Adobe RGB: Pre-Calibration Testing

Our first tests involve measuring the Philips 276E6 prior to any calibration being performed. These are arguably the most relevant numbers for this sort of monitor, as it's highly unlikely that consumers purchasing $300 monitors will also happen to have a colorimeter or spectrophotometer laying around to calibrate it. 

While no calibration is performed before this test, the display is set to a brightness level of 200 nits to keep results comparable between reviews. The white point setting has been left in the default Adobe RGB mode, which for some reason happened to be more accurate than the mode explicitly labeled 6500K.

Greyscale

On this unit, greyscale accuracy was where I would expect a low priced wide gamut monitor to be. What's strange is that the gamma was closer to a target of 2.6 despite the fact that I left the monitor in its default power 2.2 gamma mode. It's worth noting that the original unit had a gamma that was much closer to 2.2, but the RGB balance for grey shades was much more red shifted and so the greyscale DeltaE was actually higher than this unit. It looks like there's a fairly large degree of variance between different units of the 276E6.

Saturation Sweep

Due to the shifted primaries and improper gamma, the Philips 276E6 doesn't perform as well as it could in our saturation sweep test. Magenta is pulled toward red, while cyan is pulled toward green. For many levels of saturation there are fairly severe errors in magenta and blue, and even more so in green, red, and cyan.

Gretag-Macbeth ColorChecker

In the Gretag-Macbeth ColorChecker test the Philips 276E6 doesn't perform very well. The average DeltaE is nearly five, with a number of colors actually exhibiting individual errors above six. In this case the oversaturated red primary is causing significant problems with accuracy by also shifting shades of yellow, orange, and magenta toward red. I really have no idea why Philips decided to tune the monitor in this manner, and unfortunately it makes the monitor not very useful for work that depends heavily on color accuracy.

Adobe RGB: 200 Nits Calibration

Our 200 nit calibration target is still the Adobe RGB gamut with a power 2.2 gamma. Since we can't actually alter the display's primaries due to there being no 3D LUT, there will be no way to improve upon the errors with the color gamut and any saturations or color mixtures that rely heavily on accuracy there.

While my earlier testing was done in the monitor's default Adobe RGB mode, for calibration I moved to using the user defined mode which allows control over the white point through a set of RGB sliders. This allowed improvements to be made at the monitor level before making further tweaks by performing the greyscale calibration, which helps to retain tonal range as there are fewer adjustments that need to be made in the GPU's LUT.

Greyscale

Greyscale accuracy on the Philips 276E6 improves dramatically after calibation. While the gamma is still far too high in the darkest shades, it's much more linear and now tracks fairly close to our 2.2 target rather than 2.6. The greyscale does have a number of areas where one shade will have a much higher error than the surrounding ones, and it may take moving from a 65 point to an even more comprehensive 255-point calibration to eliminate these errors.

Saturation Sweep

The average error with primary and secondary color saturations is lower after calibrating the 276E6, but there's a caveat. Due to adjustments made to the greyscale, as well as the adjustments made to the monitor directly through the white point settings before calibration, the error in cyan and green is actually much higher in the lower saturations, while the errors for blue, red, and magenta in the more intense saturations are also higher. I'm honestly hesitant to actually describe this as an improvement, and many of the errors are far too severe for the monitor to be usable in color critical applications.

Gretag-Macbeth ColorChecker

Performance in the Gretag-Macbeth ColorChecker test improves after calibration, but much of this is due to the improvement in greyscale accuracy, with many colors still exhibiting very large errors. Unfortunately, even after calibration, the Philips 276E6 isn't really suitable for use as a monitor for photo or video editing.

Adobe RGB: 80 Nits Calibration

For our 80 nit calibration we continue to target the Adobe RGB gamut, but in addition to the lower brightness we also target the sRGB gamma curve rather than a simple 2.2 power function. sRGB's gamma allows for greater detail in the darkest regions of images, and this calibration target closely reflects what one would target on a professional display where images are being edited for print.

