Westinghouse H840CK15

[dtvmcdonald]

The big problem with I-Q demod is that to get it really right required two
delay lines, one for Y, one for I. And to get them right and the filters with good phase response is not easy ... without digital (or bucket brigades).

My personal opinion is that best is to use any two axes and get the response
quite flat to about 700-800 kHz then rolling off to zero at 1.5 MHz. on both
axes. Few folks notice the resulting bleedthrough from I into Q. Using
equal bandwidths the results are identical independent of the axes used.

However, other people may seriously object to the bleedthrough.

[old_tv_nut]

Originally Posted by dtvmcdonald

My personal opinion is that best is to use any two axes and get the response
quite flat to about 700-800 kHz then rolling off to zero at 1.5 MHz. on both
axes. Few folks notice the resulting bleedthrough from I into Q. Using
equal bandwidths the results are identical independent of the axes used.

However, other people may seriously object to the bleedthrough.

Yep. It was less disconcerting to have the color saturation of details reduced than to have the wrong hue - and if the level of quadrature distortion was low enough, at normal viewing distance in the home it appeared to add some color detail without being obviously wrong. With the engineer's nose to picture tube, the quadrature was visible as a rainbow effect on small strongly colored objects, but it also tended to be hidden by the luminance overshoot due to video peaking.

Since later solid state sets usually simplified the chroma bandpass such that they had this gradual chroma cutoff rather than the classic double-tuned response, it meant that they could actually benefit from composite video input that had double sideband chroma, meaning that the color detail would be somewhat better with a digital source (DVD or digital broadcast converter) fed into the baseband composite input than was possible with RF feed.

[benman94]

Originally Posted by dtvmcdonald

The big problem with I-Q demod is that to get it really right required two
delay lines, one for Y, one for I. And to get them right and the filters with good phase response is not easy ... without digital (or bucket brigades).

My personal opinion is that best is to use any two axes and get the response
quite flat to about 700-800 kHz then rolling off to zero at 1.5 MHz. on both
axes. Few folks notice the resulting bleedthrough from I into Q. Using
equal bandwidths the results are identical independent of the axes used.

However, other people may seriously object to the bleedthrough.

This raises an interesting question. What exactly is "wide bandwidth" chroma demod? Where does one draw the line in the sand and say this set is narrow band whereas this other set is wide band? I guess it is more accurate to say that some sets are wider or narrower relative to one another...

Take for instance the CTC-2 vs the CTC-5N. Both are "wideband" but the CTC-2 demodulates along I/Q and has extended I response. The CTC-N uses difference demod and starts rolling off at about 600 kHz on both axes, yet both are described as having wide chroma bandwidth on Ed's site. The CTC-2 has wider bandwidth, than the 5, but compared to the CTC-4 they both would qualify as "wide".

[miniman82]

The actual terms are next to meaningless, but I recall someone telling me it’s actually more correct in most cases to call narrowband circuits ‘equiband’ since most difference circuits have about the same frequency response in both axis. Maybe it was wayne?

[old_tv_nut]

Originally Posted by miniman82

The actual terms are next to meaningless, but I recall someone telling me it’s actually more correct in most cases to call narrowband circuits ‘equiband’ since most difference circuits have about the same frequency response in both axis. Maybe it was wayne?

Yes, the major difference is non-equiband (I/Q) vs. equiband; a secondary difference is how wide the equiband demods are. Generally, if the demods are "high level" (have high gain and are coupled to the CRT without further amplification) they will have more roll-off due to stray capacitance than a "low level" demod that is followed by an additional stage of amplification (like the RCA X/Z system). The X/Z system was truly a well-optimized trade-off of cost and performance, which was not improved on for many years.