Video Generation (non-NTSC)

[uu10.psi.com!presrch!CCHOST!SANDFORM (MARK SANDFORD)]

In Jeff Fox's message he indicated that PAL would require a 1/4
color period as opposed to the 1/2 period skip in NTSC.  I have done
quite a bit of work with video cameras and displays in NTSC and have 
found that in most cases, monitors support a wide range of input signals,
even those that deviate significantily from the "STANDARD".  I have 
found that many devices take various sort cuts to simplify thier video
generation in violation of the specs, but most monitors have internal
compensation to sort things out.  One project that I work on has a
color NTSC output that looked pretty good but had an annoying flicker,
so we rented a video vectorscope to test the video.  The vectorscope's
tolerances were such that it said the signal was not NTSC and refused to 
read it.  The purpose of the vectorscope is to meassure the video 
characteristics so it should have a fairily wide range outside the specs
or atleast one would assume, so that large problems could be meassured 
and
fixed.  After a fair amount of work by one of our engineers, we got
the vectorscope to recognize the signal and take measurements.  We
found that the engineer who designed the device had taken numerious
short cuts and violated many on the rules, but for the most part
there was no visual difference between the correct and short-cut video.
Only a couple of changes seemed to affect the visual quality.

Since we had the equiptment in house for a month I brought in my MuP21
development board and tested the video.  The results were much better 
than
our product.  The signal was recognised immediately and all video 
measurements looked good although I seemed to be getting a fair amount of
power supply noise.  (I was powering the board from one of those litle
wall socket battery eliminators).  I have the print out of the video 
signal
for the color bars if anyone is interested.  

I guess, I got a little long winded, but my point is that most monitors
will tolerate significant variations in video signals and adapt as best
they can within realistic limiations.

Mark Sandford
msandford@delphi.com
sandform@presearch.com