Originally Posted by Clarence
Looks great. Slight difference in the white paint in the lower left corner... you've got a blue push or Mike's picture has a red push. The true reference is probably in the middle.
Slight washout or clipping in his collar and the shadow under his collar... slightly overexposed?
You've captured more texture detail in the goosebumps/pores/whiskers in his neck.
Interesting how your clouds look almost transparent... I'd have to see the true reference frame to know whch is more accurate.
But for the most part, a very comparable screenshot.
Bond's shades are crisper on the BenQ, esp. at the white straight line of reflection of what's under them on the lower edge of the visible arm of the glasses. The reflection of Bond's skin thru them, in the upper part of his right eye, also shows more delineation. As does the shadow area under the visible part of Bond's face, just below the chinline.
The upper photo shows Bond's head as I think it really is, while the lower photo shows it somewhat squashed - a circle would be not a circle, if shown on the lower screen. Same for the scoop under the prop on the plane's engine. It's more oval on the upper picture, more round on the lower picture. Mike, you might want to check your vertical height - the h to w ratio is different between your pic and his, and the BenQ's is most likely more accurate, being fixed pixel. My eye tells me his is more accurate than yours, even without a circle to refer to.
A reference circle is hard to come by in HD, on pattern generators. The algorithms it takes for displaying them accurately are thru the roof according to Greg Rogers, creator of the Accupel, which is why his HDG 2000 and 3000 do not contain them, even tho the excellent patterns they do contain allow for stellar calibrations. As a calibrator, up until HD discs you had to spend serious $ to have access to circles via pattern generators, to use for H vs. V sizing and linearities in HD.
Now, with HD discs, you can find circles in movies to use. Or with a DVR, capture either ABC's or CBS's logos, to use.
Perfect circles in your HD video material allow for perfect sizing of horizontal width vs. vertical height, which have to be in perfect proportion to each other for a circle to look like a circle.
When a circle looks like a circle, only then you can trust that everything else in your video material will sustain the maximum amount of suspension of disbelief possible, in the geometry of your image structure. That's why the better patterns contain not just a master circle, but also a circle in each corner.
Be aware also that "blue push" has nothing to do with grayscale, and thus cannot accurately be used to talk about white, which is in the grayscale domain of black and white video material. It's tempting to use, and SOUNDS like it should be an accurate description of what's happening.
But the word "push" only kicks in when colors are present and are thus being decoded. As in "red push": where whenever there is color on the screen being discriminated out, there's significantly more red intensity being shown than there is green or blue color intensity, on each color involved.
Red push is the same as blue/green diminish. With a red pushed scenario, when fleshtones are being prioritized - which is the ONLY way to fly, in our endless quest for suspension of disbelief - if color is being decoded and you don't have the critical color decoding registers in your scaler, the blue isolation/filter test becomes invalid and untrustable anymore, and completely unusable. What's the point of having blue be perfect when red is overdone, making fleshtones look ruddy and sunburned, ruining your suspension of disbelief? In a red pushed scenario, the entire color intensity level has to come down till you're at accuracy in fleshtones. All the other colorations have to take a back seat.
As a result, after reworking color and possibly tint as well, the fleshtones finally come in at accurate and perhaps even breathtaking - but with red push the vivid blues and greens of a peacock feather in the same scene become wilted, and limp. Not vivid, like they are supposed to be.
It's too late once you're at the pj, in pure RGB. If present, relinearization of it has to be accomplished somewhere along the way. Luckily enough, the better scalers don't mess with the color decoding like the manufacturers of RPTVs do, and red push is not a factor. In HD passthru on a scaler, again you get whatever the mix was that the HD telecine guys designed in, which is usually completely accurate.
What you are talking about on the white of the plane is there, I can see it and I agree - but the more accurate terminology in calibration terms is that the white you are referring to is more bluish on one and more reddish on the other.
I liked using the term "push" for that purpose also until my eyes were opened by Paul Carleton, creator of Home Theater SPot.com. He used to be a video cameraman, and they would realign the color linearity of their cams each day, before each shoot for the day.
While in Seattle doing a Mit 73" back when they had the critical "push" registers in their sm, using the AVIA test patterns he showed me the coloration errors in the factory setup and demo'd how to realign them, which can be done in HD with the proper patterns as well. AVIA was the best resource for such things back then, still is, at least on regular DVD.
He used a scene from Shakespeare in Love to demo the before and after - the scene where Kit is talking with Shakespeare in the basement bar just after the announcement of the new play, the closeup on Kit's face and outfit - and my jaw just dropped when he showed me the difference between red pushed color rendition and completely linear color rendition. I had had no idea, up till that point, what "red push" was all about, and naively thought that it could be cured via color and tint, or grayscale.
NOT. The difference between accurate colorations and red pushed colorations was absolutely staggering. The doubletake left a crick in my neck...
I was a whiz at grayscale and image structure by then, and thought I knew it all! But this has nothing to do with image structure and is AFTER the grayscale has already been taken care of, to D6500K accuracy.
It was a new curve I was being thrown, and decided I'd better master it on the spot - or should I say, SPot...