Twitter: Stay up to date: @TFTCentral |    Forum: Discuss this article  | RSS Feed: Review Alerts

Introduction

The "gamut" of a display refers to the range of colours it can display, and for LCD monitors this is normally related to the backlight type that is used to light the screen from behind the LCD panel. There are several common reference colour spaces that you will no doubt have heard of, including the long-standing "sRGB" reference space and more recently adopted standards in the monitor market like DCI-P3 or Adobe RGB. This article will provide a hopefully simple overview of what colour gamut is, then take a top level look at the different ways to measure and quote this spec, without delving in to loads of overly technical or complicated information that might bore you too much. We will also talk about our newly improved testing methodology and equipment we will start to use for future reviews to enhance our accuracy and data we can report.
 

CIE-1931 offers the first representation of colour range

Experiments at the beginning of the last century into the human eye eventually led in 1931 to the creation of a system that encompassed all the range of colours our eyes can perceive. Its graphical representation is called a CIE diagram as shown in the image above. All the colours perceived by the eye are within the collared area. This standard is called CIE-1931 and is widely used throughout the world to be able to measure and compare colours for printed documents as well as displays. The CIE-1931 standard maps colours to a 2 dimensional x,y space as shown above. It is sometimes referred to as a "CIE-1931 xy chromaticity diagram"

CIE 1976 provides improved accuracy but is still not widely used

The CIE-1931 color space has worked well to enable accurate color reproduction in print and displays. Unfortunately, further study revealed that the severity of color differences, as perceived by humans, were non-uniform in the CIE-1931 color space. What that means is that for example two blue colours at a certain distance ‘v’ in the CIE-1931 color space will look more different to humans than two green colours at the same distance ‘v’. In 1976 a new color space, called CIE-1976, was introduced to fix this issue. The CIE-1976 color space, as shown above, was specifically designed to align uniformly with human perception. This is sometimes referred to as a "CIE-1976 u’v’ chromaticity diagram". The CIE-1931 color space should no longer really be used, but today’s reality is that it is still prevalent throughout the industry.

Be careful of how colour gamut is quoted

You may well see some display manufacturers quoting colour gamut figures in their specs and marketing, but you need to be mindful of whether they are based on the 1931 or 1976 CIE reference. This will have an impact on the coverage % and so if you are comparing different displays, you need to ensure you're comparing like for like (i.e. comparing CIE 1931 against 1931 for both displays). CIE-1931 tends to still be far more common so if it's not specified, it's likely that is what is being used.

Colour gamut standards - sRGB

Within this overall CIE area there are various standard colour spaces that have been defined over the years. In the display market these define a certain range of colours that the screen can produce and is related to the backlight of the screen, as opposed to the panel itself.

The most common colour space that has been around for a long time and is widely used is sRGB. This is the default for Windows and for the internet for instance, and is widely used across different devices and content.

sRGB coverage relative to CIE-1931 and CIE-1976

You may be familiar with images like the above which shows the sRGB colour space triangle within the CIE-1931 overall area or less commonly within the CIE-1976 area. The sRGB gamut triangle defines a certain range of colours based on the primaries of red, green and blue (RGB).

When discussing monitor colour gamut many manufacturers will quote how much of this sRGB space the screen can cover, with modern "standard gamut" LED backlights normally covering 97 - 100% of this reference space quite easily. It's important for a screen to have good coverage of sRGB at least to ensure accurate colour reproduction and so that it can display the full range of colours it is supposed to, when viewing common sRGB-produced content.

Wide Gamut

Many modern displays are now promoted as featuring "wide colour gamut" (WCG), being able to produce colours that are beyond the range of traditional "standard gamut" sRGB displays. This is achieved in a number of ways by enhancing the backlight with special treatments or films (e.g. Quantum Dot coating), or using certain backlight types in some cases such as RGB LED. The result is a screen that can offer more vivid and saturated colours that stretch beyond sRGB and try to confirm to other colour space standards, supporting content created and intended to be displayed in these wider colour spaces. Some common wide gamut reference spaces include...

DCI-P3

One of the most popular colour space standards quoted today is DCI-P3 which was defined by Digital Cinema Initiatives, and is associated commonly with cinema and movie content. It has become closely linked with modern HDR promotion also and so has become more widespread in recent years with modern TV's, monitors and content.

