LCD TV Reviews

Mitsubishi's PerfectColor offers only a single dimension of adjustment (saturation) rather than the more common practice with this sort of feature, which is to offer saturation and hue adjustments. I did find that I could move both the primary and secondary color points and thereby significantly improve the accuracy of the set's color gamut, but only by using the right test tools (see "Measurements").

The color was consistently vivid but at the same time subtle and refined. It went over the top only if the program material demanded it. In the recent ALCS baseball playoffs between the Cleveland Indians and the Boston Red Sox, everything looked right, from the deep but not excessively red trim on the player's uniforms to the green grass on the field.

But the promotions inserted at various points in the games, including clips from earlier games edited together, appeared to be processed to look flashy, with exaggerated, garish colors. And that's how they looked on the Mitsubishi.

Given a good HD source, the set was also crisply detailed. Yes, even at the zero setting of the Sharpness control there was a trace of white-line edge enhancement remaining on test patterns, but this was never obvious as unnatural-looking edges on real program material.

One of the best-looking Blu-ray Discs I have is a demo produced last year by Pioneer, with material ranging from brightly lit video material to film trailers. Every bit of it looked strikingly good on the Mitsubishi, particularly a series of short clips sourced from European high-definition programming. The detail was superb, the colors rich and believable, and the underlying black level more than sufficient to give the image a convincing sense of depth.

Two pressing issues with LCDs in general are motion blur and off-axis viewing. The Mitsubishi has a built-in demo to show what its 120Hz operation can do to cure the first of these problems, and it was impressive. But I did not rely on this alone; I used other familiar test material as well.

As it turned out, the 120Hz feature was not a complete cure; the recently reviewed 60Hz Sharp LC-52D64U did a better job—but only with its motion compensation features turned on. Some motion blur was clearly evident on both sets, but it was more easily visible on the Mitsubishi. But these torture tests are brutal at isolating this problem. In all of my viewing of normal program material on the Mitsubishi I was never distracted by motion blur.

The Mitsubishi can't completely escape LCD's typical off-axis viewing limitations, but it's better than most I've seen. While you'll get the best image within a relatively narrow cone centered directly in front of the set, the falloff in quality up to at least 45 degrees is not severe. The image remains very watchable at this angle. Most viewers will not notice the resulting degradation, visible as a slight loss of contrast.

The set's color uniformity was respectable for an LCD. I could see some slight discolorations if I looked hard enough on appropriate test patterns, but the black and white material I watched looked—appropriately—black and white.

But two problems intruded on the otherwise impressive performance of this set, one of them curable by a work-around, the other perhaps intractable.

The first of these was poor video processing from a 480i source to the set's native 1080p. It failed about half of the deinterlacing/scaling tests on the Silicon Optix HQV Benchmark DVD, and was particularly prone to jagged edges. You can see the result for yourself (depending on the resolution of your computer monitor) in Photo 1, a screen shot from one corner of a test pattern on Digital Video Essentials from Joe Kane Productions (the JKP in the photo). Photo 2 shows the same image, but this time the DVD player—a Pioneer Elite DV-79AVi, was set to output 480p instead of 480i. This is far smoother, indicating that the Mitsubishi's problem lies in its 480i-to-480p deinterlacing. (Both images actually showed somewhat more jaggedness when viewed directly than is visible here, due to the many steps they had to go through to arrive on your computer screen, including resolution reduction. But apart from that, no processing was performed on either photo.)

Photo 1

Photo 2

To be fair, the Mitsubishi's 480i jaggies were less obvious at a typical viewing distance (about 8 feet) with normal program material. But they were still visible.

If you regularly use an upconverting DVD player, however, and use the upconversion in a set-top-box for standard definition television programming, this will not be an issue for you. With the exception of direct 480i feeds into the set's antenna inputs and onboard tuners), a good external processor will cure this issue, albeit at a price.

But for such a high-end set Mitsubishi should be able to offer video processing at least as good as a two-year old Pioneer upconverting player, which itself offers none of the marketing cache of processors from the likes of Silicon Optix, Faroudja, DVDO, Gennum, and others.

On the plus side, however, the Mitsubishi's 1080i-to-1080p deinterlacing is good. Not great—it does not recognize 3:2 pulldown- but that's a limitation common to most of the sets I've tested.

While the external processing option can work around the Mitsubishi's deinterlacing shortcomings, the second problem is more troubling. On some scenes I saw a persistent flow of horizontal bands that moved rapidly from the bottom to the top of the screen. This strobing was subtle—not well defined but rather shadowy and indistinct—but once spotted could not be ignored. It was invisible on bright scenes and in very dark areas, but obvious on scenes of low-mid brightness in patches of solid color or gray. Complex foreground action rarely showed it, but backgrounds did. The strobing could also be seen clearly in medium gray test patterns.

Was this a problem with my power line, or something else specific to the test location? I don't believe so. The problem did not resemble hum bars, and I have not experienced it with any other set reviewed under the same circumstances. This included the Sharp LC-52D64U, which I was able to view side-by-side with the Mitsubishi on the same program material.

Conclusions
The strobing issue must temper my recommendation of this set. It's a problem that most users will see, and many may be mentally unable to tune out.

Which is unfortunate, since the Mitsubishi LT-46144 otherwise drew me in with its compelling performance. And while the best new plasmas I've tested can do somewhat better, particularly in producing black levels that no LCD I've yet seen can touch, those sets (particularly the new Pioneers) can cost more, raise concerns about burn-in—a non-issue with LCDs- and offer less brightness than a good LCD for those who feel the need for it.

The Mitsubishi comes closer than I ever expected from an LCD in providing a completely satisfying balance of black level and shadow detail, color, and resolution in one very appealing package.

Highs
Excellent resolution
Blacks and shadow detail impressive for an LCD
Good color out of the box, but can be improved by calibration

Lows
Horizontal bars strobe from bottom to top on mid-brightness images
Poor 480i-to-480p deinterlacing
Service menu does not provide the easiest calibration experience