Price: $25,000 At A Glance: Stunning resolution • Superb color • Glorious 2D and 3D performanceWith 4K-resolution
Ultra HD the latest and greatest star in the consumer electronics galaxy, we ink- and pixel-stained wretches of the press were all champing at the bit to lay hands on one. But at a massive 84 inches diagonal, 216 pounds with its floor stand, priced high enough to put you in a nice new car as long as your tastes aren’t too posh, and still limited in availability, Sony’s new 4K flagship made the company understandably reluctant to ship review samples to all the usual suspects.
The company therefore invited the press to spend individual time with the set at either its East or West Coast headquarters. So if (you know who) won’t come to the mountain, the mountain must come to (you know who). Or vice versa. This arrangement wasn’t equal to the extended review process we prefer, but the product was too important to ignore. So I schlepped all of our test gear to
Sony San Diego. The facilities were fine, the windowless room could be darkened, and apart from a morning briefing on my arrival and assistance when needed, Sony’s techs left me alone to work on the set.
For those still unfamiliar, Ultra HD is the latest moniker for the consumer version of the “4K” high-resolution video format that comes out of the digital cinema realm. (See “4K: What, Why, and When?” sidebar). Technically speaking, Ultra
HDTVs offer 3840 x 2160 pixel resolution, or exactly four times the number of pixels present in the existing 1080p standard, otherwise known as
full HD. While plans are afoot to crank up development and home delivery of native 4K content to watch on the new displays, for the present and probably near future, most of the program material will be upconverted from 1920 x 1080 pixels. So a huge concern for reviewers and consumers alike is whether traditional HD can be scaled from 2K to 4K in a way that truly challenges the quality of native 4K, thus giving a buyer of this set and others enough source material to make the investment worthwhile. That was the $640,000 question that I hoped to answer going into this review. So what did I find? Read on…
FeaturesIn most respects, the XBR-84X900 mimics the features found in
Sony’s XBR-55HX950. The remote and XMB (
XrossMediaBar) menus are similar. There are full White Balance adjustments here for tuning the gray scale but, typical of Sony sets in general, no color management system (CMS) for adjusting the primary or secondary color points from which all the other colors derive.
But the XBR-84X900 also differs from previous Sony flat-screen sets in important ways, chief among them its 4K capability. It accepts native 4K material and upscales lower-resolution sources to 4K employing a new processing chip that microanalyses individual aspects of each image to emulate a native 4K picture. In addition, with the use of special glasses, two-player SimulView
PlayStation games can be played with each player viewing only his or her own full screen (not split-screen) part of the action. As for native 4K, the set’s HDMI 1.4a inputs will accept 4K sources up to a 30p frame rate.
Unlike the full backlighting LED array of the XBR-55HX950, the XBR-84X900 is top and bottom edge-lit. Both sets feature Sony’s LED Dynamic Control (local dimming). This can work well, and does here (more on that later), but edge-lit local dimming is nevertheless less effective than full-array backlit local dimming that individually modulates dozens of LED clusters locatedbehind the screen. Also unlike the XBR-55HX950, the XBR-84X900 does not use Corning’s tough Gorilla Glass for its screen surface. That’s a big, wonking sheet of very reflective, expensive glass, so no playing catch, shooting hoops, or practicing your golf swing in the TV room. Time to move the kid’s Wii console into the playroom and let the foam bats, fake swords, and untethered control wands take flight at some cheaper screen.
The XBR-84X900 uses an IPS (
In Plane Switching) LCD panel. As far as we know, this is a first for Sony. IPS provides off-axis performance superior to other LCD designs, and that was clearly visible here. Even when I moved 45 degrees to the side, I saw none of the color desaturation and loss of contrast epidemic among
LCDs in that situation. But IPS typically has poor contrast, an issue that Sony’s LED Dynamic Dimming deals with effectively here.
