That side-mounted input bay includes the standard selection of jacks, although many of them double- or triple-up. Inputs 1 and 2 are dedicated to HDMI ports; Input 3 offers component video, S-Video or composite video; Input 4 a choice of component or composite video; and Input 5 (located on the front panel) a choice between VGA-style computer and composite video. Unfortunately, the PC input is limited to just 1,024x768 resolution.
The JVC HD-58S998 turned in a good overall performance, with one big exception. Our review sample suffered from fairly severe geometry errors, which, among other issues, caused horizontal lines to appear curved upward toward the edges. Otherwise we appreciated most aspects of the set's picture quality, including its solid black-level performance and relatively accurate color.
As with most HDTVs, the JVC HD-58S998 can get blindingly bright, so our first step in adjusting it for our darkened room was to attenuate the light output around 35 FTl. We ended up reducing the Picture and Iris controls almost all the way to make that happen, but the end result was a big improvement in black levels. In addition, we adjusted the Warm color temperature preset via service menu calibration. Prior to adjustment, the Warm preset came relatively close to the 6,500K standard, although it tended to get bluer the brighter the image became--a characteristic that calibration couldn't completely fix. As a result the set's grayscale tracking wasn't as good as we would have liked. For details see the geek box below, and for a complete look at our user menu settings, check out the Tips & Tricks section.
We didn't have any big-screen rear-projection TVs around to compare to the JVC, so we settled on a couple of competitively priced plasmas, the Panasonic TH-58PX600U and the Vizio VM60PHDTV, along with our reference Pioneer Pro-FHD1. We chose the Blu-ray of League of Extraordinary Gentlemen for our test material, played on a Samsung BD-P1000 at 1080i resolution.
In dark scenes, the JVC more than held its own against the competition. As Quatermain and Reed ride toward League HQ in a darkened carriage, for example, the black of the letterbox bars below the image, as well as the leather of the seat and the shadowed post along the window of the carriage, appeared nice and inky. Black areas weren't quite as dark as on the Panasonic plasma but they were definitely darker than on the Pioneer. We did notice that the JVC didn't display the best possible shadow detail, however. Regarding the leather seat, for example, we couldn't discern the darkest details unless we increased the brightness control to an unacceptable level, so we ended up having to sacrifice some shadow detail to keep those inky blacks.
We appreciated that the relatively accurate grayscale kept skin tones looking natural, from Quatermain's ruddy face to the delicate, pale skin of the toothsome Mina Harker. We were able to achieve decent saturation while staying realistic, a testament to the JVC's accurate color decoding, although colors didn't have quite the same intensity and richness that we saw on the Pioneer and Panasonic plasma TVs. We also saw that primary-color accuracy, especially the yellowish greens, left a bit to be desired.
Like many of the HDTVs we've tested recently, the JVC HD-58S998 failed the test for proper 1080i deinterlacing of film content found on the HQV HD DVD. In other words, we'd recommend using the 1080p setting on your source, if possible, because you may see some artifacts in 1080i mode. We noticed some moire, which appeared as faint concentric lines in the stands, during a pan over the stadium in HQV. We checked for similar effects during League, for example in the pan around the meeting room at League HQ or the one following the Nautilus as it slipped into Venice, but we couldn't discern any problems. The JVC resolved every detail of still 1080i sources, according to our test pattern generator, and the image looked as sharp as we expect from a 1080p HDTV.
The biggest problem we had with the JVC's picture quality had to do with its uneven geometry. Ideally you want a grid displayed on a TV screen to have perfectly straight horizontal and vertical lines, and with most microdisplays and all flat-panel sets, they come pretty close. With the HD-58S998 we reviewed, however, it was a different story. The most egregious issue was with horizontal lines in the top two-thirds of the screen, which curved upward toward the sides, the left more than the right, for an effect that kind of resembled a crooked smile. Other areas of the grid looked funky, too; vertical lines toward the edges bowed inward at the top and slightly outward toward the top middle, in contrast to the screen's lower third, which was mostly even.
The effect of the geometry errors depended largely on program material. In some shots, such as when the top area of the screen was occupied by clear sky, we couldn't notice the flaw at all. In others it was almost comically apparent; full-screen program guides, menu systems, onscreen text, and pans over anything with a straight horizontal edge were the typical offenders. Of course the letterbox bars on 2.35:1 movies, such as League and many others, also make the upward curve on the edges quite obvious. We mentioned the issue to JVC's representative and he said that the set uses a concave mirror that's responsible for the errors.
Brightness and color uniformity across the screen, on the other hand, were excellent. Looking at flat gray fields, we didn't see noticeable discoloration from one part of the screen to another, and brightness drop-off from the middle of the screen to the edges was average for a rear-projection HDTV. Unlike typically more-uniform flat-panel sets, projection-based displays are always brighter in the middle and get darker toward the sides.
We did notice some stationary screen grain, which looked kind of like tiny sparkling flecks and was visible primarily in white areas, like the snow outside the fortress in League and in flat fields like sky. Although more noticeable on the JVC than on many other RPTVs we've tested, the grain wasn't too distracting in most scenes, especially with our darkened picture settings, but of course the plasmas didn't exhibit any grain. On the other hand, many plasmas, including the Panasonics, exhibit some false contouring, whereas the JVC was basically free of such artifacts.
With standard-def tests the JVC turned in a decent--but by no means spectacular--result. On the moving diagonal lines from the HQV test DVD, played at 480i over the component video inputs, we noticed more jagged edges along the moving diagonal lines and in the stripes of the waving American flag than we'd like to see. The set did resolve all of the detail of the disc, although we detected some flicker in the color-bar test pattern. We also noticed that the JVC only engaged 2:3 pull-down detection correctly when the Natural Cinema setting was in the On position; it failed to implement 2:3 when in the default Auto position, introducing moire into the grandstands. We did appreciate that the JVC's noise reduction did an effective job of cleaning up the snowy-looking motes on the HQV disc's skies and sunsets, although we'd recommend avoiding the Max setting since it softened the image too much.
We also checked out the JVC's ability to handle PC signals via its VGA input, and the results were disappointing. First off, the maximum resolution is just 1,024x768, which is laughably low for a display with 1,920x1,080 pixels. Second, the image was relatively soft--according to DisplayMate, the set only resolved about two-thirds of the lines in both horizontal and vertical axes, which isn't a big surprise given that the TV has to scale 1,024x768 up to fit the screen. Unfortunately it doesn't fit the screen all that well anyway. There were black bars to either side and above and below, and no way to expand the desktop to occupy the entire screen.
|Before color temp (20/80)||6682/7142||Good|
|After color temp||6494/6834||Average|
|Before grayscale variation||+/- 509K||Average|
|After grayscale variation||+/- 214K||Average|
|Color of red (x/y)||0.653/0.325||Average|
|Color of green||0.295/0.689||Poor|
|Color of blue||0.143.0.052||Average|
|Black-level retention||All patterns stable||Good|
|Defeatable edge enhancement||Y||Good|
|480i 2:3 pull-down, 24 fps||Y||Good|
|1080i video resolution||Pass||Good|
|1080i film resolution||Fail||Poor|