Sharp XV-Z20000 initial impressions and numbers

I just updated my review of the Sharp 1080p projector, and I thought I’d post some numbers and impressions based on about one year of use.

Throw and lens shift

The Z20000 is a relatively long throw projector 1.85-2.5/1 screen width. It has a very flexible vertical lens shift of 1 screen height, which means that it is equally well-suited for a ceiling or a shelf installation. By default, the lens aligns with the bottom of the screen for a shelf mount and at the top of the screen for a ceiling mount. Use lens shift to adjust from there. Lens shift works like a charm, but it does reduce light output a little, so use it sparingly.

Brightness and contrast

Measuring brightness is not simple with the projector because it has so many different modes. Like the Z12000, the lamp has a Bright and a Eco + Quiet mode and it has an Iris with 3 settings, High Brightness, High Contrast, and Medium. It also has 5 Gamma presets and several Picture Modes. The default Picture Mode is Standard. With Standard gamma and Standard Dynamic Range selected in the Options menu, the gamma measured a nearly perfect 2.2 across the entire range. This offered exceptionally good shadow detail.

Initially, I did all my watching in the Eco + Quiet lamp mode, and subsequently changed to the Bright mode (with somewhat higher fan noise) as the lamp aged.

All measurements are post-calibration in the Standard Picture mode. The lumens figures were obtained with the projector set to minimum throw. For maximum throw, multiply by 88.6%. Throw range has a minimal effect on contrast.



I measured the ANSI contrast at 560:1 in the High Contrast mode and 489:1 in the Medium mode. Greg Rogers' review published in Widescreen Review reported 845:1.

The Medium Iris setting was my preferred setup. It offers a good compromise between brightness and contrast. Others with smaller or higher gain screens will no doubt prefer the High Contrast mode, as it certainly offers better contrast performance.

Color performance

The Sharp XV-Z20000 displayed the best gray scale performance out of the box that I have ever measured. By default, the color temperature is set at 7500k. I changed this to 6500k and then measured the gray scale.



As you can see, the default gray scale is very good. I entered the custom Gamma mode and tightened this up somewhat and achieved virtually perfect performance. I did run across one problem. The gray scale controls for the low end of the scale did not seem to work. The controls are there and they respond to input, but they have no effect on measured performance. Fortunately, the performance was so good to begin with, I wasn’t too concerned with this. (See more on this below)

The XV-Z20000 includes a Color Management System that allows calibrators to adjust the primary and secondary colors to industry standards. By default, the Sharp showed a familiar color profile: oversaturated red and very oversaturated and somewhat yellowish green.



After making adjustments in the CMS, I was able to get a nearly perfect color gamut.



This is a must-have feature for obsessive tweakers. In fact, this projector is a calibrator's dream. It was particularly gratifying to get accurate green for a change. I don’t think I’ve ever worked on a digital display that didn’t exaggerate green. Sharp’s color management system works exactly as advertised, and I was much more impressed with it than I was with a similar system provided by Optoma’s HD-3000 external processor (the HD81 uses this) and by the one offered on the Epson TW1000.

The Value parameter in the CMS works as a color decoding adjustment, but the color decoding was nearly perfect out of the box (no red push here), so minimal adjustment was required.

Here are some before/after screenshots that show what the CMS can accomplish. The original is on top. The CMS processed image is on the bottom.





Optics

This projector also has the best optics I have ever encountered. The images show great clarity and precision with no evidence of chromatic aberration. There is ZERO color fringing even at the extreme corners of the screen, as shown in this image of the upper right hand corner of my screen.



In addition to vertical shift, the lens has manual zoom and focus. For convenience I would have preferred a motorized system, but on the other hand the lens assembly has a very solid and well-engineered feel. Sharp obviously devoted a lot of resources to the optics. This is an aspect of projector performance that gets far too little attention. Really good optics are expensive, unsexy, and thus difficult to market. Nonetheless, a really good lens can make a profound difference to the quality of the image with respect to sharpness and clarity.

