TEST BENCH

Primary Color Point Accuracy vs. SMPTE HD Standard

The Dream'E's Theater picture preset delivered the best contrast and black level, and the Medium color -temperature setting delivered the most accurate color reproduction. Grayscale tracking before calibration averaged 6,490 K from 30 to 100 IRE, deviating only 121 degrees at most from the 6,500 -K standard. This is about as good as color tracking can get; none of the instruments commonly used for calibrating displays is accurate enough to improve on this result. I did try, though. I shifted to the User color temperature mode, which starts from an average color temperature of around 7,500 degrees, and adjusted the red, green, and blue gain and bias controls. I am ashamed to say I wasn't able to match the out-of-the-box results, much less beat them.

Color decoder tests for the HDMI inputs revealed that red, green, and blue were all within a percent or two of perfect. The red was about 3% low through the component video input, but I was able to correct this by accessing the red -saturation control in the hidden installation menu. As compared to the SMPTE HD specification for digital TV colors, the set's green and blue primary color points were almost perfect at the factory settings (which are shown in the chart). The red was a little off; adjustment of the red saturation and value controls in the installation menu had little effect.

The Dream'E offers iris settings in 10 steps from 10 to 100. No auto iris function is available. It has two lamp -brightness modes, Normal and Economic. In the Normal mode with the iris set at 100, the projector delivers an eye-searing maximum output of 60.3 footlamberts. In my darkened room on my rather small 72-inch-wide 16:9 screen, I got the best compromise between black level and maximum light output in the Economic mode with the iris set at 40. With these settings, the native contrast ratio after calibration measured 3,720:1 and maximum output measured 18.6 ftL. The black level at this setting is acceptable; at brighter settings, the blacks and dark grays start to wash out.

The Dream'E's Contrast control does not operate correctly. Contrast controls are supposed to control white level, but this one actually seems to control black level in reverse - i.e., black level decreases when you turn it up . It has little effect on white level. It also has some abrupt steps. For example, when you go from 49 to 50 on the Contrast control's 100-step scale, the black level (not the white level) jumps dramatically - this one step brings the black level from 0.005 ftL to 0.07 ftL. The Brightness (black level) control exhibits similarly coarse steps. However, there are also Black Level and White Level controls (they're actually labeled that way) for the input signal. These can be set differently for each input. They work backward - i.e., when you turn the White Level control up, it goes down. However, they do control what they're supposed to control, and in more precise steps, too.

By using these and ignoring the main Brightness and Contrast controls, I could set black and white levels accurately and easily. The projector offers four preset gamma level settings, plus one user-adjustable gamma setting. In this setting, gamma can be adjusted separately for seven different signal-level ranges, such as 0-15 IRE, 16-29 IRE, etc. Gamma can also be adjusted for all colors together or separately for red, green, and blue. However, I didn't need these controls, because the Gamma 4 setting looked great for watching movies in my darkened room.

With 1080i/p-format high-definition signals and overscan switched off, the overscan measured 0.5% at top and bottom, 0% at left and 1.5% at right. With overscan activated, it measured 3% at top and bottom, 1% at left, and 2.5% at right. The projector displayed 1080i/p and 720p test patterns with full resolution on both the HDMI and component-video connections. The Converted 16:9 mode reduces horizontal resolution of 1080p material by about 24 percent. There's also some horizontal resolution loss with 720p material, although the effect is less noticeable. Standard-def material was unaffected, as was vertical resolution. The detail enhancement control produced slight ringing (white halos at the edge of on-screen objects) even at its lowest settings, so I left it off.

Brightness uniformity was very good on both white and black fields. On both, it looked slightly darker at the bottom and in the upper right corner. Likewise, the color shifted only slightly across the screen. It was a tad greenish in the lower left and a tad reddish in the upper right. None of these variations was visible with normal program material.

A crosshatch pattern showed a slight convergence error on red. X- and Y-axis position -adjustment controls are provided for red and blue, but the adjustments were too coarse to correct this problem.

The projector properly de interlaced 1080i signals from the Silicon Optix HQV Benchmark Blu-ray disc, and it performed at least as well on the disc's "jaggies" tests as any display I've seen, producing no visible rough edges on diagonals and curves. It also does a fine job of compensating for 3:2 pulldown on standard DVDs.

The installation menu offers adjustment of temporal noise reduction and mosquito noise reduction, as well as an on/off setting for block noise reduction. The first two controls are set to Low at the factory, and the last is set to Off. Experimenting with the noise reduction tests from the HQV Benchmark disc, I found turning the mosquito noise reduction up to medium and leaving the other controls at factory settings delivered a low-noise picture with almost all material, without noticeably affecting detail.

The fan noise is a little high compared with most other projectors I've recently reviewed. The noise isn't particularly troublesome, but the projector is not state-of-the-art in this area.

The Panamorph UH480 anamorphic lens that DreamVision offers as an option reduced the projector's maximum light output from 18.6 to 13.1 ftL when image stretch was ignored. When I moved the projector so that the 100 -IRE white window with the lens attached was the same width as the window without the lens (arguably a fairer test), the output improved to 16.4 ftL. With the lens in place, contrast ratio dropped from 3,720:1 to 3,280:1. The lens also shifted the color slightly toward red; at 50 IRE, color temperature dropped from 6,580 to 6,491 degrees K.

The lens produces about 2% pincushion distortion at the top of the picture, making the corners arc upward slightly. The sides and bottom exhibited no noticeable distortion. When I zoomed the Dream'E's onboard lens out to compensate, I lost a bit of the picture in the upper corners. Even so, it didn't fully correct the problem, because horizontal lines that appeared near the top of the screen looked noticeably curved. This is an issue with the lens, not the projector.