We then tested the system on our reference TFT monitor. Naturally we used the same method.

Concerning TFTs, a question arises: What about latency as a function of the level displayed? I was able to observe a degradation in rise time with a gray level going from 0 to 100 or 110 as compared a move over the full range from 0 to 255. But it should also be pointed out that fall time evolves in the opposite direction: It increases as we approach 200 and becomes maximal at 255. To avoid any confusion, we chose to construct a graph representing latency in terms of the intensity of the green channel. To get as reliable an idea as possible of the actual latency, for each test we perform the measurement separately on each channel, red, green and blue.

This graph partly explains the success of the Hydis 20 ms panel. Its latency does reach 28 ms in the worst cases, but certain faster panels (some 16 ms models, for example) go well beyond these values, even though they do reach 16 ms at peaks. That's why the Hydis panels are known for being very responsive. They deviate less from their published specifications than their competitors do.