"Real" is actually a loose term.... it just means meaningful data from your capture card, and not noise or interpolated data. And I totally agree that digital pixel to analogue comparison can be tricky (thats why I had the fudge factor 105%). But even so when digitizing an analogue signal you do reach a point when you're capturing noise rather than any meaningful data. This is heavily dependant on how clean you analogue source is (type of LD and LD player in our case) and capture device.

//Snip from videohelp:
Laserdisc resolution is 528 X 576/480, but many titles in US, after 1990, are using the 544 X 480 resolution.
In Europe, the success of Laserdisc was minimal, so the few released PAL titles, continue to use the official resolution for PAL (528 X 576). In theory, there is a 544 X 576, but I never saw a Pal laserdisc using this resolution.
//End snip

So I think at the resolution of your raw cpature you are introducing some interpolated and noisy data from your analogue source. Now this is good because you always want to oversample for any A/D conversion because you can filter out the data you want. Now what I'm doing is like you said, is some noise and interpolated data reduction by downsampling. When I upsample back, I believe I have more contol than your capture card via algorithm choice and configurabilty.

Did this in anyway help clarify and answer your questions from above?

I'll be honest I'm not a video capture expert, but I do have a pretty good background in electrical engineering and deal daily with various analogue sources that we have digitize to take readings from. But any A/D conversion is pretty much the same, its just how you process and display the data. For video its a time-dependant, color coded 2D matrix. Look out nerdspeak!

BTW I still can't wait to see your Empire and Jedi...keep up the awesome work