Unlike all previous, 3D-capable Sony sets, the XBR-84X900’s 3D is passive, not active. Passive 3D reduces a set’s vertical resolution to half of its native vertical pixel count. In a typical
3D HDTV, this results in a less than full HD resolution of 1920 x 540 at each eye. It also produces a grid of black horizontal lines in the image that, depending on the size of the screen, can be visible at normal viewing distances. With a 4K set, however, these issues disappear. The horizontal lines are gone, and while the vertical resolution presented to each eye is still reduced by half, with a 4K display the overall 3D resolution seen by each eye becomes 3840 x 1080, vertically equivalent to full HD. The passive glasses are also lighter and (usually) cheaper than the active variety, and without the alternate masking of the left and right eyepieces (required for 3D using active glasses), the 3D image is brighter and there’s far less chance of
3D crosstalk (ghosting).
The set’s audio is produced via removable speakers on either side of the screen. There’s no denying that the Sony’s sound is a step up from what you get from most flat-screen HDTVs. But from what I heard—making allowances for the large, live room and a placement not conducive to the best sound—it’s certainly no substitute for full surround sound and, as with every flat-screen set we’ve tested, unlikely to challenge the audio from even a modest outboard two-channel system.
Features not tested here, due to time limitations, were Motionflow and Smooth Gradation (Sony’s motion smoothing features—left off for the entire review) and the set’s Internet functionality.
First, the Tech Talk
I’ve got much good to talk about here when it comes to picture quality, but it’s worth noting up front that, surprisingly for such an expensive TV, the Sony failed both our SD (standard definition) and HD (high definition) motion adaptive (MA) deinterlacing tests, producing more rotating bar jaggies than would permit a passing grade. It also failed the 3:2 SD pulldown test and rolled off the highest-frequency burst in our Chroma Resolution test. Then, in attempting to check for how serious the deinterlacing issues might be on real-world 480i material (using the DVD release of Titanic), I discovered a general softening of 480i and 480p sources to a degree that was dependent on the setting of the Sony’s sharpness control. However, even with sharpness at the technically optimum setting, the visible result from letting the Sony do all the upconversion work for standard-definition content was clearly inferior to letting our Oppo BDP-103 Blu-ray player do the 480i/p-to-1080p conversion before handing the signal over to the Sony for the final step up to 4K. This softening was more of a concern to me than the jaggies and in any event would have obscured subtle deinterlacing artifacts on real-world sources. At that point, with my time limited, I elected to chase the 480i issue no further.
Fortunately, the admittedly limited number of Blu-ray Discs I brought with real-world 1080i material on them (something other than test patterns) didn’t show any obvious consequences from the set’s failure to pass all of our formal HD deinterlacing tests. Overall, however, and to be consistent, our 2D performance rating was downgraded slightly from five stars only because of the Sony’s deinterlacing issues and softness with 480i/p standard-def content. And these went away when I let the Oppo BDP-103 Blu-ray player upconvert all interlaced material, HD and SD, to 1080p. Both 1080p and native 4K source material, with which I spent most of my allotted time, were unaffected by these results. Neither required scaling by the set that might produce deinter-lacing artifacts. And there were no visible issues from the chroma rolloff, which is not uncommon in many sets and is often done deliberately for valid technical reasons.
The Sony does offer a blizzard of five different adjustments that provide varying degrees of detail control: Sharpness, Reality Creation, Detail Enhancer, Edge Enhancer, and SBM (Super Bit Mapping). I left the Enhancers off and the SBM on for the entire test (SBM appeared to have no negative impact), but I initially vacillated between two combinations of Sharpness and Reality Creation. With the Sharpness on 50 (maximum 100) and Reality Creation off, the set produced the cleanest sharpness patterns, with little or no visible enhancement when viewed from mere inches away. With Sharpness on 10 and Reality Creation on Manual (Resolution 10; Noise Filtering 0) and with my nose nearly on the screen, I could see a trace of white line edge enhancement. But this was invisible at my viewing position, where this combination of settings produced a subtle but useful increase in subjective detail. This was the setting I used for my testing and viewing.
(After our access to the set had expired, Sony suggested different sharpness and Reality Creation control settings
(see "Settings" page) that they say would have yielded a sharper image with 480i content. These are included on the Settings page. Sony did acknowledge that the jaggies we detected were endemic to the set.)
