QD-Oled, Mini-Led: how Samsung TVs trick test procedures

It was Vincent Teoh of HDTV.co.uk last April who discovered the first of the rather strange results measured on the Samsung S95 TV – the brand’s first QD-Oled model. The latter displayed a higher peak of brightness on a standard-sized window than on a custom-sized window… a hitherto unseen behavior. These strange results were corroborated by our colleagues at FlatpanelsHD when testing another Samsung TV – the Neo Qled QE65QN95B – in early June.

We have just received this Samsung Mini-Led TV (QN95B) with the latest firmware and we can confirm that its behavior is still unusual. The Samsung QE65QN95B thus displays a different peak of brightness depending on the size of the window used to within a few %, which the competing models do not do. This has an impact on the brightness peak, but also on the color accuracy, which makes the Samsung TV impossible to test in good conditions. Here are the results we obtained with a window of 10%, the rendering being then excellent, then with a window reduced to 9%. The difference is telling.

With a window of 10%, the results in Filmmaker mode are very good. The TV smoothes the signal from 50% luminance up to the device’s maximum capabilities. The peak of brightness thus culminates at 1800 cd / m². The Mini-Led system is very efficient here, and clearly makes it possible to display a higher luminosity than that measured on the best Oled televisions such as the Panasonic 65JZ2000, flashed at 962 cd / m², or even Sony’s QD-Oled television — the Sony Bravia XR-55A95K – whose peak brightness was also measured at 965 cd/m².

With a window of 10%, the colors can be considered accurate and faithful to those sent by the source, since the Delta E is less than 3 — the threshold below which the human eye no longer perceives color nuances.

As you will have understood, with this 10% window, the HDR rendering is very good and the peak of brightness makes it possible to display an HDR rendering benefiting from a very beautiful dynamic, with a more visible intensity than on the OLED models by example. However, things are quite different when changing the size of the window. The results are also identical with a window of 5 or 18% (the window sizes predefined in the Calman software).

Custom window of 8, 9, 11 or 12%: a peak brightness of only 976 cd/m² and less accurate colors

With a very close window, but whose size is customized to 8, 9, 11 or 12% for example, the results are totally different. The EOTF curve is now much too smooth, since from 10% luminance, the television is well below the reference curve. Above all, the peak brightness is now limited to 976 cd / m², the equivalent of the peak brightness recorded on a very high-end Oled television or on a QD-Oled model. It should also be noted that our colleagues from FlatpanelsHD did not obtain the same results at all, since with a personalized window they observed a luminance that was too high, above the reference EOTF curve, but also a peak in brightness set back by compared to that obtained with the 10% window.

Using a custom window size has a negative impact on color rendering, since the average Delta E then rises to 5.9. The colors can moreover be considered as perfectly faithful to those sent by the source.

With this custom window, the HDR rendering is therefore totally different. Even if it is not bad and remains above average, thanks in particular to a peak in brightness higher than the majority of OLED televisions, it is clearly disappointing for a high-end Mini-Led television.

A model impossible to test without a firmware update

Unfortunately, the editorial offices of the Digital are moving and, to date, we have not been able to compare this television with another model on real content. Without a point of comparison, it is always difficult to draw conclusions on a strictly visual basis. In particular, it is quite difficult to determine whether the television delivers more than 1500 cd/m². With some HDR content, we measured a peak in brightness of 1400 cd/m²: this is more than the 976 cd/m² recorded on the custom-sized charts, but less than the 1800 cd/m² obtained with the chart at 10 %. The HDR image is bright, but the model may be limited to around 1400 cd / m² in content.

As we saw during the test of the Sony Bravia XR-65X95K Mini-Led television, a peak brightness of 1400 cd/m² is more than enough to obtain a very good HDR rendering, especially thanks to the good depth of the blacks produced by the Mini-Led backlighting system. Therefore, we do not really understand what drives Samsung to absolutely want to go up to 1800 cd / m² in the test results, except to want to pass in front of its direct competitors. The Korean brand is not the first to want to modify the behavior of its devices on a test bench. The most emblematic case is probably that of the diesel gate : Volkswagen, the world’s largest car manufacturer, had installed a device on its engines to reduce CO emissions₂ when they got on the test bench for the certification tests. This problem is recurrent in the field of smartphones as well, the publisher of the Geekbench software having excluded several models (Xiaomi X11, Samsung S20/S21/S22, OnePlus 9) which modified their behavior when launching the tests.

Like all of our colleagues, we will now measure the peak of brightness with a custom window size always between 5 and 15% since it is on this surface that televisions are designed to deliver their best peak of brightness. We are now waiting for the release of new firmware with identical standard behavior over the entire measurement range before continuing to test the Samsung QE65QN95B.