 
DCI-P3 coverage relative to CIE-1931 and CIE-1976

In simplistic terms where high coverage of DCI-P3 is achieved, a buyer knows that it can offer wider colour range than a standard sRGB gamut screen, and will allow them to better display content that has been produced based on this standard. Modern HDR movies, videos and games are good examples of where DCI-P3 is used for creation, and where you ideally therefore need a DCI-P3 supporting screen to show them properly.

Many modern displays will therefore quote a DCI-P3 coverage %. Again you need to be careful of whether this is relative to CIE-1931 or 1976 and ideally you'd want it to be as close to 100% coverage as possible if you want to be able to accurately produce the intended colours from DCI-P3 content. Nowadays this DCI-P3 coverage provides a fairly useful comparison spec when comparing wide gamut screens, since sRGB is no longer wide enough unless you refer to over-coverage figures, which can sometimes be misleading or can get a bit silly.

Adobe RGB

   
Adobe RGB coverage relative to CIE-1931 and CIE-1976

The Adobe RGB (1998) color space was developed by Adobe Systems, Inc. to improve upon the gamut of the sRGB color space in primarily cyan-green hues, for better conversion to the subtractive CMYK color model used by printers. It uses the same red and blue primaries as the sRGB color space, but the green primary is more saturated and also of a much deeper hue of green (sRGB uses a rather yellowish hue of green). This colour space is more commonly referred to with the more professional-orientated monitors aimed at printing, colour critical work etc. It doesn't carry the "multimedia" connotations that DCI-P3 does and so is less widely used nowadays for many "wide gamut" screens.

Rec.2020

 
Rec.2020 coverage relative to CIE-1931 and CIE-1976

The Rec. 2020 standard is actually more than just a color space. It is the standard for Ultra High Definition Television (UHD-TV), which also includes resolution, refresh rate and colour depth considerations. Focusing on the colour gamut side of things for the purposes of this article, the Rec. 2020 offers significant improvement over the Rec. 709 standard when it comes to the color gamut: nearly twice the size of its predecessor. It stretches considerably beyond Adobe RGB and DCI-P3 as well. This is not commonly used as a monitor spec at the moment, but expect this to become more widely quoted in the future as manufacturers reach beyond the now popular DCI-P3 reference, and look for new ways to promote and compare their wide gamut production.

Over-coverage of sRGB as a spec

One thing you will often see in specs and marketing is a colour gamut spec relative to sRGB, but quoted beyond 100%. We also currently provide this in our reviews as we feel it's quite a useful comparison figure between different displays. For a wide gamut screen this basically shows how far beyond sRGB the colour space extends, so you will see figures such as 130% sRGB and so on quoted.

This is a reasonable spec to include for comparison as you could have one screen which has 120% sRGB for instance, compared with another that can do 160% sRGB. This tells you that the latter will have a wider colour range, and should offer more vivid and saturated colours. Some people like that added boost to colours, regardless of whether they are conforming to certain colour spaces or standard, or whether they are "accurate" for the content they are viewing. Some people just like the boosted and accentuated colours.

The problem with this over-coverage spec is that it could be a little misleading or prone to error. For instance in the example CIE-1976 space above you can see an example monitor gamut plotted against the sRGB reference. In this case the monitor goes beyond sRGB for the green and blue primaries, but just falls short for the red primary. The total area of the monitor gamut is bigger than sRGB, but reporting a >100% coverage for the sRGB standard would be technically incorrect because not all sRGB color can be accurately displayed.

Quoting an over-coverage of sRGB is technically possible only if the sRGB gamut is fully covered at 100% on it's own. We will continue to provide some measurements and context in our reviews in this area in the future anyway as it can be a useful reference point to compare different screens.

Monitor Specs and EDID Reporting of Gamut

Most manufacturer websites and marketing brochures will list the colour gamut of the display nowadays. We've already mentioned earlier about being careful whether they are quoting CIE-1931 (the most common still) or CIE-1976 coverage specs. You need to compare like for like at least. Some of the specs they quote might be based on their own internal testing at the factory which should be reliable but probably still needs validating by third parties or reviewers. This factory measurement is especially likely to have happened if you see rather specific specs like "97.6% DCI-P3" for example. Other times it might be a more generic figure based on the intended panel/backlight spec of the part they are using, which is often just quoted broadly by panel manufacturers as something like "~95% DCI-P3" or ">90% DCI-P3" or similar. Take marketing specs with a pinch of salt as ever.

It is possible to get some information about a monitors colour gamut from the display's EDID information too. Extended Display Identification Data (EDID) is a metadata format for display devices to describe their capabilities to a video source (e.g. graphics card or set-top box). The data format is defined by a standard published by the Video Electronics Standards Association (VESA).

