Choosing A Monitor

[xpatUSA]

First your image of the specs of the shown monitor.
It says it has a certain color gamut. Meaning it can show colors within a range of x to y wavelength, analogue.

You can't say that, sorry.

Although a wavelength does have a specific color, that color does not have units of wavelength in the normal model expressions of the term "color". Instead, a color is expressed in 2 or 3D terms, e.g. CIELAB a*, b* (2D) or e.g. HSB (3D). And the gamuts of Manfred's monitor are normally expressed as an area on a 2D chart, or as a volume in a 3D chart.

Just so we're clear, a monochromatic hue of "red", say 630nm, has many possible colors extending from the gamut's white point outward to the gamut's boundary.

I'd post a diagram, but . . .

[xpatUSA]

Just so we're clear, a monochromatic hue of "red", say 630nm, has many possible colors extending from the gamut's white point outward to the gamut's boundary.

I'd post a diagram, but . . .

"I'd post a diagram, but . ."

. couldn't resist, sorry, here's an example showing a yellowish-green hue (the purple line), approx 560nm:



On top of an Adobe Wide Gamut (not Adobe RGB [1998]) I've sketched a tristimulus gamut and also an hypothetical "round" gamut. The four dots crossing the gamut boundaries and the limit line-of-human-vision are all the same hue but do all have different colors as specified by CIE. That is to say that, according to CIE's original definition of any specific color (XYZ chromaticity co-ordinates), there are four different colors specified by the dots (five if you include white).

Furthermore, as George's wavelength line extends beyond CIE's 1931 xyY boundary, it becomes invisible to humans and is, therefore, not even a color at that point.

[Manfred M]

First your image of the specs of the shown monitor.
It says it has a certain color gamut. Meaning it can show colors within an range of x to y wavelength, analogue.
Further it says that it can show that range in parts of 2^10 peaces of a single channel. Using a 10 bit D/A conversion.
At last it says that combining the 3 channels R,G and B will produce (2^10)^3= 1.07 billion different gradients WITHIN that gamut.
Definitions are important for a good communication. I'm not aware of any definition of gamut that includes the bitdepth.

About your reaction on the link I posted, I don't understand what you are trying to prove. I'm not going to defend him, but I must say he has a strukturell approach of the subject.
I just don't see much difference between your link and mine. So I'm very curious what points are wrong in your opinion.

George

It could be just a language issue George.

My understanding is that you are theoretically correct; one can implement any colour space using any bit depth, but we are stuck with whatever implementation the computer screen manufacturer has decided to work with. There is a technical issue with using lower bit depth with a wide gamut colour space, and while I have not seen it when working in 8-bit Adobe RGB, I have when using 8-bit with ProPhoto. The colour range between two data points can be so large that "blocking", i.e. poor colour transition can occur. It is one colour space where using more than 8-bit is mandatory. I suspect the reason that the screen manufacturers have gone with 10-bit implementations is likely for the same reason, although the colour space is smaller than ProPhoto RGB, it could still occur.

So while there is no mathematical reason to suggest a specific minimum bit depth when using a specific colour space, there is definitely a practical one.

With respect the Arnaud Frich's website. I have seen it before and came away unimpressed because he does not appear to understand some of the basics and we see this in his screen recommendations. Unless the spec sheet specifically mentions that it is sRGB (or AdobeRGB compliant), it is likely not, and the colour output cannot be guaranteed to be accurate. Colour accuracy should be of prime importance to someone who is suggesting he is an expert on the subject. As an example, my Dell P2210 screen specs claim that it shows 16.7 million colours, but when I did some detailed testing, it is NOT sRGB compliant and the colours, especially the blues are not rendering correctly.

[george013]

It could be just a language issue George.

My understanding is that you are theoretically correct; one can implement any colour space using any bit depth, but we are stuck with whatever implementation the computer screen manufacturer has decided to work with. There is a technical issue with using lower bit depth with a wide gamut colour space, and while I have not seen it when working in 8-bit Adobe RGB, I have when using 8-bit with ProPhoto. The colour range between two data points can be so large that "blocking", i.e. poor colour transition can occur. It is one colour space where using more than 8-bit is mandatory. I suspect the reason that the screen manufacturers have gone with 10-bit implementations is likely for the same reason, although the colour space is smaller than ProPhoto RGB, it could still occur.

So while there is no mathematical reason to suggest a specific minimum bit depth when using a specific colour space, there is definitely a practical one.