The Sony’s white-field uniformity was good. Not perfect—no video display is. But I was never distracted by oddities on predominantly white or near-white images. When viewed in a darkened room, LED edge lighting can also produce serious black-level uniformity issues in very dark scenes. But this was not the case here—with a single exception. When the source goes completely dark, the Sony’s screen goes very nearly full black as the LEDs shut off. But when the screen goes completely dark except for a small point of light, such as the pause bug the Oppo player places in the upper left-hand corner, the bug lightens that area of the screen. Because the edge lighting is top and bottom, you’ll see a fuzzy, lightened gray vertical stripe on the screen from top to bottom, above and below the bug. But this was rarely visible on normal program material.
The set will accept signals with color depth of up to 12-bit, though all the material on a 4K Sony movie server I had available (more on that below), as with all current and most future consumer video source material, was limited to 8-bit color. The color gamut on the server material was Rec. 709, the same as the current 2K HD standard. The XBR-84X900 does not offer a wider color gamut, which is not an issue as long as the source is no wider than Rec. 709—and no consumer sources (apart from some color photography) currently are. The movie server content was also delivered and digested by the XBR-84X900 with 4:2:2 color subsampling, which is a fancy way of saying that less digital compression of the color signal was used than is normally found in consumer video content.
4K: At LastAs explained in the accompanying
sidebar, a server loaded with 4K movies and shorts comes with the XBR-84X900. The quality of this material varied. Among the movies,
The Amazing Spider-Man,
Battle: Los Angeles,
Salt, and
Total Recall (2012) were of reference quality.
The Bridge on the River Kwai looked its age, and
Taxi Driver was grainy, but that’s likely in the film source.
[Ed. Note: Indeed it is; I saw this stunning restoration projected digitally in a theater, and it preserved the hard look that Scorsese effectively used to impart the grittiness of the crime-ridden New York City of the mid-1970s.—RS] The remaining titles (
The Other Guys,
The Karate Kid,
Bad Teacher, and
That’s My Boy) looked fine but weren’t up to the level of the other four.
Sony’s opening orientation brief included comparisons of native 4K material and the same material in 2K, but upconverted by the set. They used two XBR-84X900s side by side, both in their Cinema 1 default settings. I guessed wrong as to which was which—and anticipated up front that my guess would be wrong! The simulated 4K looked very subtly sharper when I focused on the tiniest details in the picture. This was likely due to a little overeagerness in the upconversion algorithms, but it did not look edgy or processed.
After the briefing, we turned off all but one of the XBR-84X900s, and I was left on my own. In addition to the Sharpness and Reality Control settings described earlier, I turned the backlight control down to 2 for a peak 2D brightness of just over 37 foot-lamberts. The default backlight setting used in the briefing was 5 (maximum 10), producing a peak white level of 48.7 ft-L—too bright for my taste in a fully darkened room. I did most of my viewing at about 9 feet from the 84-inch screen, for a horizontal viewing angle of 37.4 degrees—about 2.6 screen heights, or 1.5 screen widths.
My calibrations, performed with 2K test patterns (no 4K patterns were available) translated well, subjectively, to the native 4K sources. The lack of a color management system was a downer, but the set could nevertheless be tweaked to adhere closely to the standard Rec. 709 HD color gamut. The color points were very close to correct, though a little work was needed to get the color brightness readings as close to the standard as possible. More on this in
“HT Labs Measures.” Visually, the colors were never less than outstanding, with spot-on fleshtones and bright, natural hues in every source I watched.
On difficult, dark scenes, the XBR-84X900 produced solid blacks. They weren’t quite the equal of what we’ve seen on sets with full backlighting with local dimming, including Sony’s own XBR-55HX950, and they certainly weren’t up to the standard of my reference Pioneer Kuro plasma. But the dark interior scenes in Stargate Universe (a television series that’s a movable feast of dark scenes) were appropriately gloomy. Some of the exterior starfield shots looked a little gray, but others were fine. The brief starfield shots at the beginning of Prometheus (another black-level torture test), however, were more impressive, as were the shadow details in the dank and dim caverns on the menacing alien planet.
But it was the native 4K material from the server that really grabbed my attention. It was as naturally sharp as I could hope for, with an uncanny combination of detail and creamy smoothness, almost as if the regular HD pixel structure I’m used to on 2K displays was adding a subliminal graininess to the image that native 4K eliminates. If you get really close to the screen, you can see the very tiny pixels, and they were uniformly sharp from corner to corner and edge to edge.
