Samsung's Bread and Butter HDTV - With Some Jam on Top Page 3

Test Bench

Color temperature (Night setting, ACC and ACM off, before/after calibration

IRE

Temp Before

Temp After

20

7,017

6,492

30

7,036

6,454

40

7,126

6,464

50

7,064

6,436

60

7,091

6,437

70

7,199

6,531

80

7,315

6,555

90

7,264

6,674

100

7,151

6,628

Brightness (100-IRE window): 74.3 / 62.1 ftL Primary Color Point Accuracy vs. SMPTE HD Standard

Color

Target X

Measured x

Target Y

Measured y

Red

0.63

0.629

0.34

0.336

Green

0.31

0.310

0.60

0.565

Blue

0.155

0.154

0.07

0.074

Cyan

0.225

0.225

0.329

0.329

Magenta

0.321

0.320

0.154

0.155

Yellow

0.419

0.419

0.505

0.487

Grayscale for the Samsung's Movie preset initially tracked slightly blue in the key 30- to 100-IRE brightness range, measuring as much as ±815 K off the 6,500-K standard at 90 IRE. Calibration in the set's White Balance user control menu brought tracking to ±174 K from 20 to 100 IRE, and for all but the brightest windows it was ±63 K - very good performance. Note that, while it's not desirable, it's not unusual for LCD TVs to show some lack of grayscale linearity at the brightest or darkest extremes of the brightness scale.

The primary colors measured very close to the HDTV standard before any tweaking and were brought into near-perfect alignment with the Color Space controls in the advanced user menu. Before color point calibration, I had done a standard calibration of the basic controls - including contrast, brightness, sharpness, color, and tint - after which the color decoder test revealed +5% red error, +2.5% green, and 0% blue. However, the picture still showed a subtle but noticeable red/orange lean that particularly affected skin tones, and greens looked a bit undersaturated . After I tweaked the color points, the image looked slightly oversaturated (and especially so in the reds). So I turned down the overall color saturation to affect a more natural color balance, which in turn desaturated blue as measured by the decoder check (for which the final result was -5% red, +10% green, and -10% blue). But the blues and greens looked just fine in all the high-def material I watched, and the postcalibration color with Blu-ray Discs like I Am Legend and good-quality HD cable shows struck me as quite natural and a big improvement from where I'd began.

Screen uniformity for this Samsung was very good, with minor hot-spotting showing up only on a very dark 20-IRE window. The set's viewing angle was somewhat narrow, with the image showing minor but noticeable loss of contrast and color accuracy when moving as little as 20° off center in either direction - the equivalent of moving only a couple of feet from the center seat of a couch positioned 9 feet from the screen. However, after this initial dropoff, the image remained stable for a fairly wide window - so although off-center viewers don't get a perfectly optimal image, the picture doesn't shift noticeably as you move around the room. On a related note, the screen's high-gloss surface, which Samsung uses to affect better contrast, functions like a mirror with ambient light and could be a problem in bright viewing conditions.

The LN46A750 fully resolved 1080i/p and 720p signals via HDMI, and 1080i/720p via component video. It performed well on the Silicon Optix torture tests for deinterlacing HD and SD signals. The TV also performed well with the moving-resolution tests on the FPD test disc, which present graphic characters and objects moving rapidly across the screen (such as close-ups of a scrolling map or cars whizzing by). The 120-Hz processing cleaned up blurring of these images quite effectively, and unlike the processing on some other sets I've seen, it never introduced any easily detectable artifacts, even when set to High. Likewise, the Digital Noise Reduction circuitry cleaned up random video noise on standard- and high-definition material without further softening the image, but it seemed to do little for mosquito noise that afflicts the edges of objects, and I've seen HDTVs that did a better job in general with standard-def programming.

ARTICLE CONTENTS

X