Hi Castellen,
Thanks for the images, good reference! I've received the CIAs and replaced the suspected one and it's all good now!
I've done some testing on the dark image. The pot on R217 didn't help and the oscilloscope plot showed good white voltages of about 1V, but bad black levels, hence the darker tones are too dark. I'll restore the 470 ohm resistor, seems to be a value that's fine. I could get the brighter values more bright but the black level didn't come up.
At some point discovered that I have 74LS166 on U23 instead of 74F139. The latter wasn't on the BOM either. It's a wonder it gives a picture at all. When I hooked up the oscilloscope to pin 12 of U23 probably because of the 50Hz EMI around my workbench it's giving a much brighter but distorted image. Replacing it with the proper IC should solve this issue. Hopefully a 74HCT139 will work as the F version seems to be obsolete. It's about 10x slower (3-4ns vs 40-50ns) but for 50Hz/15KHz sync signals this should be fine.
Man what a stupid mistake, I should've realized sooner. Sorry for the inconvenience and thank you for your patient troubleshooting, much appreciated!