White field uniformity was also very good, a hallmark of the DLP technology. The brightness varies by no more than 10% across the entire screen. I measured color temperature at about 150k lower on the right side of the screen. This is a measurable difference but not one that’s visible to the naked eye (at least not to mine), even when looking at a uniformly white 80IRE field.

Processing

Using the Silicon Optics Benchmark DVD and HD DVD, I was able to test the XV-Z20000's standard definition and high definition processing. It is excellent. It passes full resolution for 1080i film and video sources. It also correctly deinterlaces 1080i for film-based sources, applying 3:2 pulldown. For 480i sources the 20K passed all of the tests the DVD threw at it. It took about a half a second to lock on to the racetrack scene contained in the film detail test, which evaluates whether the display's processor can correctly apply inverse telecine for standard definition film-based material.

One test the XV-Z20000 did NOT pass is the deinterlacing torture scene at the beginning of Chapter 8 of the MI:3 Blu-ray or HD DVD. The staircase scene reveals some moire and considerable line twitter. Interestingly, this scene looks better using 1080i output from the Sony BDP-S1 Blu-ray player than it did with the Toshiba XA2 HD DVD player, indicating either some difference in the way the 2 players output 1080i material or in the respective masters. Even the 1080p output of the XA2 showed problems. However, the 1080p output of the Sony was artifact free.

After its initial release, Sharp made a firmware update available that allowed the projector to accept 1080p/24 input, which it would output at 1080p/48. I am not particularly sensitive to the judder artifact that 1080p/60 supposedly induces so I cannot evaluate its success at eliminating it. However, the update did resolve some bugs, and the 480i processing seems improved insofar as prior to the update the XV-Z20000 did not pass the film detail test on the Silicon Optics Benchmark DVD. This firmware update is thus a valuable addition for this reason alone.

In the Dot by Dot mode, the projector maps the incoming signal directly to each pixel without any scaling. It is in this mode that 1080i and 1080p material look the sharpest. Using a 1080i multiburst signal from the Accupel test signal generator, I was able to verify that the HDMI input resolved the full limit of detail offered by HD sources. The component inputs, however, did not. This was apparent even with regular HD program material, which looked visibly softer over component than when fed the same material via HDMI.

Perhaps the best praise I can offer for the XV-Z20000's processing is that I no longer felt the need to use the Vantage-HD Realta external processor that had proven itself so valuable at improving the performance of the Optoma H79.

Problems

I ran across a few problems.

• As I mentioned above, the RGB Bias adjustments for tweaking the low end of the gray scale seemed not to function. As it turns out, they do work, it is just that they work on only 10 IRE or below signals. This is a very strange--in my experience unique--implementation of a gray scale control that I fear will confuse a lot of owners and calibrators. Fortunately, you can use the RGB All Colors adjustment instead to adjust the gray scale in the 20-40 IRE range.
• Page 40 of the User Manual, the page in which these controls are discussed, is reproduced in French! This is obviously a printing error.
• The fan noise was a little higher than I would have liked and the light output was a little lower than I would have liked.
• Some have reported seeing more rainbows on the Z20000 than they see on other high-end single-chip DLPs, but since I am insensitive to RBE, I cannot evaluate this.

Subjective impressions

My first thought upon seeing the image from the projector was: “I need a bigger screen!” The high pixel count really would shine more on something bigger than my 92-inch DaLite matte white screen.

I subsequently upgraded to a 100-inch Stewart StudioTek. I was surprised by the degree of improvement in image quality. First, I was able to get a larger image that did not, because of the true 1.3 gain, sacrifice any brightness. Second, this screen is able to resolve greater detail than the DaLite, so now the Z20000 seemed even sharper and more revealing than before. The DaLite imparted a subtle amount of screen grain to the image that disappeared with the StudioTek. This screen is a great match for this projector. If you wanted to go up to a larger 110" screen I would recommend the Stewart Ultramatte 150. For larger screens still, many users have reported very positive experiences with the DaLite High Power.

