Thanks, madshi, for taking the time to explain things to me. The reason I ask is that I have been able to detect very little (if any) differences from one scaler to the next, including my HTPC, which theoretically has some of the best algorithms available, especially when ffdShow is used. To take your points one at a time:
- how sharp is the image? - At 120" screen size, I have found virtually NO difference unless the scaler is of particularly poor quality, like the one in the Optoma H-79. Every other scaler I have seen, including the ones in upscaling DVD players, is identical in this regard.
- how much mosquito noise is visible? I didn't know this was directly related to scaling and instead thought this was caused by MPEG encoding. I also thought it was directly controlled by noise reduction and was related to the amount of tradeoff one would be willing to accept - that is, the more noise reduction, the less mosquito noise but also the softer the image.
- how much EE is visible? Here is an area in which I have been seeking the Holy Grail, but so far I have not found it. In standard DVDs, I can't get rid of it no matter what I have tried, but I have admittedly not had the pleasure of seeing a Lumagen unit at work. If Lumagen has actually solved this problem and gotten rid of EE, then I will gladly purchase one of their products, but I have my doubts at this point, as NO SCALER I have ever seen, including the expensive Teranex VPs I've seen at shows, has ever accomplished this feat.
- are any jaggies or other artifacts visible? I thought those were deinterlacing artifacts, not scaling artifacts. The only such scaling artifact that I was aware of was "line twitter", something that again very seldom is a problem even with low cost solutions.
- overall subjective PQ impression - Well, yeah, I am always concerned about overall PQ, but I am trying to nail down what parts of the PQ are directly related to scaling and what parts are related to deinterlacing, optics quality (sharpness), and display technology (mosquito noise vs. low level dithering).
I'm not trying to give you a hard time , madshi, but rather to find out what specific attributes are directly related to scaling. My theory here is that a lot of people are mistaking some of the other problems and artifacts inherent in video processors and displays and simply blaming these problems on "scaling", when "scaling" as a single concept is actually very much the same from one unit to the other.
Now maybe the problem is that I just haven't seen the Lumagen scaling, and maybe they know something that the HTPC guys and Teranex don't know, so I will reserve final judgement until I get a chance to evaluate the Lumagen scaling for myself.
So one question still remains - Is there some specific test pattern or video clip that will clearly demonstrate the superiority of one scaler over the other? You know, some sort of grueling scaler torture test that one scaler will fail while another passes, much like the HQV disc does for other aspects of video reproduction? Right now it seems like people just look at image quality in general and then proclaim one scaler to be better than another, unless of course they are using the term "scaler" when they actually mean "video processor", as that term would incorporate many more parameters than simple scaling alone.