With respect the Arnaud Frich's website. I have seen it before and came away unimpressed because he does not appear to understand some of the basics and we see this in his screen recommendations. Unless the spec sheet specifically mentions that it is sRGB (or AdobeRGB compliant), it is likely not, and the colour output cannot be guaranteed to be accurate. Colour accuracy should be of prime importance to someone who is suggesting he is an expert on the subject. As an example, my Dell P2210 screen specs claim that it shows 16.7 million colours, but when I did some detailed testing, it is NOT sRGB compliant and the colours, especially the blues are not rendering correctly.

Definitions are the basics of language, be it gamut or just the color red. If you state there are more definitions then communication is becoming difficult.

As far as I know there's no objection by connecting a wide gamut screen on an "old " video cart when there're the appropriate connections. From what I understood is that when the card is 8 bit and the monitor 10 bit, the less significant bits are truncated.

Back to the thread. That article is dealing 8 different points to think off when buying a new monitor and he's dealing with them in a right way. What struck me was the impact a wide gamut monitor will have on the pc. All the programs including the os will be shown on that wide gamut screen. And all of those that doesn't have color corrections will show them in a wrong color. So using a wide gamut screen on a multi purpose pc will add another problem.

As this monitor was mentioned before, he has a review of that Benq SW2700(PT?). He also mentioned the eventually problem with homogenity. https://www.color-management-guide.c...or-review.html

George

[Manfred M]

Definitions are the basics of language, be it gamut or just the color red. If you state there are more definitions then communication is becoming difficult.

I don't think there is an issue with the term "gamut". My issue is with the term "wide gamut"; where there is no clear definition of this term.

As far as I know there's no objection by connecting a wide gamut screen on an "old " video cart when there're the appropriate connections. From what I understood is that when the card is 8 bit and the monitor 10 bit, the less significant bits are truncated.

I am not aware of any modern 8-bit cards, so the argument is a bit moot. 10-bit cards have been around so long that I suspect that this is probably part of the PCI-E standard. Some screens still have VGA connectors, although I don't understand why anyone would use them; they were designed for use with the obsolete CRT (picture tube) screens that have not been around for a long time. I suspect that the case you make is more academic than practical, from a hardware implementation standpoint. Neither my graphics card nor my screen accepts VGA connections.

Again, I don't quite understand your point. The theoretical issues of 8-bit versus 10-bit can be discussed all you want, but in the real world we have to deal with how the hardware manufacturers have implemented their solutions. Wide-gamut = 10-bit.

Back to the thread. That article is dealing 8 different points to think off when buying a new monitor and he's dealing with them in a right way.

He misses the most important aspect of the whole exercise. My number one question is "it the screen 100% sRGB compliant". If the answer is NO, then that computer screen should NEVER be considered for image editing because one cannot guarantee colour accuracy. The fact that he does not seem to understand this, is extremely worrisome. A screen that does not meet this very basic requirement will never display the colours properly and no amount of calibration / profiling can bring it into compliance.

The other critical point with a screen for photo editing is the contrast ratio. This does not seem to be important to him either.

What struck me was the impact a wide gamut monitor will have on the pc. All the programs including the os will be shown on that wide gamut screen. And all of those that doesn't have color corrections will show them in a wrong color. So using a wide gamut screen on a multi purpose pc will add another problem.

We are looking at a non-issue here George. A wide gamut screen has no issues displaying a narrow gamut image properly, so long as the software is capable of identifying the colour space that the image has used. This data is embedded in the image file and as long as the browser (for internet based images) or display software for anything else recognizes the colour space, the operating system takes care of the mappings for us.

There are two issues; not all software is equally good at handling colour spaces correctly and mistakes can be made when assigning the appropriate colour space in an image. The first issue is easy to take care of by using software that is properly designed.

The second issue can be more problematic when viewing images on the internet; if the wrong colour space is assigned, the colours will not be displayed properly. Experienced users would not make this mistake; inexperienced users, that is something we can't do much about. In reality I run into this very rarely in real life. I do know there are some "experts" that suggest sticking to sRGB is the easiest way to avoid that issue; but I definitely don't agree with this approach.

As this monitor was mentioned before, he has a review of that Benq SW2700(PT?). He also mentioned the eventually problem with homogenity. https://www.color-management-guide.c...or-review.html

He mentions that this is a reported problem with some units. I did check mine out and it is not one of them.

There are other reviews out there that are not in line with his findings. In fact when I was doing the research before buying my screen, I found none that identified this problem.

http://www.tomshardware.com/reviews/...or,4374-8.html

Unfortunately, he seems to be and Apple / Eiso / NEC "fanboy" and that is obvious in his writing. These companies do make fine products, but they are not the only ones. We should be wary of writers who do not seem to understand the subject matter they are covering and who do not seem to understand their own biases.