Greyscale

As expected, the 276E6 struggles when calibrated using the sRGB gamma target. In the darkest shades of grey the errors are quite high, descending from a DeltaE value of roughly 9 at black to an error of 3 at roughly 25%. Beyond that point the calibration is actually fairly good, and the average overall error is below our target value of 3.0. However, the severe inaccuracy in the dark regions makes it fairly evident that the monitor won't be usable in applications that require strict conformance to the sRGB gamma and luminance specifications.

Saturation Sweep

Surprisingly, the Philips 276E6 performs much better in the saturation sweep with our 80 nit calibration than it does at 200 nits. This may be due to the modifications I have made at the monitor level to correct the white point before doing a greyscale calibration in CalMAN, and given that the white point settings on my two samples were very different it's likely that results with different calibration targets will vary a great deal from unit to unit.

Gretag-Macbeth ColorChecker

Another surprise is the fairly low average error in the ColorChecker test. I can really only ascribe these differences to the changes made when altering the monitor's white point, but it really is interesting to see a display perform better with this arguably more difficult calibration target than with the 200 nit calibration with a simple power 2.2 gamma. 

Contrast, Brightness, and Gamut Display Uniformity
Comments Locked

51 Comments

View All Comments

  • willis936 - Friday, April 29, 2016 - link

    I'm skeptical of your first claim without seeing data.

    As for the second that's why packages lime dispcalgui exist.
  • Brandon Chester - Friday, April 29, 2016 - link

    Again, that doesn't help the fact that software needs to support it. I think you're confusing greyscale calibration and color management here. If there was some easy way to fix color management across all Windows programs this would not be such a long standing issue.
  • UrQuan3 - Thursday, May 12, 2016 - link

    "Cheap colorimeters are so inaccurate that they're basically useless."

    I'm going to have to go with willis936 on your first comment. It sounds rather like someone driving a Ferrari saying that a Mustang has so little horsepower it is useless. To the average car owner, they're both godlike. In practice, a little $100-200 colorimeter makes a large improvement on almost any monitor. Expensive calibration for expensive monitors. Of course, use the best gear when doing a review.

    I wonder how you would review calibration tools? That does not sound easy.
  • Pork@III - Thursday, April 28, 2016 - link

    Too bad against full cover CCFL
  • Azurael - Friday, April 29, 2016 - link

    It's possible to get a 27" 2K display for $300 equivalent in Europe... I've got a Hannspree HQ271HPG which even with VAT is £200. I wouldn't say it's the best thing in the world (stuck with HDMI 1.4 & DL-DVI and hiding >1cm behind a piece of glass) but it is IPS, it calibrated up nicely (to sRGB) and the backlight consistency is much better than most cheap monitors on my sample (although it does have a bit of bleed visible at the very edges on a totally black screen.)
  • Gunbuster - Friday, April 29, 2016 - link

    I know it's a cheap monitor but dear lord, did they have to make the bezel so chunky that it looks like a 22" in photos?
  • Haravikk - Friday, April 29, 2016 - link

    Why does this include a VGA port?

    I'd also much prefer down-facing ports, and some kind of cable management, monitors that don't include these always confuse me.
  • zodiacfml - Saturday, April 30, 2016 - link

    Thanks for always including a tutorial and in-depth look of color management. I quite understand the challenges of the industry.

    You are correct that Philips should be applauded for taking the first step as this will take time to improve as OLED/AMOLED of Samsung has improved throughout the years. For now, the Philips seems useful for increasing saturation/vividness of content for entertainment.

    Questions:

    1) Isn't better for Philips to target a higher color space despite coming short for now (as conversion from a bigger space to smaller seems straightforward)? The Adobe RGB doesn't improve from the sRGB space in the "reds" where the most benefit from quantum dots can be had. I believe this primary color should be given attention as content to show this is widely available in photos such as flowers, sunsets, and red sports cars. I have seen too many red subjects looking flat like plastic.

    2) How does color spaces Rec. 2020 and Pro Photo RGB relate to each other? They seem to have the same coverage but obviously for different applications.
  • zodiacfml - Saturday, April 30, 2016 - link

    I did some reading and found the problem already which is color bit depth. What are the currently supported bit depths supported by video cards and monitors?
  • Oxford Guy - Monday, May 2, 2016 - link

    AdobeRGB is obsolete.

Log in

Don't have an account? Sign up now