[Argus25]

It is interesting, the switching transients around the color bars, I'll have another look tonight and take a photo of the output from mine.

Of course partly what you see will depend on the scope's bandwidth. With the 2465B scope, running it on its 400MHz bandwidth, it always produces a less clean & less sharp looking waveform (wherever the waveform originates from) than switching on its 20MHz LPF, which makes any video waveform look cleaner regardless where it came from.

I think most likely the +/- power supplies to the analog part are ok as for the most part the amp circuits are differential design and the supply rejection is good, but it wouldn't hurt to try to add extra bypassing there.

More likely it is a common grounding issue of some of the logic and analog grounds and a common impedance there back to the supply earth.

Try adding a thick separate earth connection from the tracks in the analog area directly back to the supply's earth, bypassing the supply's connector.

Can you see any high frequency transients on the power supply outputs ?

(The caps on that supply often dry out too).

Can you also add a picture of the White waveform with the chroma switched off.

PS: I have also found many encapsulated DC/DC converters shocking for transients and noise, the "quiet" thing is mostly marketing hype. Also these spikes arise from a low Z source and are difficult to filter off without moderately big inductors and filter caps, which would have not fitted inside the small units!