SIM2 Grand Cinema HT 300 E-LINK DLP Projector Testing and Calibration

Testing and Calibration

Thomas J. Norton

On 720p resolution test patterns, the SIM2 Grand Cinema HT300 E-LINK reproduced a full 720 lines per picture height. The Faroudja DCDi deinterlacer built into the SIM2 performs its function as well as any deinterlacing technology on the market, and better than many. In any event, neither I nor FM found any reason to complain about the SIM2's built-in deinterlacing and scaling, on those occasions when we used it. Both of us did much of our viewing and testing with a 480p or 720p DVI or HDMI source, which bypasses the projector's deinterlacer.

My DVD testing was conducted with a DVI output from the Marantz DV8400 DVD player (using an adapter cable into the SIM2 video processor's HDMI input). For reasons I was unable to determine, the SIM2 processor refused to accept an HDMI output (at either 480p or 720p) from the Pioneer DV-59AVi DVD player, though that output works flawlessly into other projectors.

My measurements were divided into two segments: before FM and after FM. With the projector fresh out of the box, before FM started his viewing, I measured a peak contrast (peak full-field white vs. video black) of 1472 (11.78 foot-lamberts peak, 0.008fL video black) on my 80-inch-wide Stewart FireHawk screen. The pseudo-ANSI contrast (using just the four center squares on a 16-square checkerboard) measured 215.

With 12 different gamma settings on tap, the user could be understandably confused as to the best option. I found that 10 (very CRT-like, though a bit lighter at the low end) worked well for me.

The red, green, and blue color points on the SIM2 were not very different than we've measured on most digital projectors. Green was very close to spot-on. I agree with FM that the SIM2's greens look more natural than those from many competing digital displays. The SIM2's reds looked noticeably richer than from some projectors, and indeed the red point turned out to be positioned slightly deeper into red than is strictly correct. Blue was a little closer to pure blue than the purplish blue required by both the NTSC and ATSC standards. The measured red and blue deviations did not affect the SIM2's subjective color quality. And these color points did not drift significantly from the first time I measured them, with the projector fresh out of the box, to the time I get the unit back after FM finished his review, with about 200 hours on the lamp.

I wish I could say the same thing about the color-temperature measurements. If you're a calibration professional or a power videophile with good test gear, you may find the following section useful. For those who may not need or want such a detailed explanation, the section following it, "In Brief," states the important points concisely.

Calibration Follies
When I first tested the projector, I did not have access to the service menu with its full set of six color-temperature calibration controls: drive (high end) and cut (low end) for red, green, and blue. But even without that, a grid of 34 different color-temperature options is available to the user—so many options that you would have to be extremely lucky to pick the correct one by eye. After testing a number of them, I found that #22 on our sample was respectably close to the D6500 standard across the brightness range. The result is shown in the Before curve of Fig.1, taken with our Photo Research PR-650 colorimeter on a Stewart FireHawk screen.

Fig.1 Before: Low lamp hours, setting #22 on User color temperature After: 200 hours on lamp, setting #30 on User color temperature plus service-level calibration

While FM was working on his review, we obtained the access code for the service menu to enable a more detailed calibration. After the viewing tests were complete, the projector came back to me for further measurements. When I set up the projector again on the Stewart FireHawk screen, the #22 setting was no longer anywhere near correct; in fact, the color temperature at this setting had drifted upward to between 7700K and 8000K.

What had happened? I blame the projection lamp, which now had approximately 200 hours on it. I spot checked all of the 34 available settings (at 60IRE). Only three of the settings, at just below 7000K, produced a result reasonably close to the D6500 standard. I chose setting #30 as the best candidate from which to start a more complete service-level calibration. (This won't necessarily be the best choice with all samples of the projector.)

At this point, I entered the service menu. Unfortunately, you can't do a calibration in the menu from an external source, only from the projector's internal 20 and 80 IRE full-field test patterns. I prefer to calibrate with test patterns from my most frequently used source (the best way to get the most accurate color in a specific installation), which in this case was the Marantz DVD player. Fortunately, it was relatively easy to jump into the service menu, make a change, jump out to take a reading on the Photo Research from a DVD test pattern, and then jump back into the menu for further changes. Tedious, yes, but not a trigger for calibration-rage.

What was rage-inducing, however, was the touchiness and unpredictability of the individual calibration controls. Sometimes, increasing a given control by one step would increase the measured value, while at other times, the same change would decrease it. If you increased or decreased a control from a given setting, it would not necessarily produce the same measured result when you returned it to its original setting. For example, at one point, I obtained the following three sets of x,y color coordinates with the calibration controls at precisely the same settings:

x=0.302, y=0.350

x=0.301, y=0.315

x=0.310, y=0.328

The only action I took between these measurements was to change one of the calibration controls slightly, then change it back to its original position before taking another reading.

Once I discovered this quirk, I was able to work around it—tediously—by making very small changes between measurements, edging gradually toward the desired D6500 standard. Finally, I obtained the After result shown in Fig. 1.

But based on the variations I experienced initially, I decided to check this setup for stability. I shut the projector down after taking the measurements shown, let it cool off for an hour, then turned it on again. After a brief warmup (about 30 minutes), I re-checked the measurements.

