I decided to do just that when I played BioShock Infinite at 4K resolution. The results were excellent. High-end PC gaming is the main reason I can see wanting to buy a 4K display like this. At about 4 feet, the closest I could hack it, the image was sumptuous, with an unreal sharpness to the graphics, like the clothing and textured straw hat of the guy getting a shoeshine on the Options page. I flipped back and forth between that and 1080p and the difference was obvious -- 4K looked much smoother and more detailed, while 1080p from this distance appeared with jagged edges and much less overall crispness. Yes, certain graphics tricks like anti-aliasing can help those issues, but it's tough to argue that from 4 feet, games that can take advantage of 4K resolution will look better than at 1080p.
I did see some slight tearing at times but overall the graphics held up well. Gameplay was fine with an Xbox 360 controller as long as I avoided the Very High graphics preset -- no matter, High moved along nicely and still looked great. When I went to use the keyboard and mouse, however, it was tough to get the sensitivity right, and the game either felt a bit too jerky or a bit too laggy; overall, however, I could get used to it and it was still totally playable. For fun, I ran the built-in benchmark and the game-PC combo came in at 41 frames per second at that resolution. In case you're curious, the same system reached 130fps with 1080p resolution.
The TV also seemed to lose the signal and go to black more often than other sets I saw, for example when switching inputs and sources etc.
The rest of the review comprises the results of tests I performed with 1080p sources. I used CNET's standard TV testing procedure, placing the Seiki in a lineup of comparable TVs.
Click the image at the right to see the picture settings used in the review and to read more about how this TV's picture controls worked during calibration.
|Comparison models (details)|
|Vizio M3D550KD||55-inch LED|
|Panasonic TC-P55ST60||55-inch plasma|
Black level: The Seiki was the worst in the lineup in this category. As I mention in my calibration notes, the TV was unable to achieve decent shadow detail unless I increased the brightness control quite a bit, which really hurt black levels. Watching Chapter 7 of "The Amazing Spider-Man" for example, where Peter wanders back alleys foiling crimes, its letterbox bars and black areas were brighter (worse) than any of the other sets' by a wide margin. Reducing that control to the level dictated by my standard test pattern delivered markedly deeper blacks, but it also obliterated details like the bricks behind the mugger, the trash bags and the graffitied wall. All were either invisible or appeared so murky that they couldn't be made out.
A sort of unhappy medium can be achieved by using the Movie preset, whose black levels were better but still the worst in the room, and whose shadow details were still obscured, but better than what I achieved with the standard test pattern.
Color accuracy: The Seiki was again the worst in the lineup. Skin tones, from Peter's face in the dark as he experiments with web-slinging devices (52:03) to Gwen's face when talking to Peter in the well-lit school hallway (36:33), looked too ruddy, with that almost sunburned effect. Bold primary colors, like the trees and grass outside the school (32:59) or Spidey suit itself and the glare of the police lights (57:14) also looked too vibrant and unrealistic. Black and near-black areas, as usual for an LED with light black levels, were also tinged blue.
Overall, however, as evinced by the mostly Good results in the Geek Box below, color wasn't nearly as inaccurate as it could have been. In fact, color was the Seiki's strongest suit.
Video processing: As a 4K TV the Seiki has to convert all sources, including 1080p, to its native panel resolution. That conversion is by nature imperfect. In the Seiki it proved more imperfect than usual, but I'm sure it could be worse. That said, you'll likely achieve better results with an external scalar or a Blu-ray player with a 4K output.
Here's where the Seiki's relatively small screen size is an asset compared with the 84-inch behemoths from LG and Sony: upconversion errors (like all artifacts) are more difficult to spot in smaller screen sizes. According to test patterns the Seiki didn't resolve every line of a 1080p signal; there was some softness in a multiburst test pattern from my generator. On the other hand the highest-resolution lines from the 1080p, 50-inch Toshiba showing the same pattern became indistinguishable at a seating distance of around 64 inches to my eye. Even from my normal seating distance of about 90 inches, the finest of lines on the Seiki looked less sharp than on the 1080p set with 1080p material. Switching over to 720p, however, those differences were much less obvious.
As always, however, differences in sharpness and resolution are much more difficult to spot with real program material than with test patterns, especially stationary ones. I used "Samsara," a spectacularly detailed Blu-ray with plenty of slow shots that allow comparison. In most scenes from my normal seating position the Seiki didn't look any less detailed than the others (the best comparison again being the like-sized Toshiba). In some areas I did see a bit of softness, for example in the rocky hillside under the town at 8:52 or the grains of color in the monks' mandala (11:31) the Seiki did look a bit less sharp, but it was close enough that I wouldn't notice the Seiki's softness outside of a side-by-side comparison. As usual, the differences in color, black level, and other areas were much more obvious than differences in apparent sharpness and resolution.
The Seiki passed our 1080p/24 cadence test, delivering the proper smooth but not too smooth motion expected from film-based sources when I watched the Intrepid sequence from "I Am Legend." That's what I expect from any 120Hz set not using a dejudder/smoothing mode.
The Seiki did produce a worse motion resolution test result, with more visible blurring, than the other 120Hz sets in my comparison (the Sharp and the Toshiba) when I engaged their 120Hz modes. The numerical result of 300 lines of motion resolution is imperfect, however, because the test pattern is a 1080p source, so the Seiki first upconverts that to 4K, which can interfere with the results.
But when I watched material designed to show motion blur, such as the stationary shots of a moving metronome, swinging hammock, and passing license plates from the FPD Benchmark disc, the Seiki was the blurriest in the room, and significantly worse than the other 120Hz TVs, not to mention the plasma and 240Hz Vizio, which were even better. As usual I looked for obvious instances of motion blur in program material, for example the fight scene and web-swinging of Spider-Man, and it was more difficult to discern.
Uniformity: The screen of the Seiki was less uniform than that of the other LED sets, with markedly brighter sides than the middle. The issue was most visible in letterbox bars, but any dark scene, such as the cityscape at the beginning of Chapter 7 in Spider Man, also showed it. From off-angle the Seiki's dark areas lost fidelity quicker than any of the others aside from the Vizio, and the bluish tinge became even more apparent.
Bright lighting: The matte screen of the Seiki handles reflections well, and did a solid job of preserving whatever black levels the TV could produce. Its light output isn't as high as that of some LED TVs -- about 70fL in Dynamic mode compared with 94fL on the Toshiba, for example -- but still plenty for just about any room lighting.
|Geek Box: Test||Result||Score|
|Black luminance (0%)||0.007||Good|
|Avg. gamma (10-100%)||2.34||Average|
|Avg. grayscale error (10-100%)||3.089||Average|
|Near-black error (5%)||0.421||Good|
|Dark gray error (20%)||3.56||Average|
|Bright gray error (70%)||2.915||Good|
|Avg. color error||3.546||Average|
|1080p/24 Cadence (IAL)||Pass||Good|
|1080i Deinterlacing (film)||Fail||Poor|
|Motion resolution (max)||300||Poor|
|Motion resolution (dejudder off)||N/A||N/A|
|Input lag (Calibrated mode)||38.3||Good|