Someone mentioned using Fullscreen to help repair specific scenes in the 2004 SW dvd's. On the theory that the P&S region would often be where the eye would be drawn. I know there was some work done, but I haven't had the chance to dig it out and evaluate it, yet. Perhaps the P&S region's contribution could be de-emphasised on a gradient, to make it less obvious?

You mention an algorithm... at the risk of taking it on a tangent, which I will... The P&S & Widescreen, taken together, would contain information that falls "between" the pixels scanned for either. But perhaps of more benifit - I have wondered if an algorithm could be used to pull extra detail out of the combination of PAL and NTSC widescreen? (Or fullscreen tv shows, for that matter). Assuming, that is, that the PAL and NTSC versions were telecined independantly. It wouldn't work if the PAL was upconverted from the NTSC. On the other hand, it might work if the PAL and NTSC were each downconverted from a high-rez master. Might be a way to get approaching-Anamorphic-quality out of letterbox... Am I rambling? Thought so.