Anyway, I made some captures to try and help me decide whether to use the composite or s-video output from my LD player.
Frame A: composite
Frame A: s-video
OK so these shots are way too dark; this is something I'll have to fix in the capture card driver settings. But look at the two pics; the difference is only subtle but the composite one is slightly sharper. It's also got more noise on it; but a decent AVISynth denoiser should be able to filter this without losing too much of the detail.
So it looks like composite is better? Well, take a look at this shot:
Frame B: composite
In the magnified area you can see awful combing artefacts. This is much better on the s-video capture:
Frame B: s-video
I'm quite pleased with the level of detail this shot is showing, and I would be interested to see the same frame from an NTSC capture, to see if it demonstates the advantage of using the PAL resolution.
Just for the record, this is what it looks like with th HQ noise reduction circuit turned on:
Frame B: s-video HQ on
Bloody awful!