Test Bench

For the control settings used here go to soundandvision.com.

While using them might be effective with a different sample, unitto-unit sample variations might render them less precise than an individual calibration.

All the measured results here were taken using Calman measurement software from Portrait Displays, together with X-rite i1pro3 Plus and Klein’s K-10A color meters, and a Fresco 6G pattern generator from Murideo.

Some of today’s Hdr-capable TVS, such as this Sony, are designed so that after you calibrate for SDR the set itself automatically tweaks the settings to optimize the results for HDR when you switch to an HDR source. The visible SDR settings in the menus remain unchanged in this situation; the electronic tweaks needed by the HDR material are performed inside the set and invisible to the user. But I had some trouble with this, likely due to my own fumble-itis when it comes to computers. So I did it in reverse, first doing a full calibration in HDR and then checking SDR to see if this process works in reverse. At least on the Sony, it did.

HDR10 FULL-ON, FULL-OFF CONTRAST RATIO: 448,455:1

The maximum HDR output I measured from the X93L was 986.6 nits (this was the second X93L sample we checked and the first was only a little different). The set’s ECO mode was turned Off on both samples. I also confirmed this by disconnecting most of the test setup and used just the Murideo test pattern generator connected directly to a Minola LS-100 luminance meter, which measured 1084 nits in Hdr--only marginally different from the full test setup (Murideo pattern generator, computer running Calman 2023 for Business, and a Klein colorimeter profiled against an X-rite i1pro3 Plus for color accuracy. Overall these results are about 30% lower than has been reported elsewhere. The only plausible explanation for this discrepancy would be batch-to-batch variations, though this is unlikely.

On the upside here, keep in mind that human vision isn’t linear; perceptually, a luminance increase of 25% is relatively small. If this were not the case, we’d all go blind in the transition from dawn to noon. In addition, many (but not all!) HDR sources are limited to 1000 nits. I also ran an HDR test showing the peak light output vs window size, and the X93L peaked at just over 1000 nits on a 15% window, smoothly tapering down to just over 500 nits at the lowest (small window) and highest (full screen) extremes. Most HDR HDTVS we’ve tested are less uniform than this in their brightness vs window size results.

Delta E is a figure of merit that indicates how closely a display adheres to the color standard. Experts generally agree that at

Delta E levels below 3.0-4.0 (most argue for 3.0) the result is visibly indistinguishable from perfect tracking. Delta E values are also typically used to characterize the accuracy of either white balance (grayscale), or individual primary and secondary colors.

Post HDR calibration, all of the color Delta Es were below 3.2 (cyan) with none of the others over 2.8. The luminance Delta Es (from 10% to 100%) averaged 1.3 with a maximum of 2.4.

On a color-match color checker test, measuring the Delta Es on a wide range of colors, the average without luminance was 0.79 and the peak 2.49. With luminance errors, the average was 2.06 and the maximum 6.2.

BT.2020 saturation sweeps produced a respectable result apart from green and cyan, which were poor as they are on most consumer displays. But no present or anticipated consumer sources can achieve BT.2020). The results were far better on P3 color inside BT.2020 with Delta Es averaging 2.4 with peaks at 4.4--the latter only at the 100% level (they never exceeded 3.4 at 80% and below).

Using the 1976 standard, the Sony X93L covered 72.3% of BT.2020 color and 94.4% of P3 color.

SDR FULL-ON, FULL-OFF CONTRAST RATIO: UNMEASURABLE

The SDR results were produced with the Brightness set to 35 and the Contrast to 90, as they arrive out of the box. This produced an output level of 414 nits, or 121 foot-lamberts. The brightness was 0.000 nits, resulting from the unmeasurable result shown above. As discussed in the review, you can turn these two controls down if that result is too bright for you (as it was for me). Lowering either or both of these control for SDR didn’t materially alter the calibrated results.

After calibration, the SDR luminance Delta E values of the X93L never exceeded 2.9 (at 20%; with a low of 0.6 at 100%). The color Delta Es remained at or below 1.8 except for Blue at 4.0. A color checker produced an average Delta E of 1.4 and a peak of 4.8 (the latter for 100% blue).