- Written by Chris Eberle
- Published on 15 May 2013
The Mitsubishi HC8000D-BL 3D DLP Projector On The Bench
The HC8000D-BL was installed upside-down eleven feet from a Carada Brilliant White screen; which has a gain of 1.4. Readings were taken with an X-Rite i1Pro spectrophotometer off the screen at a distance of 10 feet. Patterns were provided in both 2D and 3D by an Accupel DVG-5000 signal generator and everything was controlled by CalMAN 5.0. Luminance measurements were taken by an X-Rite CA-6X tri-stimulus colorimeter controlled by Progressive Labs software.
The preferred starting point for both out-of-box and calibrated performance is the Cinema mode. This mode utilizes the best presets for color temperature and gamma. The only exception is the color gamut which defaults to Wide. It is slightly over-saturated so I changed it to Normal before beginning the tests. The bulb was at its Normal setting which is the higher power mode. The fixed iris was set to High Contrast and the auto iris was turned off.
The grayscale is slightly warm but tracks pretty evenly from black to white. DeltaE values are between three and five except for 90 percent which is slightly higher. This is a barely visible error and represents very good performance. Gamma tracking is superb right out of the box. Even though there is a multi-point gamma control, I had no need to use it.
Here is the eight-point CIE chart.
This is only a fair result. Red is a little over-saturated and blue is a little under. This affects the magenta secondary which is also under-saturated. Green, yellow, and cyan are much closer to the correct points. Luminances are also quite good for those colors.
To see the rest of the story, check out the saturation sweep.
The colors that perform well on the previous CIE chart also look good on the saturation sweep. They track properly from white all the way to the 100 percent saturation point. Blue and red however are not quite as accurate.
After calibration, I measured the following for grayscale and gamma.
This is one of the best charts I've ever seen. DeltaE values are under one for all levels except 20 and 90 percent; and they are both under two. The hump in gamma at 90 percent is a tiny one, only representing a level that's 0.13 fL too dark. This is about as good as it gets folks.
After spending some time with the color management system, I was able to improve the gamut measurements.
Red is now much closer to its correct point, as is magenta. Blue however is still in the same spot. No manipulation of the saturation or hue controls would bring it into line. I was also unable to get yellow any better. Like most projectors, Mitsubishi's CMS is effective with luminance but not saturation and hue.
When I see any color points more than a bit off the Rec 709 spec, I like to check the positions of the secondary colors relative to the actual primaries. This tells us whether or not color phases properly from one primary to the next.
You can see the secondaries are right where they should be. If you're a CalMAN user, you can do this yourself with the color target editor. Simply create a custom gamut using the editor; then select it as the target. The CIE chart will be recalculated automatically.
Here is the post-calibration saturation sweep.
All colors are now quite close to their points at all saturation levels except for blue. And the luminance levels are superb. This has the greatest impact on perceived color accuracy and balance. I was happy to see the luminance controls work so well. Aside from the issues with the blue primary, the HC8000D-BL is a very accurate projector.
A separate 3D calibration was performed with the glasses placed over the meter. Color and gamma were largely the same in 3D, but grayscale required a fair amount of adjustment. The end result was an average DeltaE of 1.86 which is excellent. No adjustments were needed in the CMS to achieve a similar chart to the ones above.
Utilizing the crosstalk pattern from the Accupel, I measured an extremely low 0.21 percent. This is well below the visible level. DLP usually provides the best 3D performance with the fewest artifacts and crosstalk. The HC8000D-BL is a great example of this.
Light Output & Contrast Performance
The HC8000D-BL is not an exceptionally bright projector. Even in its high-lamp mode, the best it could do was 17.215 fL with the fixed iris set to High Bright. There is no penalty for opening the iris. Gamma, grayscale, and color are unaffected. In the High Contrast mode, the max reading was 13.838 fL; still perfectly acceptable on my 92-inch screen.
In 3D the brightness drops to 3.043 in the High Bright iris mode and 2.433 in High Contrast. These numbers sound low, and they are, but the 3D experience was pretty good thanks to the HC8000D-BL's superb clarity, excellent contrast, and complete lack of visible crosstalk. My only concern would be as the bulb ages, it will dim somewhat. These light levels are just bright enough to enjoy good 3D.
The most impressive thing that came out of my testing was the black level. DLP is not known as the king of contrast but this projector cranked out some impressive blacks. I measured .001 fL in High Contrast, and .002 fL in High Bright iris mode which makes the contrast ratio either 13,838:1 or 8607.5:1. Of all the DLP projectors I've reviewed, the Mitsubishi has the lowest black level. That it can maintain it even with the fixed iris open is even more impressive. This excellent contrast adds to the 3D experience as well. In the High Contrast iris mode, the black level was immeasurable.
The HC8000D-BL did pretty well in our standard battery of video processing tests. Below-black was an unusual failure. To render the pattern properly, I had to input an RGB signal and set the HDMI Input option to Enhanced. This option is grayed out with component signals. That means with a 4:2:2 or a 4:4:4 input, all information below video level 16 will be clipped. This won't impact image quality but it will make it harder to set the correct black level. There was no clipping above white regardless of mode or signal type. The best resolution is seen with RGB signals. While component signals pass the burst test, RGB renders the 1-pixel pattern more clearly. The failure on the Mixed Content Vertical test is due to line twitter in the scrolling text.