[xpatUSA]

My issue is with the term "wide gamut"; where there is no clear definition of this term.

Looks like you're right. One found in Wiki-land is no help at all:

https://en.wikipedia.org/wiki/Gamut#Wide_color_gamut

By using the phrase as a descriptor, Adobe has perhaps created a de facto standard, a sort of Chuck Norris thing. Here it is at top right, along with some better-known actual standards:



Rec.2020 seems similar to Adobe's effort, except for the white point. It is aimed at fancy TV gamuts.

Rec.709 is the basis of sRGB.

I've read that the green primary in Adobe RGB (1998) was a mistake, not actually intended to make greenery look pretty . .

Wide Gamut = 10 bit

With all due respect, that could have been phrased differently. This thread has already discussed the matter of 10 bits, and the statement runs counter to that discussion; it implies that wide gamut is only 10 bit and can not be displayed by any other bit depth. I do understand the hardware angle, of course.

I view "wide gamut" in same light as "resolution":




P.S. Only in the world of photography would "wide" (a 1D linear or angular measure) be used to describe a 2D or 3D object !!

[george013]

Post #18

It's all about definitions George. If wide gamut is defined as a colour space that is wider than sRGB and bit depth greater than 8-bits, you have your answer. This is definitely the definition that is used by manufacturers of display equipment. Take a look at the specs of wide gamut screens and you will see this is true

That was and is my problem with the definition. I thought that has been solved in post#23.

George

[dje]

Looks like you're right. One found in Wiki-land is no help at all:

https://en.wikipedia.org/wiki/Gamut#Wide_color_gamut

By using the phrase as a descriptor, Adobe has perhaps created a de facto standard, a sort of Chuck Norris thing. Here it is at top right, along with some better-known actual standards:



Rec.2020 seems similar to Adobe's effort, except for the white point. It is aimed at fancy TV gamuts.

Rec.709 is the basis of sRGB.

I've read that the green primary in Adobe RGB (1998) was a mistake, not actually intended to make greenery look pretty . .

Yes I think wide gamut is really a fairly general term however with present technology. a monitor that covers Adobe RGB is considered wide gamut as a result of common usage of the term.

For me, the interesting thing about the two color spaces on the right hand side of Ted's photo, ie AdobeWideRGB and the ITU Rec2020 one below it, is that the primaries are defined as monochromatic colors ie single wavelength. This is indicated by their location on the very edges of the CIE xy chromaticity diagram (which is a locus of single wavelength colors). To achieve these primaries in real life with LCD/LED pixel technology would be quite difficult I think.

Dave

[Manfred M]

Yes I think wide gamut is really a fairly general term however with present technology. a monitor that covers Adobe RGB is considered wide gamut as a result of common usage of the term.

I tend to agree Dave when it comes to screens, but when discussing colour spaces, I rarely hear AdobeRGB referred to as a wide-gamut colour space, whereas ProPhoto RGB and ColorMatch RGB are. Throw in modern photo ink jet printers with their additional colours above the standard CMYK base, these can exceed the AdobeRGB colour space as well.

Hence some of the confusion.

[xpatUSA]

For me, the interesting thing about the two color spaces on the right hand side of Ted's photo, ie AdobeWideRGB and the ITU Rec2020 one below it, is that the primaries are defined as monochromatic colors ie single wavelength. This is indicated by their location on the very edges of the CIE xy chromaticity diagram (which is a locus of single wavelength colors).

Dave

There seems to be some misunderstanding as to how color space primaries are defined.

The definition of a primary is by x,y co-ordinates on a CIE xyY diagram. x,y co-ordinates do have a "dominant" wavelength but a wavelength does not define any particular pair x,y co-ordinates.

https://en.wikipedia.org/wiki/Wide-g...GB_color_space

https://en.wikipedia.org/wiki/Rec._2020

Here's a rhetorical question: a green color-space primary has a wavelength of about 550nm - what are the CIE x,y co-ordinates of that primary?

Rhetorical, because it can not be answered. Those co-ordinates could anywhere on a straight line from the white point to the edge of the human visual gamut.

A less rhetorical question:

A monochromatic primary has co-ords of x,y = 0.3, 0.6 and a white point of D65 - what is it's wavelength?

Answer can be found here, thereby avoiding a tiresome calculation:

http://www.brucelindbloom.com/index....alculator.html

Put more simply: "wavelength is a property of x,y" but "x,y is not a property of wavelength". Sorry, Dave.

