[old_tv_nut]

The problem was not solvable by modifying the receiver IF, only by modifying the Q channel filter in the encoder. If the Q channel bandwidth would be more sharply limited to 500 kHz, there would be no lower Q sidebands (symmetrical to the 4.5 MHz sound trap) to produce quadrature crosstalk.

The encoder Q filter circuit originally published by RCA had some phase correction for better transient response. It's possible that the coupling could also have introduced a sharper high frequency rolloff than the NTSC spec or even a partial null, but there is no mention of this in the literature. I have never seen service info for the early RCA encoder and have no idea what the alignment instructions were, if any.

You also have to realize that NTSC specs were predicated on a flat-top IF with sharp sound carrier trap. The phase/group delay characteristics of this standard IF were built into the FCC spec by a phase-compensating "Fredendall" filter at the transmitter. However, set manufacturers soon abandoned this IF design for a "haystack" response that had some of the standard phase response (especially near the sound trap), but was not identical. PAL specs did not include a phase compensation at the transmitter, expecting that phase distortions on color transients would be cancelled by the PAL technique.