I wish I could say they hadn't changed, but they had. The gray scale now had a slightly pinkish cast across the mid-brightness region, which could only be eliminated by turning down the contrast to a very low level, a level too low too produce an acceptably bright image (even on my 1.3-gain, 78-inch-wide Stewart Studiotek 130 screen—a screen identical to the one FM used).

More significantly, the color temperature had dropped by about 500K across the entire brightness range, with a measurable excess of red. The Before curve shown in Fig. 2 is identical to the After curve in Fig. 1—that is, the best calibration I was able to achieve, and an excellent one at that. The After curve in Fig. 2 shows how this was altered simply by this turn-off/cool-down/turn-on cycle. No other actions were performed on my part between these two sets of measurements, aside from a slight adjustment in the contrast control to keep the whites out of clipping (which made no difference in the color-temperature readings). I also entered and left the service menu once to confirm that the calibration settings from the last set of measurements had not mysteriously changed (they had not), but I made no alterations to these settings.

Fig. 2 Before: Same as After in Fig. 1 After: Same calibration settings as Before (in this figure) after cycling projector off, then on Color Temperature numbers

I took one last shot. I turned the projector off yet again, this time for several hours, then turned it back on. Or I should say, I tried to turn it on. First I pushed an input switch on the remote (I used 0)—the normal turn-on procedure from standby mode. The green light on the processor came on, but the blue light flashed and the projector would not start. To get it to turn on, I had to go through a convoluted procedure. First I turned off both projector and processor from their respective power switches. Then I switched the power back on, first the projector, then the processor. The blue light came on at the processor, and the projector's fan turned on, without the lamp. After a minute or so the fan turned off. I then pushed the input button again and, finally, the projector came on and functioned normally. The projector demanded that I follow this tedious procedure the last few times I used it.

With the projector finally on, I took another set of color temperature readings. Again, on this last try, I changed nothing that would significantly alter the color temperature. Now the pink shift in the gray scale observed in the last attempt was gone, but the color temperature had increased to just over 7000K (not shown in either figure)! At this point I made no further attempts to adjust the projector.

In Brief
The results I ended up with, based on three separate sets of post-calibration color temperature measurements, produced three very different curves: one admirably close to the correct D6500, another at around 6000K, and a third (not plotted in the figures) at approximately 7000K. All were taken with the same color-temperature calibration settings, changed only by a turn-off/cool-down/turn-on/warm-up cycle and adjustments to the contrast and brightness controls to keep the black levels correct and the whites from clipping.

The three sets of readings I obtained were different because of varying levels of red. The green stayed relatively constant and very near to the correct green color point for all of them.

Light and Day
I had measured nearly 12fL from the projector on an 80-inch-wide Stewart FireHawk screen prior to FM's tests and with the projector fresh out of the box, but by the time it came back to me, it was down to under 8fL on the same screen. This falloff in light output was comparable what I've seen in the past on other projectors using similar projection lamps. The lamp lifetime claimed by most manufacturers is impressive, but what is left unsaid is that you may lose a substantial portion of the projector's as-new light output in the first few hundred hours. This appears to be a generic issue with the projection lamps used in most home projectors, not a problem specific to this projector or to SIM2.

I also re-measured the SIM2's peak contrast ratio (100 IRE full field/video black) with the final calibration settings discussed above, on both the Studiotek 130 and FireHawk screens with a 78-inch wide (not diagonal) image. On the Studiotek, it measured 1300 (9.1fL peak white/0.007fL video black). On the FireHawk, it measured 1272 (7.63fL peak white/0.006fL video black). These are not particularly impressive readings for a top-grade single-chip DLP projector, especially on a smallish screen with some positive gain. The black levels were respectably low—though merely average among the many DLP projectors I've tested in the past year—but the peak contrast was limited by the relatively low light output from the lamp with 200 hours on its timer.

One point that FM did not comment on is the SIM2's light leakage. Some light trickles out of the bottom of the case (the top in a ceiling-mounted installation). It wasn't excessive nor a distraction for either of us since we both sat slightly forward of the projector in our separate setups, but you should be aware of it as it may an issue in some installations.

I agree with FM about the SIM2's sharp, richly saturated images. It is a very good-looking projector. Fred lived with the projector for several months and never commented negatively on the color, at least not after his initial color-wheel glitch was cleared up.

But during the final calibrations I did feel like I had entered the Outer Limits, where someone else was controlling the horizontal and the vertical—or in this case, the color temperature. The fact that I could not obtain a consistent, stable gray scale that would hold up through these activities is troubling in a high-end projector.

Nevertheless, it's important to note that the color temperature shifts did not randomly occur during normal viewing, where they would have clearly been a serious visual distraction, but only during specific user actions, such as turning the projector off then on again later. The color was always stable—though not always precisely the same as I had calibrated—once operation had begun.

The SIM2's 9fL on my 78-inch Studiotek 130 screen (the same type and size as FM used) was adequate for normal use, and I can understand why Fred was pleased with the brightness of the image. I also suspect the aging CRTs on his DWIN projector are producing considerably less output. But the SIM2 offers no headroom for the added light loss that might occur with further use, as do a number of competitors with their more powerful lamps and light control features like multiple lamp and iris settings. I strongly recommend that the SIM2 Grand Cinema HT 300 be used on a relatively small screen (no larger than 7-feet wide) with a modest gain (such as the 1.3 of the Studiotek 130). Used within those limits, it does produce images every bit as compelling as FM reports.—Thomas J. Norton