The free VESA DisplayHDR test tool is a useful tool for some testing elements, but also provides some information read from the monitor EDID. Below for example is the report from the LG 38GL950G (a Native G-sync screen):

In this section in the middle it reports the RGB primary xy numbers and from there it accurately calculates the gamut coverage for several popular and common colour spaces (relative to CIE-1976) including sRGB, Adobe RGB, DCI-P3 and Rec.2020. We have validated that these calculations are correct incidentally using the tools discussed later in this article. This could in theory be a useful and quick way to establish the colour gamut capabilities of the display, although the crucial thing here is to consider whether you can trust the EDID information or not?

The EDID information is provided by the display manufacturer so in one sense you might want to put as much faith in to that spec as you might in to any other spec they market for the display! This information could potentially be very hit and miss from display to display and from one manufacturer to another. It is always a good idea to test the gamut independently if you can, or refer to review sites like TFTCentral who will measure it as well independently.

In some cases you might be able to put more faith in this information contained within the EDID. For instance we know that Native NVIDIA G-sync screens are factory colour calibrated as part of the strict NVIDIA process with highly accurate equipment and processes, and the information recorded in detail in the EDID. It is likely that many other non-G-sync screens might not update this EDID information accurately in the factory in this same way. There is a high chance some values could be average figures, desired/expected, or even just made up. As a result it's always better to independently verify the results if you can with a highly accurate method, or refer to review sites who can do that for you in their testing.

It's worth also keeping in mind that this EDID information is only available for a single colour space mode, usually the native backlight colour space and so even if this was highly accurate information based on real testing in the factory that you could believe in, it's not possible to know how it might perform in other modes like specific sRGB emulation modes etc. We will continue to test monitor gamut independently in various modes (using our new more accurate method described below) rather than rely on manufacturer EDID information, although we may well cross-refer from time to time to see how accurate this EDID information tends to be.

Measuring Colour Gamut Accurately

Getting the accurate gamut coverage number for a monitor is challenging. It requires both high end measurement equipment and a careful measurement procedure. In the past we have used an X-rite i1 Pro 2 spectrophotometer combined with ChromaPure's software to measure colour gamut in sRGB, DCI-P3 and Rec.2020. This provides a pretty decent measurement of colour gamut, but we believe we can improve it and be more thorough.

We are now updating our testing methodology to include tests with a new spectroradiometer device from UPRtek (their MK550T) to increase our accuracy and provide further measurements in our reviews. Look out for an article soon where we look at this new device, its software and the various other useful and new things it will allow us to do. For now we will just talk about how it can help improve our colour gamut measurements and accuracy. This is an expensive but well regarded device and should provide improved levels of accuracy when measuring colour spaces..

By measuring the x,y primaries of RGB produced by the monitor with the high levels of accuracy that this spectroradiometer device can provide (in various modes if necessary), we can use a special calculation tool to work out the true gamut coverage for both CIE-1931 and CIE-1976. Combining the super high accuracy of the UPRtek device with this new method produces the most accurate measurements of the monitors true gamut coverage.

Example Tests

Example tests with the LG 38GL950G

We carried out the tests on the LG 38GL950G which offers a wide colour gamut Nano IPS panel. It also has a working sRGB emulation mode. This is a Native G-sync module screen which also gives us a chance to sanity check the EDID information against our measurements which we believe should be accurate given the factory calibration and measurement process NVIDIA perform. We measured the screen with our new UPRtek device and gamut calculation tool as described above. The following results were obtained:

x,y measurements taken with the UPRtek spectroradiometer in native gamut mode were:

# color, x, y (CIE-1931)
Red , 0.6840, 0.3086
Green, 0.2569 ,0.6691
Blue , 0.1475, 0.0584

and in sRGB emulation mode:

# color, x, y (CIE-1931)
Red , 0.6426, 0.3269
Green, 0.2987 ,0.6035
Blue , 0.1492, 0.0586

		Old measurement method		Measured CIE-1931 coverage new method		Measured CIE-1976 coverage new method
Colour space	Manufacturer spec quoted	Native gamut mode	sRGB mode	Native gamut mode	sRGB mode	Native gamut mode	sRGB mode
sRGB %	135%	130.9%	97.7%	100.0%	100.0%	100.0%	100.0%
Adobe RGB %	-	-	-	89.5%	74.8%	94.8%	86.2%
DCI-P3 %	98%	96.5%	72.0%	95.1%	75.2%	98.1%	81.9%
Rec.2020	-	69.2%	51.7%	70.9%	54.0%	75.9%	59.7%

Let's talk through the results to explain what they all mean and what they tell us.

