[rec.video] Kell factor, part II

dya@unccvax.UUCP (York David Anthony @ W06BF-TV, Concord, NC) (09/14/88)

	Many people have written to ask about the so-called "Kell" factor
particularly about the statement that a "ccd array has a K-factor of 1.0."
This is my response (I thought it would be of general interest):

	The Kell factor for electron-scanned pickup tubes and CRTs comes
from the fact that there is a degradation due to the energy distribution
of the scanning spot.  The ideal scanning spot would cause a modulation
transfer function (MTF) which was unity up to the Nyquist limit for
the number of theoretical pixels or lines in use, and zero thereafter.
However, the scanning spot has a Gaussian distribution. The Kell factor
attempts to account for the falling amplitude response with an increase
in spatial frequency, and is actually pretty poor at characterising
video systems. Nevertheless, the Kell factor is a good first approximation
in visual system response to square wave gratings and television systems.

	In vertical resolution, the Kell factor also is a "fudge factor"
designed to describe the degradation due to vertical interlacing. In the
vacuum tube pickup tube, lateral leakage between frames (where there are
rows of undriven-to-cathode-potential regions due to interlace) causes
a loss of vertical spatial resolution. Aliasing components with the
interlace structure degrade the apparent vertical resolution even
more.

	What I meant about the CCD array having a Kell factor nearly 1.0
is that each individual sensing element could be (and generally is) 
rectangular, and thus, a sine wave grating at or near the Nyquist limit
for the CCD array in use would have an amplitude response of 1.0 because
the scanning "spot" is synthetic and is shaped damn nearly perfectly.
Networks outside the pickup array can provide the necessary shaping
to ensure perfect sin(x)/(x) shaping with alternate zeroes on adjacent
lines, thus giving perfect interlace. Horizontally, the perfect sampling
is provided by the electrical and optical isolation of adjacent elements.

	In this regard, the CCD array is a perfect pickup device.

	The interested student should probably read "Image Quality:
A Comparison of Photographic and Television Systems" by Doctor Otto
Schade.  This gives a much more precise and theoretically correct 
treatment of scanning spot energy distribution and its effect on
TV spatial resolution.  It was reprinted in the entirety in the June, 1987
SMPTE Journal; it was also published by RCA in 1975.

Hope this helps...

York David Anthony
DataSpan, Inc