I spent a couple of hours looking at DVD, HD DVD, and broadcast (HD and SD). My only recent frame of reference was the SIM2 HT3000. They looked very similar, except that the Sharp seemed a little, well, sharper, a perception that I attribute to its superior optics. After tweaking, the Sharp certainly offered better measured performance. It provides better color performance and higher contrast. In fact, the only area of measurable performance where I thought that the Sharp needed improvement is light output. The High Brightness mode is quite bright, but the black level is unacceptably elevated. The High Contrast mode offers great black level, but it is too dim for my taste. The Medium mode is just about right, but by 670 hours I was getting only 9.5 fL so I had to get a new bulb. I was surprised how little visible change I saw after installing the new bulb, so I think I could have gone until at least 800 hrs. before the lack of brightness really began to impact viewing pleasure.

The biggest improvement of the 1080p pixel count is smoothness, depth, and a difficult-to-describe improvement in realism. The illusion that you are looking through a window viewing actual events is better conveyed with 2,073,600 pixels than it is with 921,600 of the little buggers. There’s clearly more picture information on the screen, which is evident even with standard DVD. However, I think that I was even more impressed by the Sharp’s optics, color, and gamma performance. The CMS system works wonderfully and it really shows on the screen.

There is actually one downside to 1080p projection, at least when combined with a unit with optics as good as the Z20000. As audiophiles have long known, one of the unintended consequences of using a very high-end sound system is that it is ruthlessly revealing of the source material. That is also true with video. The Z20000 did a much better job of revealing differences in quality between a variety of HD broadcast sources, differences that were not as obvious with my Optoma H79. I saw all sorts of flaws in these images that I hadn’t noticed before. However, with clean sources, such as HD DVD, the results are startling.

Despite the advantages of 1080p, resolution is only one measurement of performance and arguably not even the most important one. If the Z20000 had not also provided exemplary colorimetry, optics, and contrast, then the 2 million pixels wouldn’t matter much. Furthermore, to the extent that it does matter you need a large screen to appreciate it. I wonder how much obvious benefit users can expect from 1920x1080 50 and 60” rear projection displays.

Conclusion

In response to market pressures, Sharp has recently reduced the price on the XV-Z20000 to $6999, which is a truly great deal for a projector of this quality. Jason has these units in stock and can quote you a very attractive price, so if you are in the mood to upgrade you should give him a call.

To sum up, the XV-Z20000's strengths are:

• World-class color available with the built-in full-featured CMS.
• Excellent optics with zero convergence and chromatic aberration problems.
• About the highest ANSI contrast available.
• Very good native on/off contrast in the Medium or High Contrast modes.
• Nearly perfect gamma provides excellent shadow detail.
• Very good processing, especially for high definition sources.

Its weaknesses are:

• Relatively low light output in High Contrast and Medium modes limits screen size.
• Potential problems with RBE for those sensitive to this artifact.
• Not appropriate for short-throw installations.
• Lower on/off contrast than what's available from some similarly priced alternatives.
• More fan noise than many competitors in this price range.

I have some good news to report. It turns out that the lens shift affects brightness and contrast more than I thought. When I took my initial measurements it was with the Z20000 on a top shelf. I put it there only because that's where the H79 had been and I was lazy. Tonight I moved the projector down 15" to a lower shelf. This required much less lens shift. I then took my lumens and contrast measurements again, and this time I included measurements from the economy mode as well. I have updated the table in the original review. Both lumens output and CR improved significantly. With zero lens shift they would improve even more. Too bad that this convenience feature has such a performance penalty.

To answer some questions:

The fan noise is worse than the H79, but better than the Sony HS20. I haven't heard the Z12000. I would rate the noise level low, but noticable. The H79 was virtually silent. I hear no color wheel sound at all.

I am using the Vantage HD scaler. It ALWAYS makes HD look better. I haven't checked out the Z20000 without it, but I'm sure it wouldn't look as good.

The Auto Contrast is a mystery to me. It certainly isn't an auto iris and, like Jason, I couldn't see that it had any measurable effect on CR.

OK, take a look at the table now. My attempts to use fancy formatting was a flop.