, your
eye-tracking is such that the CRT flashes the image on your retina just
exactly as you have tracked to the right place and the right time.

You can think about this the other way around: instead of thinking of
your eye's gaze point traversing pixels, think of pixels traversing the
back of your retina.

This issue has not been a problem in the past, because computer
displays don't have motion, and television displays have such a
restrictive viewing angle that tracking is minimal. Cinema potentially
has the problem, and sometimes you see it, but there is a highly
sophisticated, knowledge-based, neural-network front-end processor that
compensates. He's called a cinematographer.

Regards,

C.

Charles Poynton
vox:   +1 416 486 3271
fax:   +1 416 486 3657
poynton@poynton.com [preferred, Mac Eudora, MIME, BinHqx]