Test Bench: Mitsubishi WD-57831 57-inch DLP HDTV
Brightness (100-IRE window, before/after calibration): 184.1/85.0 ftL
Grayscale (color temperature, before/after calibration):
|IRE||Temp before||Temp after service menu|
Out of the box, with the Mitsubishi WD-57831's video mode set to Natural and color temperature set to Low, image brightness was a fairly hot 185 foot-lamberts, which would be fatiguing to many viewers in a room with subdued light. A basic disc calibration followed by subjective adjustments with program material brought this down to a more reasonable 85 ftL.
Despite the manual's statement that the Low color temperature setting adheres to the 6,500-kelvin industry standard for grayscale, measurement with a Photo Research Spectrascan colorimeter confirmed that it did so only at the dark 20-IRE window. As the image got brighter, the set tracked fairly consistently but slightly warmer (redder) than 6,500 K through the middle of the brightness range, then got even warmer on the two brightest windows. This is better than the best preset on many TVs and would be acceptable for many viewers but still falls short of the neutral gray tones on an accurately calibrated set. A service-menu calibration brought grayscale to within ±80 K for most windows, excepting the 90-IRE window, which remained 186 K off.
Color accuracy out of the box was off, but the PerfectColor and PerfecTint color-palette controls in the set's user menu allowed fine tuning of individual primary colors to get them in line. The accuracy of the red, green, and blue color primaries was about average and slightly improved with the PerfecTint control, though the controls lacked the range to achieve perfect alignment with the industry standard for color points. Color decoder error measured +12% red out of box, which was very noticeable in program material as exaggerated flesh tones, and so low on green as to be virtually unmeasurable, meaning more than 25% down. These errors were essentially zero'd out with the PerfectColor saturation controls, then further adjusted subjectively with a variety of program material. Blue decoding looked spot-on.
Picture uniformity on full-field gray test patterns at varying brightness levels was outstanding, and a welcome relief after looking at a series of bigscreen LCDs, all of which seem to suffer in this area to some degree. Hotspotting was virtually nonexistent, though the very darkest patterns (10 and 20 IRE) did exhibit some video noise (a not uncommon phenomenon). Picture geometry, assessed with crosshatch and linearity test patterns, was essentially perfect.
Overscan (on both the component and HDMI inputs) with HD signals measured 3% in Standard mode and 14.5% left/right, 3% top/bottom in Wide Expand mode (the only two format options for HD programs).
Multiburst resolution patterns delivered at 1080p, 1080i, and 720p from a generator as well as a Blu-ray Disc player were fully resolved, though the projection-screen "grain" common to all RPTVs was a bit more obvious on bright, uniform backgrounds than on some other models. Horizontal viewing angle was fairly good, allowing the viewer to be reasonably far off-center before a noticeable drop-off in picture output, but not so great in the vertical plane, where standing up or down at a distance of 8 to 10 feet could subtly affect image brightness.