Native wide mode

• sRGB coverage - Firstly the manufacturer quotes a 135% sRGB coverage in an effort to provide a comparison point against other screens quoted in a similar way, so you can tell that it has an even wider colour space coverage than a screen that might be 120% sRGB for example. It also highlights the fact that it is a decent way beyond a normal "standard gamut" sRGB screen (typically quoted as 97 - 100% sRGB). We've talked earlier about whether or not quoting over-coverage in this way is appropriate or can lead to errors, but our view is that it's a useful additional spec. In our original review we were able to confirm using our i1 Pro 2 device and ChromaPure software a 130.9% sRGB coverage so this was close to the spec. Measurement with the UPRtek spectroradiometer device and gamut calculation software will not allow over-coverage calculation in the same way, so we only confirm here that 100% coverage of the sRGB space is achieved regardless of whether you are considering CIE 1931 or 1976. This was not checked with our old method, as only an over-coverage spec was calculated, without consideration for if the sRGB gamut was covered exactly and fully. The UPRtek device allows us to confirm this more thoroughly. So in this instance, given 100% is covered, the 130.9% over-coverage spec is valid too. So we can summarise that the 38GL950G in its native wide gamut mode can successfully cover 100% of the sRGB reference, and extends beyond that to around 130.9% for comparison purposes with other screens.
   

• Adobe RGB coverage - this isn't quoted by the manufacturer or measured with our old i1 Pro 2 device/ChromaPure process, but using the new device and method we can confirm that the native gamut covers 89.5% (CIE-1931) or 94.8% (CIE-1976) of the Adobe RGB reference. This is a useful additional measurement now possible and relevant to those doing colour critical work/photo work/printing.
   

• DCI-P3 coverage - The manufacturer quotes 98% coverage. Using the UPRtek device we can confirm that 98.1% of the CIE-1976 DCI-P3 space is covered. It looks like LG spec is based on CIE-1976 although they don't stipulate, perhaps because this gives you the higher number? This means that there is a small portion of the reference space that is not produced, so it can't cover 100% of DCI-P3. Our i1 Pro/ChromaPure tests measured a smaller 96.5% which must be down to small inaccuracies in the i1 Pro device relative to the UPRtek spectroradiometer
   

• Rec.2020 coverage- LG don't provide a spec for this. Our original method calculated 69.2% coverage and we saw similar results from the UPRtek device as well, although a bit higher at 75.9% relative to CIE-1976.

sRGB mode

The screen has an sRGB emulation mode as well which is designed to restrict the full backlight gamma and provide a more "standard gamut" appearance for those who might want to work with sRGB content, or avoid oversaturation or unrealistic colour tones for certain uses.

• sRGB coverage - using our original method we recorded a 97.7% coverage in this mode. With the new method this was more accurately recorded 100% coverage regardless of if you are referring to CIE-1931 or 1976. This is is a more accurate method of measurement thanks to the new more advanced device which confirms a slightly more accurate sRGB emulation than we had originally measured.
   

• DCI-P3 coverage - In this sRGB emulation mode this greatly reduced the colour gamut of the screen. Our original method recorded 72.0% coverage, now more accurately calculated at 75.21% (CIE-1931 comparison) or 81.9% (CIE-1976). You can see why it's important to know whether CIE-1931 or 1976 are being used in specs and measurements as the figures are very different. Again the new device offers improved accuracy.

We won't provide all these measurements in our future reviews, we will think of a simpler and easier way to present figures that you can easily refer to and rely on.

Below this is the information reported from the EDID in the VESA Display HDR tool which provides CIE-1976 coverage calculations. This is listed as the Native gamut as well:

The EDID information being reported by the screen is based on the Native gamut mode based on the figures listed, and our own measurements. There is a very good correlation between the manufacturer provided stats in the EDID and our own measurements which is great news.

Example tests with the Asus ROG Swift PG329Q

We carried out the same tests on the recently reviewed Asus ROG Swift PG329Q which had a particularly wide colour gamut. This is an adaptive-sync screen (i.e. not a Native G-sync module screen like the LG 38GL950G is) which also gives us a chance to sanity check the EDID information again and see whether that is accurate, or is less reliable than the NVIDIA calibrated EDID information from Native G-sync screens.

