The grayscale as delivered measured extremely cool (10,000K) even in the LOW setting when Standard or Dynamic mode was selected. Theater mode LOW dropped this considerably closer to the D65 (6500K) standard, but tracking wasn't linear and the darkest part of the grayscale went distinctly greenish. Calibration eliminated the greenish tint but in order to optimize the critical dark part of the grayscale below 35 IRE, an excess of blue had to be allowed into the brighter part. Excess blue, especially measuring less than 7500K, is hard for the eye to detect—far harder than excess red or green. With only one set of color temperature controls in the factory service menu, perfect grayscale tracking was impossible, though the end result was definitely acceptable. In the color temperature chart, you'll notice very little difference between before and after calibration. That's because Kelvin measurements tell a lot about red and blue, but mostly ignore green in the white balance, and green was precisely what was in error.

The JVC displays "below black" information through all inputs. Using the pluge pattern on the Sencore VP-403 generator to set brightness (black level) seemed straightforward unless Dynamic Gamma was switched on. Switching Dynamic Gamma on, however, made things a bit more difficult as it seemed to demand that BRIGHTNESS be reduced to minimum, which was much too dark with real-world sources and pictures. But even with it off, having to reduce the brightness control 25 clicks (almost all the way down) from factory default seems unusual as well as unlikely, yet that's where it measures right when using the HDMI inputs. That alone might prevent many owners from ever seeing this set at its best. HD from DISH Network looked best with BRIGHTNESS set at least down to –20. Satellite/cable boxes aren't consistently "on-standard" but two other competing sets I had on hand looked correct with the DISH receiver (HDMI connection) using their brightness controls set at or near factory defaults. Oddly enough, the HD DVD player (also over HDMI @1080i) looked correct at the center (0) brightness setting. A Mitsubishi set on hand required little or no change in the brightness setting when going from the Sencore generator to the HD DVD player into the HDMI input. Strange. Component inputs were correct at +5 of the brightness control. [I've encountered this numerous times, and one of the best cheats I've found is using my DirecTV HD TiVo to record a high quality HD broadcast of a movie that I have on DVD. Since industry standard test discs are available on DVD I adjust an input correctly using the DVD player, and then compare a dark scene from the movie between the DVD player and HD TiVO inputs and adjust the image until it matches as closely as possible. –Ed].

The real surprise came when I noticed that the flashing 10 IRE/0 IRE bars of the PLUGE pattern caused not just the black level but even the bright staircase box in the center to go up and down in level along with it. Even more mysteriously, when the brightness control was dropped down so low that the flashing 10 IRE bar completely disappeared, the residual black level and the bright staircase box continued to go up and down in sync with the (now invisible) flashing bars. This too was a phenomenon particular to the HDMI inputs. Since this occurred even in Dynamic mode, it shouldn't have been due to the dynamic iris, which, according to JVC, is defeated in Dynamic mode.

Changing from a black screen to a faint 5 IRE window pattern had the same effect on residual black level in surrounding areas, actually increasing it 4 fold. This is something I might have expected Dynamic Gamma to do, but it was turned off. Theater mode suffered less of this than Standard and Dynamic, and the component inputs suffered from this shifting black level as well. This explains why the set could fade to a full-screen black that was significantly darker than black areas in dark scenes. It also partially explains the poor ANSI contrast ratio discussed below. I would really have wanted to attribute this to the dynamic iris opening prematurely but once again, it's not supposed to be working in Dynamic mode.

Maximum light output (Theater) on a white window varied from 78fL to 40fL depending upon the manual IRIS setting. Turning Smart Picture off didn't change the maximum light output from the small rectangle, but it did allow a full white screen to go much brighter (57fL, IRIS fully stopped down). This is the first time I've seen a display get considerably brighter with the whole screen lit compared to a smaller white box. Standard and Dynamic modes could even reach 100fL on a full white screen if you opened the iris up. This sort of dynamic range is unusual and likely undesirable. Keep Smart Picture ON and compensate for room lighting with the iris control.

Minimum black level (HDMI input, no video signal) varied from .01fL to .05fL depending upon IRIS setting. Peak contrast ratio measured 4,000:1 with the iris fully stopped down, though as noted above, dark scenes will almost never see black levels this low. ANSI (checkerboard) contrast was only 105. ANSI contrast ratio is always poorer in RPTVs than direct view flat-panel displays due to light scatter inside the cabinet and in the optics, though this set measured poorer than most. Fortunately, the eye is easily deceived by bright areas next to dark ones, making contrast appear better than it actually measures. Don't let the far superior ANSI contrast ratios of flat panel LCD and plasma sets lead you to believe they'll beat this set on dark scenes. They won't, and their contrast advantage on bright scenes will be subjectively subtle.

Both maximum light output and minimum black level are considerably lower than last year's set—a wise and welcome tradeoff. Darker blacks are much more desirable than excessively bright whites.

Color primaries measured the same as last year's set with red (.630/.337) and blue (.150/.055) close to HDTV color points and green (.291/.681) significantly oversaturated like most displays these days. Thankfully, this set showed little tendency toward screaming fluorescent-looking green athletic fields.

The demanding focus pattern was reproduced as sharply as I've seen (with excellent panel convergence, excellent geometry, and uniform focus across the screen) and the 1080i horizontal resolution pattern clearly showed all lines in the full 1920-line source. There was no softening, little rolloff in video frequency response, and no other artifacts so often seen on lesser displays. White balance uniformity across the screen, which has been a serious problem with Sony's SXRD sets, was very good.

JVC's scaler did pretty well on most of the Silicon Optix test screens, and I suspect its performance on these torture tests may be tops among current RPTVs. Jaggies were well controlled though the set didn't lock onto the 3/2 pulldown sequence. Further testing with Star Trek Insurrection indicated that 3/2 pulldown was indeed working with that disk at 480i. 3/2 pulldown provided by the set, which is essential for minimizing artifacts from film based material, wasn't much of an issue after progressive scan DVD players emerged since the DVD players were forced to do that function. Now, with HD DVD sending the sets interlaced 1080i, this important process is once again done by the set, not the player.