Projector Reviews

I'd like to add two comments before discussing the measurements. Shane mentioned how quiet the HC5000 is. I not only second that, but found it to be the quietest projector I've ever had in my home theater—quieter even than the Sony "Pearl." I saw a fairly lodge dust blob shadow, visible only on the very darkest scenes, just to the right of center-screen.

The accuracy of the primary and secondary color points of the Mitsubishi, while not exceptional, ranged from fair (green, magenta, and yellow) to very good (red, blue, and cyan). Instead of the usual before/after curves in the accompanying chart, we've plotted two curves for the post calibration results: one in which the test patterns filled the screen, and the other in which the white patterns were windows—that is, the white areas are in small boxes in the center of the screen surrounded by black. The readings were taken in this box or window, but the average picture level on the screen (the level to which the Auto Iris responds) was always a combination of the white and black area—and thus lower than the level in the window or the level in the equivalent full screen white pattern.

In both sets of measurements shown in the chart, the iris was in Auto Iris3, the Gamma was in Cinema mode, and the lamp was on Low. As you can see, these measurements, taken with our Photo Research PR-650 spectroradiometer, confirm the results discussed in the main measurement section, above.

The calibration was performed with full-screen patterns, which is why that result is the most accurate. In fact, the x/y coordinates of the white point at each step from 30IRE to 100IRE varied by no more than 0.003. With the window patterns, there is a significant increase in the color temperature at low brightness levels. That will affect the accuracy of the projector's color temperature in dark scenes.

One possible reason for the full/window measurement discrepancy is that with the auto iris in operation, the actual light output at low levels with a window pattern drops far faster than with the same input level, full screen. That is a result of the way an Auto Iris responds: to the average picture level. A 20 IRE full screen pattern produced a light output of 0.23 foot-Lamberts in my test, but a 20 IRE window pattern was too dim for the Photo Research to read. A 30 IRE full screen measured 0.86fL, whereas a 30 IRE window measured 0.46fL. If the HC5000 could maintain its color temperature down to levels well below 20 IRE, it's likely that the differences between the full and window results would be less significant.

Before calibration, the Middle setting of the Color Temp control was the most accurate, but still off from the desired D6500. The pre-calibration results for all three factory color temperature settings are shown below. The low brightness readings were taken at 30 IRE, the High at 80 IRE. All the readings are in Kelvins:

You can see the same pattern here as in the calibrated result: the numbers are fairly consistent when comparing the full vs. window results at high brightness, but the results vary significantly at low brightness levels. It is a little puzzling that the discrepancy between the full and window readings increases with higher color temperature settings, ranging from 1321K in Warm to 4986K in Cool. With the iris Open, these full/window discrepancies dropped to a maximum of less than 150K. When the projector was calibrated using the full screen white field patterns with the iris on Auto Iris3, and the iris was then switched to Open, the full screen color temperature changes were not visually significant. So if you get a Mitsubishi calibrated with the iris in one of the Auto settings and later decide you'd rather use the Open setting, you shouldn't have to arrange for a new calibration. The latter is fortunate, given the visible pumping of the Auto Iris in any of its settings.

But switching from Auto to Open Iris carries severe peak contrast ratio penalties. In Auto3, I measured a peak contrast ratio of 4185:1 (16.74fL peak white, 0.004fL video black, on my 78-inch wide, 16:9, Stewart Studiotek 130 screen). With the iris on Open, this dropped to a mediocre 886:1 (16.85 peak white/0.019 video black).