[dje]

I tend to agree Dave when it comes to screens, but when discussing colour spaces, I rarely hear AdobeRGB referred to as a wide-gamut colour space, whereas ProPhoto RGB and ColorMatch RGB are. Throw in modern photo ink jet printers with their additional colours above the standard CMYK base, these can exceed the AdobeRGB colour space as well.

Hence some of the confusion.

Indeed Manfred. ProPhoto RGB has a very wide gamut, probably wider than the Adobe Wide gamut RGB space referred to by Ted above. It achieves this wider gamut by using imaginary colors as primaries (see the green and blue primaries in the diagram below) and is fine for allowing a wider range of colors during photo editing. However it can never be directly implemented in an RGB monitor because it's primaries simply can't be produced in an actual device.

Dave

[dje]

There seems to be some misunderstanding as to how color space primaries are defined.

The definition of a primary is by x,y co-ordinates on a CIE xyY diagram. x,y co-ordinates do have a "dominant" wavelength but a wavelength does not define any particular pair x,y co-ordinates.

https://en.wikipedia.org/wiki/Wide-g...GB_color_space

https://en.wikipedia.org/wiki/Rec._2020

Here's a rhetorical question: a green color-space primary has a wavelength of about 550nm - what are the CIE x,y co-ordinates of that primary?

Rhetorical, because it can not be answered. Those co-ordinates could anywhere on a straight line from the white point to the edge of the human visual gamut.

A less rhetorical question:

A monochromatic primary has co-ords of x,y = 0.3, 0.6 and a white point of D65 - what is it's wavelength?

Answer can be found here, thereby avoiding a tiresome calculation:

http://www.brucelindbloom.com/index....alculator.html

Put more simply: "wavelength is a property of x,y" but "x,y is not a property of wavelength". Sorry, Dave.

Ted you might be reading more into my statement than what i intended to convey. Here is a quote from the Wiki article on rec2020.

"The RGB primaries used by Rec. 2020 are equivalent to monochromatic light sources on the CIE 1931 spectral locus.[7][8] The wavelength of the Rec. 2020 primary colors is 630 nm for the red primary color, 532 nm for the green primary color, and 467 nm for the blue primary color.["

Perhaps I should have used the words "equivalent to" when talking about the relationship between primaries and wavelengths.

Dave

[xpatUSA]

Ted you might be reading more into my statement than what i intended to convey.

And maybe I was too pedantic into the bargain.

Here is a quote from the Wiki article on rec2020.

"The RGB primaries used by Rec. 2020 are equivalent to monochromatic light sources on the CIE 1931 spectral locus. The wavelength of the Rec. 2020 primary colors is 630 nm for the red primary color, 532 nm for the green primary color, and 467 nm for the blue primary color."

Perhaps I should have used the words "equivalent to" when talking about the relationship between primaries and wavelengths.

Dave

"equivalent to" works for me, I reckon!

[Manfred M]

However it can never be directly implemented in an RGB monitor because it's primaries simply can't be produced in an actual device.

I suspect that the OOG colours that don't exist are more of a theoretical rather than real issue. As we are dealing with real world cameras, they can only capture frequencies that exist, so even though the colour space allows for it, how can one capture colours that do not exist? Those values could, in theory, be generated by PP algorithms, but wouldn't this be somewhat akin to OOG colours and using rendering intents to deal with them?

[dje]

I suspect that the OOG colours that don't exist are more of a theoretical rather than real issue. As we are dealing with real world cameras, they can only capture frequencies that exist, so even though the colour space allows for it, how can one capture colours that do not exist? Those values could, in theory, be generated by PP algorithms, but wouldn't this be somewhat akin to OOG colours and using rendering intents to deal with them?

Manfred I don't think the imaginary colors are of any real concern and yes rendering intents should deal with it. I was just making the point that for a system based on three primaries, the boundaries of a color space on the CIE xy diagram are a triangle linking the three primary points. So if your primaries are generated by a real device such as a monitor, the best you can do is something along the lines of the Adobe Wide gamut RGB space above where the primaries are located on the boundary of the visible color area and positioned to cover as much of the visible space as possible. Even this would be difficult to implement in a real device because the monitor pixels would have to generate monochromatic light.

Dave

[xpatUSA]

I suspect that the OOG colours that don't exist are more of a theoretical rather than real issue. As we are dealing with real world cameras, they can only capture frequencies that exist . .

"frequencies", Manfred? of the light? If so, why the change from more easily-understood wavelength? If not, what frequencies?

. so even though the colour space allows for it, how can one capture colours that do not exist?

