HD-DVD testing: After setup, we proceeded to watch some of the best program material we had available, which in this case was Batman Begins on HD-DVD played via HDMI at 1080i from our Toshiba HD-A1. On hand for comparison purposes was the competing JVC HD-56FN97, a 56-inch LCoS display, along with a couple of smaller high-def sets, the Sharp LC-46D6U LCD and the Panasonic TH-50PH9UK.
The darkness of Batman served to highlight the Sony KDS-R60XBR2's excellent black-level performance. In the very beginning, when the kid Bruce Wayne stares into the black hole at the bottom of the well, the darkest areas of the image looked inky and deep. Our measurements indicated that the Sony was noticeably darker than the JVC, although not quite as deep as the best flat-panels we've tested--the Panasonic and Sharp included. In general, flat-panels have an advantage over projection sets in maintaining a constant deep black level, because light from one area of a projection screen can affect another. Overall, however, the KDS-60XBR2 delivered as deep a level of black as we've seen from any rear-projection set, including the A2000.
Sony's Auto Iris setting is designed to improve contrast by automatically adjusting the iris depending on program content. We found that the Auto1 setting hampered black level performance too much, but Auto 2 was intriguing, in that it actually increased the contrast ratio in certain scenes, making bright areas brighter while keeping dark areas around the same. When Bruce Wayne enters the ninja temple, for example, the door opens up to reveal white sky that was brighter in Auto2 than it was in our default Min Iris setting. To our eyes the jump in brightness was actually a bit much--we preferred the more constant brightness levels of the fixed Iris controls, namely Min--but some viewers may like the extra punch.
Shadows throughout the film had plenty of detail and very little noise. When the adult Bruce returns to explore the hole and create his Bat cave, for example, we saw all of the shadowy nooks and edges in the rocks, and the spill of light from above as it faded into shadow, as well as the light from Wayne's glowstick looked clean and natural, with no trace of false contouring. We saw a tiny bit of contouring in these areas on the JVC and the Panasonic, but none on the Sharp.
Color accuracy, as you can see from the Geek Box below, is one area where the Sony KDS-R60XBR2 could improve. Its primary colors of both green and red were pretty far outside the HDTV colorspace, but were not overly tinged with yellow or orange, which is common on some other sets. Other aspects of the 60XBR2's color, namely its grayscale accuracy and color decoding, were superb, so overall color still came across as lush and vibrant. When young Bruce runs through the garden, for example, the green of the trees and plants looked rich, if a bit too green, and in tones throughout the film, from the ruddy police chief to the delicate face of Rachel (Katie Holmes), appeared natural. Nonetheless, we wish the Sony had some way to improve the accuracy of its primary colors, which would certainly be worth an extra performance point.
Details were sharp, and according to test patterns, the Sony handled every detail of a 1080i source from our HDTV signal generator (this was an issue on the A2000 we tested, but Sony claims to have addressed it on subsequent A2000 models). We did notice that on some scenes, however, certain areas of the Sony's image appeared a hair softer than on the JVC. Inside Wayne's mansion, for example, some of the bric-a-brac on a table and designs in its leg looked very slightly less defined than with the JVC. We chalk this up to the fact that the focus of the review sample we tested was slightly softer than that of the JVC, an effect that was visible more on the sides than the middle of the picture. As always, focus issues in rear-projection HDTVs can vary from one unit to the next.
The Sony's uniformity across the screen was very good for an LCoS display. As usual with a rear-projection set, the middle of the picture appeared brighter than the outside, but the resulting hot spot was less noticeable than on the JVC, for example. The extreme bottom edge of the KDS-R60XBR2's image appeared slightly brighter than the adjacent area, and in some mid-gray fields we saw that the bottom third appeared very slightly greener than the rest of the image. We didn't notice any of these issues in normal program material though. As we expected, the Sony suffered a bit when seen from off-angle, becoming dimmer and slightly discolored as we moved to the extreme edges; when viewed from above or below, the screen became discolored much more quickly. These effects are typical in rear-projection sets, and the Sony handled off-angle viewing better than many we've seen.
Geometric distortion on our review sample was nearly nonexistent, with straight edges on all sides and almost no trace of bowing. Details in sides and corners appeared slightly softer, namely less focused, than on the JVC. Fringing around white lines was virtually nonexistent, and as we'd expect on any 1080p display, it was impossible to discern individual pixels from further away than a few feet.
One issue we've complained about in the past is stationary screen grain, where subtle, tiny sparkles in the screen become visible, especially during pans or flat fields. The KDS-R60XBR2 showed less of this kind of grain than the JVC, for example, and although spots did crop up occasionally, such as in a green shade as the camera panned across a courtroom, they weren't nearly as prevalent as we've seen on many other rear-projection sets.
1080p vs 1080i: We tested the Sony primarily with a 1080i source, mainly because we trust the Toshiba's image quality more than we do the Samsung BD-P1000 Blu-ray player, the only relevant source we have capable of outputting a 1080p signal. But we did watch some 1080p material and, surprisingly, found that the Sony did indeed look better with a 1080p source, at least in one instance.