I also sampled a number of Blu-ray Discs I brought with me, and as upconverted by the set, none of them disappointed. In fact, despite their 2K roots, they had enough of the 4K look to make me less concerned about the current shortage of native 4K source material. One of the movies on the Sony server was Battle: Los Angeles. I knew this ahead of time and brought the Blu-ray version for comparison. Once you locate a scene in this film without shaky-cam syndrome, it’s a sharp, crisp BD transfer. While this comparison was done by switching inputs, with a slight delay, on the same set rather than side by side on two sets, my earlier impression was reinforced: The upconversion of a good 2K source was very nearly indistinguishable from native 4K on this set. Longer comparisons than I had time for might qualify this conclusion, but I was impressed.
I was able to do a full 3D calibration, but the time I had for actually watching 3D was limited. Nevertheless, what I did watch looked spectacular. With a peak brightness around 23 ft-L, 3D popped off the Sony’s big screen in a way 3D rarely does even on much smaller sets. And the resolution, even though it was an upconversion from 2K 3D on Blu-ray and limited to 1080 lines vertical by the passive 3D process, was easily the most impressively sharp and detailed 3D I’ve yet experienced from a consumer display. It was so much fun to watch that I regretted not bringing more 3D discs apart from Avatar and Wreck-It Ralph. But they were more than enough to confirm the Sony’s impeccable 3D credentials.
Conclusions
How do you rate the value of a product that has limited competition in its market segment—even if the price is enough for a comfortable down payment on a house? You can’t, particularly if that product excels at what it does. Just how much of the Sony XBR-84X900’s outstanding performance—whether the source is 2K, 4K, or even 3D—is due to its native 4K, and how much can be credited to the extra margin the price gives for the designers to elevate every other aspect of the set’s design to a higher standard? We can’t say as yet for certain, but we can say that 4K certainly doesn’t hurt.
Yes, I would have preferred to spend several days with the set after calibration, as we usually do in our in-house reviews. Long-term viewing often turns up unexpected information that the eager flush of a first infatuation cannot. But the time I had was sufficient to raise covetousness to a steamy level, steamy enough to mandate a visit to a confessional ASAP. If only I had the space and the ready cash!
Type: LCD
Screen Size (diagonal, inches): 84
Native Resolution: 3840 x 2160
3D: Yes (passive glasses)
HDTV Tuner: OTA
Backlight: LED edge-lit with local dimming
Wall Mount or Stand Included: Floor and tabletop stands
Dimensions (W x H x D, inches): 76 x 44.75 x 3.25 (without stand); 76 x 59.5 x 22.5 (with floor stand)
Weight (pounds): 160.2 (without stand); 215.7 (with floor stand)
Price: $25,000
Connections
Video Inputs: HDMI (4, 4K on HDMI 2 and 3), PC D-sub 15-pin analog RGB (1), component video (1), composite video (2, 1 shared), cable/antenna (1)
Audio Inputs: Stereo analog (2), PC/HDMI (L/R minijack, common with HDMI 3)
Audio Outputs: Optical digital (1), headphones (1, analog)
Additional: USB/DLNA (2), LAN, Audio Return Channel (ARC) on HDMI 1, RS-232C (1)
4K: What, Why, and When?
It was at the movies where we all first encountered a 4K display. Not all digital theaters are 4K today, but many are, with Sony and Christie Digital the leading U.S. suppliers of 4K digital cinema projectors and systems. The theatrical standard for 4K resolution is 4096 x 2160 pixels. For the as yet infant home 4K market, however, the trend is toward 3840 x 2160, or basically double the horizontal and vertical resolution of the best current HD. This is done because a uniform doubling greatly simplifies the upconversion process for regular 1080p full HD to 4K and greatly minimizes the potential for artifacts. That amounts to 8,294,400 pixels, versus the 2,073,600 pixels (maximum) in our current HD format.
Proponents of 4K, which includes virtually all current TV makers, have adopted the Ultra HD moniker to differentiate it from traditional full 1080p HD in the marketplace. It is said to offer significant advantages over crusty old HD. With four times the number of pixels, the pixel structure will be virtually invisible, even when you sit close to a very large screen. Skeptics argue that 4K will be beneficial only in jumbo screen sizes. At 50 inches and under, 4K would not appear to make much sense—either visually or economically. But none of us has yet spent hands-on time with 4K sets in the largest currently popular and affordable consumer sizes. That’s what it will take to judge whether or not there will be any advantage to 4K in, say, a 65-inch set. Stay tuned.
