In an effort to be precise, fair, consistent and, quite simply, professional, how we rate TVs – and indeed all the other products that pass through our hands – is not a question of chance or simple intuition. Ratings are arrived at by averaging around a hundred different criteria, tested and grouped in 4 main test categories: Build quality and design represents a quarter of the overall rating. Display quality counts for almost two thirds and Sound quality represents 10%. Energy consumption makes up the few remaining percent.


Build quality and design

Here the available sockets, menu design and the remote are all rated. Anti-reflective coating is also evaluated. We don’t rate the type of panel itself (matte or glossy) but rather the intensity of reflections. PC compatibility is also studied (display quality and the input lag – display delay).

Many other aspects are also rated, including PIP (Picture In Picture) and PAP (Picture And Picture), backlighting settings, the thickness of the frame to be fixed to the wall or whether or not it has a rotating base.
One innovation that illustrates a trend among manufacturers: we currently test a full sample of photo, audio and video files to evaluate compatibility with TVs that claim to read these files. Among the formats used are Divx5, Xvid SD and HD, with or without subtitles, DVDs and VOBs, AVC and VC1 in TS, M2TS and MKV containers, with AC3 and DTS tracks.

A big gap in readers currently built into TVs: generally they will only read drives, keys and cards that are formatted in FAT32. This file system has a 32 GB limit. You can however go higher thanks to certain pieces of software. We therefore test discs between 32 GB and 200 GB. In all cases the files tested are limited to 4 GB, a constraint of FAT32.


Display quality

We carry out a first series of tests that allows us to evaluate a TV’s display quality at default settings. These readings are not rated but sometimes allow us to underline particular points in our articles. You can also find them in our ''Face-offs''.
The settings are then calibrated for Home Cinema type viewing; with brightness at 200 cd/m², we try to get the best out of the most common settings (backlighting, brightness, contrast, sharpness, gamma curve, colours, dynamic contrast and other destructive filters disactivated, colour temperature and so on).
Next we move onto a series of measurements that are then rated, weighed up and agregated.
Without going through all of the dozens of scores that we allocate, let’s mention the most important:

• Contrast and black levels: they are calculated using two different test patterns, illuminated at average whilte levels of 1% and 35%. These test patterns have been specially developed by Digital Versus to detect any dynamic contrast that may influence brightness and image quality. You can in the same way differentiate contrast in dark and light scenes. The higher the contrast, the better the image. It’s when you get beyond 1500:1 that the image is really set off nicely. Moving on to black levels, they're very deep for values lower than 0.07 cd/m². Depending on our needs, we use HCFR software or Colorfacts Pro 7 from Datacolor.

• Colours: this is measured according to the international standard DeltaE95 index. For this we use an Eye One sensor coupled with Blue Eye Pro from LaCie. We generally grade anything under an index of 3 as excellent. An experienced eye will sometimes be able to detect the difference between a DeltaE of 2 and 3, but this won’t be the case for the vast majority of viewers. Colours are passable when the DeltaE is around 6 but poor for anything beyond 8.

• Gamma: We measure the distribution of levels of brightness on 10 values of grey that go down from black to white. For a distribution that we qualify as good, the average gamma of greys needs to be 2.2. It can vary around this value as long as the curve is linear. An average of 2.2 with great gamma variations of some greys is much more disturbing for the viewer. This is why we pay particular attention to the shape of the gamma curve.

• Viewing angles: Since 2009, we have been carrying out a precise evaluation of viewing angles by taking a reading of colours at a 45° angle to the side of and above and below the screen. We then compare these readings with those taken straight on to quantify reduction in viewing angles. We use the same DeltaE94 index as for colours. In addition we measure light leakage on a dark image. This allows us to evaluate how much blacks lighten when you move away to the side of the screen. This phenomenon is particularly disturbing on LCD screens with tight viewing angles.

• Clouding: This is a fault linked to certain panels that allow light to filter through when the image is supposed to be black. We evaluate it by measuring how much light escapes through the panel. We consider that any light leakage at an angle of less than 10° can be considered clouding as it stops you from viewing a homogenous black image straight on unless you place yourself a good way from the screen.

• Homogeneity: the score for a panel’s homogeneity is obtained by measuring the brightness of whites at 28 points around the screen. Then the differences are agregated and a rating from 0 to 5 is worked out.

And many other readings too...

Many other readings are taken and evaluations carried out to judge display quality. For example, a panel  of viewers grades scenes in practice (films, DVDs and Blu-ray) and judges the effectiveness of video deinterlacing, of the movement compensation system or ghosting… without even mentioning the photos we take and make available in our ''Face-off''.

Sound quality

This is evaluated by our panel of viewers on pieces of music and films. 2010 will see us refine this section with the addition of readings resulting from the purchase of a specially designed testing device. This section will then be updated and expanded. Watch this space…


Energy consumption

The score for energy consumption is the result of an average worked out between standby consumption and consumption when on a standard test pattern display, with the screen calibrated at 200 cd/m² - when settings allow. The readings are taken with a Voltcraft Energy Monitor 3000. Consumption when turned on is compared with consumption according to display surface area, so as not to penalise larger screens. This allows us to compare consumption on different sized TVs.


Product survey: 40’’ to 49’’ TVs