In the beginning of Chapter 8 in the Mission Impossible 3 Blu-ray disc, the camera pans over a set of stairs leading down into a party. When we set the Samsung player to 1080p, the stairs appeared relatively solid, but when we switched to 1080i and watched the Sony, the stairs sort of strobed and flashed as the camera moved, and a couple concentric lines of moirÃ© patterns appeared. The JVC exhibited none of this behavior when played through the scene (although, if we chapter skipped backward to the beginning of the chapter, the JVC would exhibit similar issues for second or two, and then the stairs would appear solid after it locked in). Based on this behavior, we believe the Sony is not properly implementing 2:3 pulldown with 1080i material. We couldn't find any other instances of this happening, however, and we currently don't have another 1080i source to test for 2:3 pulldown. While incidents like the strobing stairs are noticeable, they're rare enough that we don't consider this a major picture quality issue--we didn't notice any similar effects during Batman Begins, for example, even in the shot where the camera moves across the stairs of Wayne's mansion. We didn't notice any other major differences between the Sony's handling of 1080i and 1080p.
Standard-def performance: One of the few differences between the KDS-60A2000 and the KDS-R60XBR2 is the latter's supposedly improved video processing. Sony calls the step-up version "DRC-MFv2.5." In one of our tests it actually improved image quality somewhat (2:3 pulldown during Star Trek: Insurrection; see below), but overall the KDS-R60XBR2's standard-def performance, meaning its picture quality via component video at 480i, as well as S-video and composite video, was still relatively disappointing for such a high-end TV. Of course, if you're watching standard-def from a set-top box that's connected only via HDMI and set to output a resolution other than 480i, then the box does the processing and the comments below don't apply. The same goes for upconverting and progressive-scan DVD players.
Our main method of testing SD performance is to use Silicon Optix's HQV test disc, which includes a variety of patterns and real video to challenge various aspects of a display's processing. On the resolution tests the XBR2 did fine, as long as we turned DRC Off, but when engaged, it introduced varying degrees of flicker, more with composite than with S-video or component video. All of the inputs had a difficult time removing jagged edges from diagonal lines and the waving American flag.
One test of the HQV disc evaluates 2:3 pulldown processing with a pan that follows a race car in front of some bleachers; if the bleachers show a moirÃ© pattern, resembling concentric curved lines, then the test fails. The KDS-R60XBR2 failed this test, but did so differently from its predecessor. In component and composite video, with CineMotion engaged and DRC on, it stayed in film mode most of the time but dropped in and out, causing brief bursts of moirÃ©. It never properly locked into film mode via S-video. We also watched the opening pan of Star Trek: Insurrection, which was an improvement over the A2000; the set successfully eliminated moving lines from the upturned boats and jagged edges from the houses via all inputs. Strangely, the KDS-R60XBR2 also introduced a combing effect in some scrolling text overlays, which disappeared when we turned CineMotion off.
Once again the Sony's detail was quite sharp with standard-def. We found that with composite-video sources especially, engaging DRC Mode 1 did sharpen details in the picture, although it also appeared somewhat crunchier (edges were exaggerated) and noisier. With S-video the difference between DRC Off and on was less noticeable, but still there, whereas the difference in detail between DRC on and Off via component video was almost impossible to discern.
Of the four levels of noise reduction, High was easily the most aggressive. It did an excellent job of squelching snowy motes in the sky scenes, but we did see a loss in detail; edges seemed less sharp. We didn't note the same loss in detail in High mode with the A2000, but since we don't have them side by side, it's difficult to confirm. The other three modes (Low, Medium, and Off) were less effective at reducing noise, but Medium worked relatively well without sacrificing detail.
Xbox 360 testing: We hooked up the Xbox 360 to the Sony via our component-video adapter (we also tried VGA for fun, but issues with the VGA input led us to abandon that) and the results were mostly good. 1080p purists might complain that the set cannot accept 1080p sources via its component-video inputs, but we didn't think that was a huge deal since the games looked great in 1080i--and as we mentioned above, the Sony resolved every detail of 1080i sources via component video.
PC performance: We also tested the Sony's VGA input and the results were disappointing. We set our PC to 1,920x1,080 and unlike the 60A2000 we reviewed, the R60XBR2 did display an image--but it was far from ideal. We could find no way to get the image to fill the screen completely; there was a black border about 8-inches wide on all sides, which is unacceptable. We were able to reduce interference by playing with the pitch and phase controls, but never completely eliminate it.
The Sony fared much better as a PC monitor when connected via HDMI. It filled the screen and resolved every line of a 1,920x1,080 source, according to DisplayMate, and text was relatively crisp. The image was overscanned significantly, however, so the task bar at the bottom of our Windows XP desktop, for example, was completely obscured along with one column of icons on the right side. According to DisplayMate, overscan measured 3 percent on the top and bottom and 2.5 percent on the sides--and we couldn't do anything with the Sony's controls to improve that. Of course, depending on the drivers in your video card, you should be able to correct the overscan at the expense of re-scaling the image.
|Before color temp (20/80)||6,128/6,350K||Good|
|After color temp||6,421/6,462K||Good|
|Before grayscale variation||+/- 201K||Good|
|After grayscale variation||+/- 53K||Good|
|Color of red (x/y)||0.674/0.324||Poor|
|Color of green||0.276/0.703||Poor|
|Color of blue||0.138/0.057||Average|
|Black-level retention||All patterns stable||Good|
|2:3 pull-down, 24fps||No||Poor|
|Defeatable edge enhancement||Yes||Good|