We measured the screen with our new UPRtek device and gamut calculation tool as described above. The following results were obtained:

Let's talk through the results to explain what they all mean and what they tell us.

Native wide mode

• sRGB coverage - Firstly the manufacturer quotes a 160% sRGB coverage in an effort to provide a comparison point against other screens quoted in a similar way, so you can tell that it has an even wider colour space coverage than a screen that might be 130% sRGB for example. It also highlights the fact that it is a long way beyond a normal "standard gamut" sRGB screen (typically quoted as 97 - 100% sRGB). We've talked earlier about whether or not quoting over-coverage in this way is appropriate or can lead to errors, but our view is that it's a useful additional spec. In our original review we were able to confirm using our i1 Pro 2 device and ChromaPure software a 157.4% sRGB coverage so this was close to the spec. Measurement with the UPRtek spectroradiometer device and gamut calculation software will not allow over-coverage calculation in the same way, so we only confirm here that 100% coverage of the sRGB space is achieved regardless of whether you are considering CIE 1931 or 1976. This was not checked with our old method, as only an over-coverage spec was calculated. The UPRtek device allows us to confirm this more thoroughly. So in this instance, given 100% is covered, the 157.4% over-coverage spec is valid too. So we can summarise that the PG329Q in its native wide gamut mode can successfully cover 100% of the sRGB reference, and extends beyond that to around 157.4%.
   

• Adobe RGB coverage - this isn't quoted by the manufacturer or measured with our old i1 Pro 2 device/ChromaPure process, but using the new device and method we can confirm that the native gamut covers 100% of the Adobe RGB reference too (again regardless of whether we consider CIE-1931 or 1976). This is a useful additional measurement now possible and relevant to those doing colour critical work/photo work/printing.
   

• DCI-P3 coverage - The manufacturer quotes 98% coverage, and this is one area where it is a little more tricky. Using the UPRtek device we can confirm that 98.1% of the CIE-1976 DCI-P3 space is covered. It looks like Asus' spec is based on CIE-1976 although they don't stipulate, perhaps because this gives you the higher number? This means that there is a small portion of the reference space that is not produced, so it can't cover 100% of DCI-P3. Our i1 Pro/ChromaPure tests confirm that the gamut extends beyond that reference area to around 116.0%. If we are being strict, this isn't quite true as it can't cover the full 100%, being only 98.1% covered. The 116.0% shows it reaches beyond the DCI-P3 space a fair bit, but doesn't tell the full picture. So we would summarise here that 98.1% of the DCI-P3 reference is covered, but if we were to consider over-coverage area it extends to approx 116.0%. The new method allows us to more accurately confirm DCI-P3 coverage, including with higher levels of accuracy.
   

• Rec.2020 coverage- Asus don't provide a spec for this. Our original method calculated 83.3% coverage and we saw similar results from the UPRtek device as well.

sRGB mode

The screen has an sRGB emulation mode as well which is designed to restrict the full backlight gamma and provide a more "standard gamut" appearance for those who might want to work with sRGB content, or avoid oversaturation or unrealistic colour tones for certain uses.

• sRGB coverage - using our original method we recorded a 95.2% coverage in this mode. With the new method this was more accurately recorded at 97.1 / 97.0% depending on if you are referring to CIE-1931 or 1976. Either way this is a more accurate method of measurement thanks to the new more advanced device.
   

• DCI-P3 coverage - In this sRGB emulation mode this greatly reduced the colour gamut of the screen. Our original method recorded 70.2% coverage, now more accurately calculated at 73.1% (CIE-1931 comparison) or 79.9% (CIE-1976). You can see why it's important to know whether CIE-1931 or 1976 are being used in specs and measurements as the figures are very different. Again the new device offers improved accuracy.

We won't provide all these measurements in our future reviews, we will think of a simpler and easier way to present figures that you can easily refer to and rely on.

Below this is the information reported from the EDID in the VESA Display HDR tool which provides CIE-1976 coverage calculations. This is listed as the Native gamut as well:

The EDID information being reported by the screen appears to be based on the sRGB mode based on the figures listed, and our own measurements. This is fair enough, although it means it can't confirm the full native gamut figures for you which is probably the more common usage scenario than sRGB mode. There is however a pretty good correlation, within 2 - 3%) between the manufacturer provided stats in the EDID and our own measurements which is good news.
 

Further reading

For lots more detailed discussion about colour spaces please also see our Pointer's Gamut article.