ouch! I could have sworn it was long cable runs...
The only other thing that could cause this apart from cascading is just a device that does not have an input/output impedance of close to 75ohms (could be a capacitor is faulty adding impedance). Or a device has some kind of AGC (automatic gain control - this includes our DVR - that doesn't like your signals).
Either way the brighter the video image the hotter the white level. This equates to more current flowing (all be it very small signal levels) which in a poorly matched signal path could push the H-sync pulse's ground up above 0v (thereby creating a more difficult to detect pulse - hence the tearing). An AGC circuit fed with overly hot video (super white) could attempt to reduce the levels on the fly thereby reducing the level of the sync pulse on those lines that contain hot levels.
Please excuse the amount of tech detail here - the reason I am up to date on video signal levels and video in general is because I am working on a video signal generator for a future product.
I have sync/black/white/63.555uS/64uS/525/625/odd fields/even fields rattling around in the brain cell at the moment and have learned a great deal about the topic literally in the last few weeks.
Please do keep me posted on how you fair Ian