Summer is commonly held to be a calm season, but this year that was an exciting time on the PC monitor market. Many manufacturers not only redesigned their older models but also came up with something absolutely new: new types of the matrix, new capabilities and, what is perhaps the most important thing for the customer, new prices.
Before I begin to deal with the new products, I’d like to remind you about our methods of testing LCD monitors. Our tests fall into three groups: color reproduction, response time and contrast ratio.
The term color reproduction is somewhat vague. It cannot be described with one number, and I have to perform a number of tests in this category. First, I subjectively evaluate the quality of the image or, to be exact, color gradients (smooth transitions from black to red, from black to green, etc). Such gradients look striped on many monitors, sometimes at any settings or at any settings other than the factory ones. Some users think that striped gradients are due to the use of 18-bit matrixes instead of 24-bit ones, but this is not exactly true. The lower color depth of the matrix may indeed lead to stripes in gradients if the Frame Rate Control is poorly implemented (this is the technology that emulates 16 million colors while the matrix itself is only capable of displaying 262 thousand), but the real reason is usually different. Before outputting the image on the screen, the monitor performs a series of calculations and transformations: color temperature correction, gamma compensation, contrast correction, etc. If the accuracy of those calculations is low, you see striped gradients. The matrix’s color depth has nothing to do with it. Even an “honest” 24-bit matrix cannot guarantee that the monitor will correctly process the data before sending them to the matrix. You can learn more about the different matrix types and their parameters from our article called X-bit’s Guide: Contemporary LCD Monitor Parameters and Characteristics.
The second step in my test program is about gamma curves. These are curves that show the dependence between the signal sent from the graphics card and the monitor’s pixel brightness. This dependence is not linear, but exponential and the exponent is denoted with the letter gamma . Without correction, an LCD matrix has an S-shaped curve rather than an exponential one, so it’s the monitor’s electronics’ job to bring that dependence to the necessary shape. The quality of color reproduction depends on how well the electronics does this job.
In each diagram I publish in the review there are three frames (one for each basic color) and two curves, a theoretical curve for gamma 2.2 and the curve that I draw basing on my measurements. In the best possible case the two curves coincide. If they do not, there may be image defects. If the actual gamma curves are much higher than the theoretical ones, the image will look whitish on the screen. If they go lower, the image will look dark and with too much contrast. If the gamma curves for the different colors go too far from each other, some tones of gray will have a slight coloring. If the gamma curves reach the axes of the diagram sooner than at the points (0, 0) and (1, 1), the monitor won’t reproduce details in lights or darks, displaying them as pure white or pure black, respectively.
The gamma curve graphs are drawn for the default settings and for reduced contrast and brightness settings. I usually publish only the former graphs because in most cases the two diagrams do not differ much. However, there are monitors which react to your changing their settings by getting free of image defects (for example, many monitors from NEC poorly reproduce details in light images at the default settings) or by acquiring them (for example, by losing details in dark images). In this case I publish the second diagram and tell you what difference is between it and the first diagram, created at the default settings.
The last test of the quality of color reproduction involves measuring the color temperature of different levels of gray. Ideally, this temperature must be the same, but in practice it is often the case that the temperature of a level of gray is hundreds or even thousands degrees higher or lower than the temperature of white. So when you set up the temperature of white correctly on such a monitor, you’ll see that some tones of gray look not exactly gray but bluish (i.e. their temperature is higher than necessary) or yellowish (i.e. their temperature is lower than necessary).
The matrix response time is measured with an oscilloscope and a photo-sensor that register how the brightness of a pixel is changing. In my reviews I publish the results in either of two formats: a flat diagram whose X-axis shows levels of gray (from 32 to 255, i.e. from a dark gray to white) and Y-axis shows the time it takes to switch a pixel from black to a gray and back again to black. Or I publish a 3D histogram that shows transitions not only from black to gray, but also between different levels of gray. In the second case the average for all the transitions is calculated; it is a value indicative of the matrix speed. Besides that, I measure the error of the response time compensation circuitry for RTC-enabled monitors. An RTC error is a miss that leads to a too-high or too-low brightness of a pixel (you can learn more about RTC and its artifacts in our article called LCD Panels with Response Time Compensation: 7 Monitors Reviewed ).
And finally I measure the monitor’s brightness and contrast ratio. It’s simple: I use a Pantone ColorVision Spyder calibrator to measure the levels of white and black for a few variants of settings, usually at the default settings, at the maximum settings and at the settings that yield a 100nit brightness of white (1 nit = 1 candela per square meter). The ratio of those levels is the contrast ratio. The higher it is, the closer the monitor’s black is to the real color of black.
If you would like to check out the monitors reviewed in Part 1 of our roundup series, check out the article called Closer Look at 20" and 21" LCD Monitor Features.