Dream Vision Starlight1 LCoS Projector


On The Bench

Equipment used: EyeOne Pro spectrophotometer, CalMAN Professional 3.7 analysis software, Accupel HDG-3000 signal generator, Spears & Munsil Benchmark Blu-ray disc.

All measurements were taken off the screen (Carada Brilliant White, gain 1.4) from the seating position (10 feet back). Video processing tests were performed using an Oppo BDP-83 connected directly to the projector and set to Source Direct mode.

The Cinema 2 preset offered the most accurate out-of-the-box color performance. If you don’t want to calibrate any further, this is the mode to use. The color gamut is the same regardless of mode. Only the luminances change. Without a color management system, the only modes offering correct color luminance are Cinema 2 and User. Since I wanted maximum control, I chose User as my starting point.

You can see below the color points are pretty close for red and blue but green is more over-saturated. Luminances however are nearly perfect; less than .1 foot-Lambert off target for every color. This is excellent performance.

Pre-calibration grayscale isn’t too bad. This would be a perfectly watchable image if left un-adjusted. Only the gamma is visibly off at 1.98 but tracking is decent except for a little slide downward at 80 and 90 percent stimulus. This is also excellent performance especially with regards to grayscale accuracy.

Post-calibration color measures the same since there is no CMS. Fortunately the luminances are spot-on. This has a greater impact on picture quality than accurate color points. You can also see when the secondaries are calculated to the measured primaries, their positions line up perfectly.

Grayscale is pretty much dead-on after calibration of the custom color temp memory. There are three slots available if you want to create other white balance presets. I like to have a 5500K setting for black & white movies. Gamma was adjusted simply by choosing the A curve and raising the correction factor. There is a multi-point adjustment available but the tracking was flat enough that I didn’t need to use it.

Video processing performance was excellent thanks to the HQV Reon VX chip utilized by the Starlight1. All the Spears & Munsil source adaptive tests were passed with the exception of 2:2. Edge adaptive tests were handled well with minimal pixilation or line twitter in either the horizontal or vertical axes. With Crystal Motion set to Off, 24Hz signals are frame-quadrupled for a refresh rate of 96Hz and 60Hz signals are left unchanged. When Crystal Motion is turned on, all content is displayed at 120Hz with frame interpolation.

I tested the Crystal Motion on a moving resolution wedge pattern. This is the most extreme test of moving-object resolution I know of. On the Low setting artifacts were minimal; just a hint of breakup in the higher frequency portions of the pattern. On High, the breakup artifacts were more severe and there was tearing of the wedges. As I noted earlier, the High setting showed artifacts in actual content while the Low setting did not.

Optical performance was superb as every image and test pattern was rendered with maximum resolution. Sharpness was very good edge to edge. This projector also had near-perfect panel alignment. I observed no color shift on any part of the screen in a full-field white pattern. Contrast performance was literally off the chart. I was unable to record a measurement for minimum black. I adjusted the iris to its smallest opening for a peak white level of 12.8 foot-Lamberts. Since I had no minimum reading, I could not compute the contrast ratio. Let’s just say the image easily competed with my Pioneer Kuro plasma TV. This is no small feat. The only other projector I’ve worked with that offers this kind of image depth is my reference Anthem LTX-500, also an LCoS design.