Samsung HL-S5688W 1080p DLP Rear Projection TV Technical
This Samsung HL-S surprised me with more technical competence than I expected. Video bandwidth at 1080i was as good as I've seen with no apparent roll-off. The most demanding part of a multiburst pattern did show some moiré, but if overscan was disabled (in the service menu), it became textbook perfect with every line sharply delineated. Taking overscan out reduces a reasonable 5% of overscan down to 1%, which unfortunately allows occasional broadcast transmission artifacts (white dashed lines) to appear at the very top of the picture—mostly during commercials. Some people may not tolerate this. That's why Samsung dialed the overscan in.
Resolution, using my Sencore VP-403 pattern generator's Focus test pattern, was outstanding, and uniform all across the screen. Unlike the Mitsubishi, which could do this trick too, the Samsung actually maintained superb resolution with real-world color pictures. There was a bit of pincushioning on the sides that made the black bars beside 4:3 pictures even more annoying. It's not correctable, but with 16:9 operation, it's not noticeable.
Light output was a blinding 157 foot-Lamberts in Dynamic mode. Movie mode (with Contrast set to maximum) measured 128fL. Black level in Movie mode measured .03fL—slightly higher than the Mitsubishi and JVC, but subjectively very similar. With the Contrast at maximum in Movie mode the contrast ratio measured 4,266:1 (very good). ANSI contrast, using a checkerboard pattern and a direct contact light meter, measured 187:1 (only fair).
The pre calibration color temperature was consistently bluish at any setting. Warm 2, the lowest setting and the setting closest to the 6500K industry standard, was over 8000K even in Movie mode. Warm 2 in other modes was over 10,000K—as bluish as some sets are in their coolest mode. With five available color temperature settings, why would Samsung make every one of them too cool? Excess blue in the grayscale always hurts color accuracy and the beauty of secondary colors. Grayscale tracking was fair (going even more blue in dark scenes). Fortunately, Samsung provides the required controls in the service menu to tweak all of this to perfection. As a note, other Samsung HL-S sets I've calibrated were as low as 6000K in Warm 2—too reddish—but some have been nearly perfect, so there's some variation among samples. Grayscale uniformity across the screen was excellent—something that usually eludes the LCoS sets (Sony and JVC).
The pre-calibration color primaries in Movie mode were nearly spot-on HD standards. Secondary colors were not, with yellow being greenish-yellow and magenta being too blue. The color decoder (balance of red, green, and blue), however, was accurate with both HDMI and Component sources.
For Movie mode, I was able to easily adjust all six colors to HD textbook perfection (earlier Samsung's were more problematic) with my Lightspex spectroradiometer. The other two instruments capable of doing this accurately are the Photo Research PR650 (and its newer replacement) and the Minolta CS-200. I tried several other lesser test instruments on hand just for kicks, but the results weren't accurate and often poorer the factory setup. I could have picked SMPTE C (NTSC) standards (which are still close to HD standards but more accurate for film-based material), but either color standard is far from the intentionally inaccurate secondary color points Samsung dialed in.
Modes other than Movie mode have a separate setup page in the service menu and their colors were all factory set to exaggerated, oversaturated values. Yellows and greens, in particular, were excessively oversaturated and magenta was far too blue, apparently to make the picture look more "colorful".
Tests using the Silicon Optix HD Benchmark test disc (on HD DVD) showed the Samsung's deinterlacer to be a real winner except for its inability to perform 3/2 pulldown at 1080i. The resolution loss test was textbook perfect and the jaggies tests weren't far from it. With 480i DVDs, 3/2 pulldown was done quite well.
Some of my ISF colleagues have experimented with placing a neutral density filter over the lens of recent Samsung DLP sets. Although a good one will cost over $100, the benefit would be a reduction in the unnecessarily high peak light output (to or less) and at the same time a deepening of black level to perhaps best in class—a welcome trade for dark room viewing. Cleaning the lens on these sets yearly is another tweak with sometimes stunning benefits. And, of course, replacing the bulb before the picture gets dim and washed out is the best tweak of all. Disabling the automatic feature of the iris and closing it down fully in the service menu will reduce light output significantly but it will not lower black levels.