[old_tv_nut]

Pritchard's 1977 paper is excellent, as all of his were. However, the measured results are for the case of a video feed (no R.F./I.F. problems) or the video output of a Conrac receiver that matches the original NTSC assumed IF response, rather than the then-current haystack response that the paper mentions. The results do generally correspond to the relative subjective impressions of the Super and Deluxe chassis. The chroma transients of the Super chassis are generally awful compared to any other RCA before or since.

In some other threads here, I have posted the investigations that I did at Zenith of the possibility of re-introducing I/Q demodulation in later sets to get rid of quadrature crosstalk. The investigation was prompted by RCA's re-introduction of I/Q in their solid state sets, which, however, puzzlingly, had low I-channel high-frequency response compared to the NTSC spec. What we found was that the loose NTSC spec on the Q channel lower-sdieband response meant that some installed encoders put too much lower Q sideband energy in the signal, which was not symmetrical to the 4.5 MHz sound notch that all receivers must have. The result was that some programs had acceptable transients and others did not. We surmised that RCA discovered this and deliberately reduced the I response to prevent visibility of unacceptable quadrature crosstalk depending on the signal source. The new use of I/Q decoding by RCA was short lived, and they went back to equiband decoding.