I may have missed the point, but it's easy. I can remove the UV/IR blocking filter from my Real World SD14 thereby allowing the sensor to capture wavelengths from ~280nm to ~1150nm. Placing an IR filter on the lens removes "light" (anything </=700nm) from the capture - leaving a capture with no colors in it but, nevertheless, a capture.

This might be of interest:



http://kronometric.org/phot/gamut/Ca...ysisGamuts.pdf

From practical measurements - not theoretical at all!

Most un-modified Real World cameras do pass some IR, i.e. non-colors, to the sensor as evidenced by those IR shots with a lens filter and super-long exposure times. Most such captures are converted in post either to monochrome or faux color of course and by so doing, end up well within any color gamut on the planet - Even Rec.601 -

HTH.

[Manfred M]

"frequencies", Manfred? of the light? If so, why the change from more easily-understood wavelength? If not, what frequencies?

Late at night moment, I guess. Both are properties of light (or any wave). Wavelength is m and frequency is Hz.

I may have missed the point, but it's easy. I can remove the UV/IR blocking filter from my Real World SD14 thereby allowing the sensor to capture wavelengths from ~280nm to ~1150nm. Placing an IR filter on the lens removes "light" (anything </=700nm) from the capture - leaving a capture with no colors in it but, nevertheless, a capture.

This might be of interest:



http://kronometric.org/phot/gamut/Ca...ysisGamuts.pdf

From practical measurements - not theoretical at all!

Most un-modified Real World cameras do pass some IR, i.e. non-colors, to the sensor as evidenced by those IR shots with a lens filter and super-long exposure times. Most such captures are converted either to monochrome or faux color of course and by so doing, ending up well within any color gamut on the planet - Even Rec.601 -

HTH.

I understand the IR and UV impact; these are parts of the electromagnetic spectrum that we humans cannot see, but our sensors are sensitive to, especially in the near IR and UV. That's what the cutoff filters are for and yes, these are not perfect so we see some of these wavelengths affecting our images, especially in long exposures.

I expect that there are technical solutions that can be utilized to exclude wavelengths we cannot see possibly through the use of rendering intents or cutoff filters during signal processing. On the other hand, as Dave has mentioned, I don't see any real major breakthroughs in display technology on the horizon that will give us the colour ranges we are looking at in this discussion. In prints, we are moving in that direction as printers are exceeding the colours that screens can display.

[xpatUSA]

I've read that the green primary in Adobe RGB (1998) was a mistake, not actually intended to make greenery look pretty . .

Found it . . .

Since the PAL / SECAM television standard existed first, it is logical to assume that the other three derived from it. I have heard the rumor that the green primary for Adobe RGB came about by the accidental use of the NTSC green primary, used incorrectly since NTSC is defined relative to Illuminant C while Adobe RGB is defined relative to D65. After the mistake was discovered, Adobe decided to keep it since their experiences with this accidental reference space were favorable.

http://www.brucelindbloom.com/index.html?WorkingSpaceInfo.html

Although users loved the wider range of reproducible colors, those familiar with the SMPTE 240M specifications contacted Adobe, informing the company that it had copied the values that described idealized primaries, not actual standard ones. The real values were much closer to sRGB's, which avid Photoshop consumers did not enjoy as a working environment. To make matters worse, an engineer had made an error when copying the red primary chromaticity coordinates, resulting in an even more inaccurate representation of the SMPTE standard.

https://en.wikipedia.org/wiki/Adobe_RGB_color_space

Two versions, one green, one red! Take yer pick . .

[xpatUSA]

. . . On the other hand, as Dave has mentioned, I don't see any real major breakthroughs in display technology on the horizon that will give us the colour ranges we are looking at in this discussion.

How about the 30-year old OLED technology?

"The model covers 100% of the AdobeRGB colors, 100% sRGB, 100% Rec. 709, 97.5% DCI-P3 and 85.8% Rec. 2020"

http://laptopmedia.com/news/dell-fin...rother-up3017/

"On April 17, 2016, Sony presented a 55 in (140 cm) 4K OLED display with the support of Rec. 2020 color space."

https://en.wikipedia.org/wiki/Rec._2020 (again).

I just posted something similar but it seems not to have got there . .

In prints, we are moving in that direction as printers are exceeding the colours that screens can display.

[Manfred M]

How about the 30-year old OLED technology?

Yes and every year we hear that the manufacturers have mastered the production process for an affordable price. And every year so far we have been disappointed. Let's see if 2017 or even 2018 is finally the year of the breakthrough. Once Samsung, LG, et al have mastered the TV screen production, we might end up seeing these in computer screens. The contrast is great the colours vibrant, but can these units do subtlety?

TV watchers have a different set of requirements than people working on computer screens in an office setting versus those working digital still images.