The larger question about 4K is when. Apart from a few short Internet 4K downloads and the output of some video cameras, 4K material must currently be sourced from a computer server loaded with 4K files. Sony provides just such a device to buyers of the XBR-84X900, preloaded with 10 movies plus a variety of shorter material, as an extended loan. The server provided to the earliest buyers will be replaced at some point, and at no charge, with a more flexible design (re-loaded with the same content) to which external storage may be added. Subsequent buyers will receive the new server, also pre-loaded with the initial launch content. Additional downloadable movies will be offered periodically for a yet to be determined charge. Sony has an advantage over other HDTV manufacturers in its ability to offer native 4K material. It owns a major movie studio and is currently engaged in remastering dozens of titles in 4K for possible release in some form. New movies are increasingly being shot digitally in 4K as well.
Ongoing developments in video compression algorithms, specifically the pending rollout of the new HEVC (high-efficiency video coding) standard that will replace the AVC standard now commonly used for Blu-rays, may allow for a new Blu-ray-like player offering 4K movies and other programming on disc (though there’s zero chance that such a format will be compatible with existing Blu-ray players). But at present, there are no publicly announced plans for such a disc format as content providers wrestle with the murky future of packaged media competing against what till now have been inferior but more convenient downloads. A company called RED, which makes professional 4K cameras, has shown a compact, $1,450 Redray server using its own proprietary compression algorithm and is said to be working with a partner to deliver content to consumers. Information on what that might be, however, remains hazy.
Even with 4K streaming or download services being launched, or new disc or other hard media options becoming available, it’s clear that, for a long time to come, the vast majority of content anyone will watch on an Ultra HD display will be upscaled 1080p. As we did with the accompanying review of the Sony XBR-84X900 Ultra HD TV, Home Theater’s evaluations of these displays will continue to pay close attention to the quality and value of that conversion.
Full-On/Full-Off Contrast Ratio: 37,140:1
The measurements below were taken in the Cinema 1 Picture Mode, through an HDMI input, and in 2D unless otherwise noted.
With the Backlight on 2, the Picture (Contrast) control at 90, the Brightness at 53 (as measured here; later reduced to 51), the Gamma on 0, and the LED Dynamic Control on Standard, the peak white level was just over 37 foot-lamberts. Since the LEDs turned (nearly) full off with the full black field used for the black measurement, the black level as seen by our Minolta LS-100 light meter was 0.001 ft-L—essentially unmeasurable with precision since this is the lowest level the meter will read. But you’ll definitely want to use LED Dynamic Control. With the latter off, the black level increased to 0.043 ft-L for a full-on/full-off contrast ratio of 865:1. The dynamic control is likely this aggressive because IPS panels sacrifice good native contrast for broad off-axis performance, which may be why most manufacturers don’t use IPS panels. But in the limited viewing time possible for this test, I never saw any obvious brightness pumping with LED Dynamic Control on Standard.
All of the color results and charts provided here, together with additional data, were obtained using SpectraCal CalMAN, SpectraCal.comDelta E is a figure of merit indicating how close the color comes to the D65 HD color standard. Values below 3 are generally considered visually indistinguishable from ideal. Pre-calibration, the set’s gray scale Delta E averaged an impressive 0.81, with a high of 1.43 at 100 percent. The pre-calibration 2D color point Delta E averaged 5.45, with red and green the major offenders at nearly 7. It was hardly worth the trouble to do a gray scale calibration, but I did one anyway with the average Delta E dropping to 0.35. While the set has no color management system, I was able to improve the color Delta E to 1.91 with creative tweaking of the Color control, since the major pre-cal color gamut offender was the brightness of the colors, not their locations on the CIE chart. The latter are shown in the accompanying diagram.
