Panasonic Viera TH-46PZ8OU Plasma HDTV Review - Performance: Color

Once again, the TH-46PZ8OU proved to be a mixed bag. Although the display showed clean, uniform whites, the colors themselves peaked too soon, leading to a loss of detail in the image. The color gamut was also a little off from the ideal, which would make greens somewhat inaccurate.

Color Temperature (8.85)
You might think that white is white on a television, but it isn't that simple. Instead, the color of white can vary as a display shows different intensities of white, and that's what we look at here. On a perfect display, the color of the white would stay the same whether it is a pale grey or the brightest white that the display can manage. In reality, that doesn't happen; because of the way the display processes colors, the color of white varies slightly, and that's what we test here; how constant the white is. We do this by showing a series of images with different intensities of white, and measuring these. This graph shows the result; the intensity (from black to brightest white) is along the bottom of the graph, and the measured color temperature is shown by the line. If it goes above the center, the color temperature is higher (which means it is warmer, and bluer), and if it goes below, the color temperature us lower (or cooler, and more yellow).

The TH-PZ80U had good performance in this test; although the color temperature did vary with the intensity, it only varied by a small amount. This means that you aren't likely to see much of a shift in the whites on the screen; the only shift that is likely to be visible is at the lower end of the intensity range; the whites get a little warm at the low end of the range. Another way to look at this test is to plot the actual color data, which we do in the graph below. The red circle on this graph indicates the minimum color change that is generally detectable by the human eye, and the dots represent the measured color values.  The bottom line here is that if a dot falls within the circle, you aren't going to be able to spot the difference between the two whites.

As you can see from this, most of the measured points are within the circle, which means that you won't be able to spot any difference in the white. Some of the spots do fall outside the circle, though, meaning that you might be able to spot a slight difference in some shades of grey. But the differences are minor; the TH-PZ80U provided mostly consistent whites across the intensity range.

RGB Curve (7.05)
All of the colors that you see on the screen are made of three basic components; red, green and blue, which are created by the individual elements of the display. To create yellow, the display turns on the red and green elements. To create violet, the display turns on the red and blue elements. To create white, the display turns on all three. So, it is fundamental that these three components of all colors are accurately represented on screen; any issues with reproducing one of the colors will affect all three. So, we test the response of each of these components, and these results are shown on the graphs below. For each of the primary colors, the intensity goes from 0 (at the left) to the maximum, for the most intense of the colors that the display can generate.

On a perfect display, these curves would be completely smooth like a child's slide, indicating that the intensity of the input signal was accurately reproduced; every slight increase in the intensity of the input lead to a similar increase in the intensity of the color on the screen. In practice, the graphs we see tend to look more like a bumpy mountainside than a child's slide. The reason for this bumpiness is the way that displays process color and control the individual elements of the display; because it is processed digitally, the signal increases in a number of discrete steps.

This processing seemed to be something of an issue for the TH-PZ80U; the graphs are somewhat irregular and don't always follow a smooth curve. There were odd jumps on the curve (particularly on the blue), and spots where an increase in the intensity of the input didn't lead to a corresponding increase in the luminance of the screen. This can lead to issues such as false contouring (where a band of color appears in a subtle gradient), and we saw some evidence of this in our tests using a variety of test photos; some images had banding and false contouring that made them look almost unnatural.

The other issue is at the top end of the red curve, where the curve flattens off. This indicates another issue; the luminance for red is being pushed too hard, and it reaches the maximum intensity too soon. In effect, the display can't produce any more red or blue at this point, even though the intensity of the signal is increasing. What this means for the viewer is that some subtle details will be lost; details of objects bathed in sunlight or particularly brightly colored objects will be lost because the display can't reproduce the subtle changes that these details require. We found this happening a lot in the images we use for testing; details like the stitching on a drum skin in bright sunlight were lost; the drum skin looked like a flat surface instead of the textured, contoured surface that it really is.

This issue gets much worse if you go for brightness and bump up the picture control. If you do that, the curves peak even sooner, and even more details are lost. In our sample photos, bumping up the picture control made the image brighter, but even more detail (such as the aforementioned drumskin and clouds on a bright blue sky) just vanished, replaced by flat white or flat color. And again, this underlines the compromise of this display, where you have to pick color accuracy and range or brightness; you can't get both.

Color Gamut (4.32)
Television signals contain a certain range of colors, called the color gamut. For high definition signals, these limits are defined in an international standard (called the ITU Recommendation .709). So, a decent display should accurately match this color gamut, and that's what we test here. We measure the red, greens and blues that the TV can produce and match them against the standard; a perfect display would exactly match the standard gamut

For the geeks amongst you, the chromaticity values of the colors are in the table below.

The TH-46PZ8OU didn't do that well here; although the blue was pretty close, the red and green were both somewhat off. In both cases, the maximum extent of the color gamut was too far out, meaning that the display can produce a wider gamut of colors than the standard calls for. While this might sound like a good thing, it's actually a problem; it will lead to inaccurate colors that will look unnaturally vivid. This was borne out with our tests with a series of sample pictures; objects such as tree leaves and red military uniforms looked somewhat lurid and slightly cartoonish.