Stewart Filmscreen Harmony G2 Acoustically Transparent Screen Review Test Bench

Test Bench

Screen Target Point White Point Luminance (ft-L)
Center (L&R, 2° Avg.) 0.312 / 0.32916.25
Center Left Side 0.313 / 0.31313.31
Center Right Side 0.311 / 0.33413.35
Center Top0.313 / 0.33316.16
Left Top0.313 / 0.33413.23
Right Top0.315 / 0.33313.42
Center Bottom0.313 / 0.33315.58
Left Bottom0.313 / 0.33512.93
Right Bottom0.312 / 0.33513.14
Center Left (from 35° Left) 0.313 / 0.32915.99
Center Right (from 35° Right) 0.312 / 0.33016.12

For measurements, I used a Labsphere USRT-99 Spectralon Reflectance Target placed on a tripod directly in front of screen center. (Many screen manufacturers use Labsphere targets as a baseline parameter for a color-neutral, Lambertian reference with 99% reflectance at any measurable angle.)

My luminance target for Rec.709 is 16 ft-L, so I began my calibration and, when finished, the white point was 0.313/0.330, and light output 18.2 ft-L. My 2.4 gamma target was a nearly flat 2.4. Setting the Labsphere aside, I refocused the projector to the screen and ran a pre-calibration pass that would highlight differences, if any, between the reference reflectance target and the screen itself.

The white point clocked in at 0.311/0.330 as light output dipped to 11.935 ft-L—not unexpected for a woven screen, though the lamp remained in Eco. Gamma remained largely corrected, averaging 2.39, rising from 40 IRE and staying at near 2.0 through 100 IRE. Calibration improved gamma tracking, dropping the average to 0.836 with the largest error coming at 90 IRE and measuring 1.53. Light output dropped a very slight 0.269 ft-L.

Color gamut pre-calibration averaged 1.309 with red the leading error at 1.159. After calibration, the average dropped to 0.678, cyan becoming the biggest offender at 0.958. Using the Color Checker feature in Calman software from Portrait Displays, a measurement of 47 different hues, the averaged error was 1.1, with a maximum 3.2 on purplish blue. Blue flower and blue sky upped the average as both measured at 2.8. A slight blue push isn't unheard of with woven screens, but Stewart has managed to deliver a nearly spectrally flat surface based on measurements made with a very capable $2,999 Epson projector.

With the projector's Natural mode selected, the calibrated white point at center screen was 16.249 ft-L and 0.312/0.329. Points measured at the perimeter varied from 12.925 ft-L and 0.315/0.333 to 15.581 ft-L and 0.313/0.334. I don't attribute these slightly lower light measurements and deviations from the white point to the screen, but instead to fall off from the outer circumference of the projector lens and associated chromatic aberrations.

This data indicates how neutral Stewart's screen is. Spectralon is a "static" reference surface material. The screen becomes reference caliber with a compatible projector (and skilled calibrator). This affords another hat-tip to Epson, because if a $3K projector plus this screen, in the standard size of 96 by 54 inches for just above another $3K, can create an image only hundredths of a percent off from reference, that's a challenge unmet a decade ago in metrics approaching six figures left of decimal point.—MPH

Stewart Filmscreen
(800) 762.4999