The 2D gamma averaged 2.36 in the 0 setting of the gamma control. The maximum was a rather high 2.62 at 70 percent brightness, decreasing to 2.29 at 20 percent and 2.22 at 100 percent, though this odd curve (with LED Dynamic Control on Standard) produced no obvious picture degradation. With LED Dynamic Control off, the gamma averaged just under 2.2 and was consistent across the brightness range (2.12 minimum at 20 percent and 2.21 maximum at 90 percent).
A 3D calibration (diagrams not shown) required considerably more adjustment but was relatively straightforward. The gray scale Delta E pre-calibration average was 9.00. Post calibration, the average was 0.64, with a maximum of 1.08 at 100 percent brightness. The color Delta E averaged 6.19 before and 3.57 after. With the gamma control on 0, the 3D gamma averaged 2.0, with a minimum of 1.82 at 30 percent and a maximum of 2.05 at 80 percent.—TJN
Unit-to-unit sample variations, the viewing environment, and the source might render these recommendations less than optimum. They are provided only as a potentially useful starting place.
The settings here that are most likely to translate reliably from one sample to another are those involving specific features with only a few setting options, such as Color Gamut, Gamma, and Noise Reduction. The ones most likely to be subject to sample variations are video controls offering a wide range of settings. This will be particularly true for color temperature, white balance, and color management adjustments (where available).
We strongly recommend that you find the optimum basic video settings for your sample by using one of the many display setup DVDs that are available, such as Digital Video Essentials(DVD) or DVE HD Basics (Blu-ray). A full calibration, particularly of the gray scale and color gamut, is best left to a trained and properly equipped technician such as those certified by the Imaging Science Foundation (ISF) or THX.
| 2D | 3D |
| Picture Mode | Cinema 1 | Cinema 1 |
| Scene Select | Cinema | Cinema /td> |
| Backlight | 2 | Max (fixed) |
| Picture | 90 | 90 |
| Brightness: | 51 | 50 |
| Color | 55 | 44 |
| Hue | 0 | 0 |
| Color Temperature | Warm 2 | Warm 2 |
| Sharpness | 10 (see text) | 10 (see text) |
| Noise Reduction | Off | Off |
| MPEG Noise Reduction | Off | Off |
| Dot Noise Reduction | Off | Off |
| Smooth Gradation | Off | N/A |
| Reality Creation | Manual | Manual |
| Video Area Detection | Off | Off |
| Resolution | 10 | 10 |
| Noise Filtering | 0 | 0 |
| Motion Flow | Off | Off |
| CineMotion | Auto 1 | Auto 1 |
|
|
| ADVANCED SETTINGS |
| Black Corrector | Off | Off |
| Advanced Contrast Enhancer | Off | Low |
| Gamma | 0 or +1 | 0 |
| LED Dynamic Control | Standard | Standard |
| Auto Light Limiter | Off | N/A |
| Clear White | Off | Off |
| Live Color | Off | Off |
|
|
| White Balance |
| Red Gain | 0 | –8 |
| Green Gain | 0 | –10 |
| Blue Gain | –1 | 0 |
| Red Bias | –1 | –2 |
| Green Bias | –1 | 0 |
| Blue Bias | +1 | –1 |
| Detail Enhancer | Off | Off |
| Edge Enhancer | Off | Off |
| Skin Naturalizer | --- | --- |
| i/p Conversion Preference | --- | --- |
| Screen Display Area | Full Pixel | Full Pixel |
| 3D Settings | All Default | |
| Ambient Sensor | Off | Off |
|
|
| Pro Picture |
| HDMI Dynamic Range | Auto | Auto |
| Color Matrix | Auto | Auto |
| SBM | On | On |
|
|
Addendum
During our final review fact-check with Sony and after we no longer had access to our test sample of the XBR-84X900, Sony suggested different global Sharpness and Reality Creation control settings that they say would have yielded a crisper image with 480i content and better addressed the softness with 480i/p programming that we observed in our review. Unfortunately, we were not able to experiment with these settings, but they are shown below. Sony did acknowledge that the jaggies we detected on both HD and SD video—an interlacing artifact—were endemic to the set.
Cinema 1 Picture Mode for HD 1080i Content
Sharpness: 50
Reality Creation: Manual
Resolution: 5
Cinema 1 Picture Mode for SD 480i Content
Sharpness: 50
Reality Creation: Manual
Resolution: Min
