It’s been a while since we last tested 19” LCD monitors on our site. Not only new monitor models but, what is more exciting, new technologies have emerged since then (alas, a release of a new model today just too often turns out to be a minor revision of an older monitor, so it is even hard to tell sometimes where the new model differs from the older one).
Trying to keep up with the times we are adjusting the methodology we employ to test LCD monitors. A two-dimensional diagram of response time measured on transitions from black to levels of gray and back again once used to provide an exhaustive description of a monitor, but when it comes to the new models with response time compensation it is important to know how long transitions between different levels of gray may take. And we want to know this not only to evaluate the response time parameter proper but to check out if there are any RTC artifacts which may arise when the compensating impulse is way too strong and the pixel brightness shoots up above the necessary level (this shows up on the screen as light shadows, rainbow patterns, etc).
In order to help you compare different monitors better, from this article onward I will provide 3D diagrams of response time and RTC error for all RTC-enabled monitors and will calculate an average response time by averaging all the non-zero values in the table of transitions (zero values correspond to “zero” transitions like “128-128” and to transitions which our measurement system lacks precision to measure, e.g. transitions between the lightest color tones). I will also calculate an average RTC error by averaging all the values in the RTC error table that have non-zero correspondences in the response time table.
I want to remind you that you shouldn’t base your judgments on numbers only because the distribution of numbers may also be important. For example, take and compare the RTC error diagrams for Samsung’s 940BF and 960BF monitors (you can find them in the text of this review): although the former model has a bigger average error, its errors are more uniformly distributed among all the transitions and thus are going to be less conspicuous visually than the huge errors of the 960BF on a few transitions from black to gray. Yes, this analysis is hard to make and is even harder to justify because there is always room for an argument like “What’s better: to have a few big errors or a lot of smaller ones?”, but we just have no choice other than to analyze deeply. If we operate with averaged numbers only, we run the risk of becoming like the manufacturers of LCD matrixes who measure response time by a black-to-white transition and come up with a number that often tells you nothing about the real speed of a particular matrix.