sRGB vs Adobe RGB How wrong am I?

[Abitconfused]

I imagine that the sRGB color space contains about 96% of all the colors a photographer might encounter in everyday life. How wrong am I?

[pschlute]

I imagine that the sRGB color space contains about 96% of all the colors a photographer might encounter in everyday life. How wrong am I?

I guess it depends on the term "encounter" , and possibly the species of the photographer

My understanding (very layman) is that sRGB can depict around 17 million colours. The average human can perceive between 1 and 3 million. Certain fish can detect up to 100 million (would love to know how they know this !).

SRGB like any "gamut" is a scientific construct. The purpose of a gamut is to ensure compliance throughout a process ...from taking a picture ...processing it...printing or publishing it on a screen.

ps. This is not my area of expertise

pps. for everyday photography (without doing extensive editing, or printing) sRGB is your best bet

[Manfred M]

all the colors a photographer might encounter in everyday life.

I have no idea what "everyday life" means and I have never seen any number suggesting that sRGB covers 96% of those colours.

When I do my gamut check, there are very few days where I don't see out of gamut issues in my images when I convert to sRGB.

[DanK]

The notion of a number of colors is a bit off. In digital photography, we discretize what I think is a continuous distribution. I think the more useful way to think about this is asking how much of the range of the color space humans can see is reproduced within the sRGB gamut. The relevant number would be a fraction of a volume. I have no idea what that number would be, although I suspect Ted knows. However, an appreciable amount of what we can see is outside of sRGB. Whether one encounters it on a given day depends on what one is photographing.

Peter wrote:

for everyday photography (without doing extensive editing, or printing) sRGB is your best bet

The reference to printing is important. Photo printers can reproduce substantially more than the sRGB color space. That's one reason why people who are really fussy about prints edit in a larger color space than sRGB.

[Manfred M]

My understanding (very layman) is that sRGB can depict around 17 million colours. The average human can perceive between 1 and 3 million.

Carefully said, your statement is a bit misleading. The 16.8 million colours for sRGB include a lot of shades that humans cannot differentiate.

When we look at the total range of human vision, we can see far more than what any of the colour spaces commonly used in photography (sRGB, Adobe RGB and ProPhoto RGG, P3) as the volume of the surface of human vision is greater than any of them. The number of individual shades that a human can see, was up to around 10 million, but I'll be frank here, the sources I've seen are all over the place.

for everyday photography (without doing extensive editing, or printing) sRGB is your best bet

I would disagree here; any colour space that is wider than sRGB is usually recommended for photo work. A lot of trained photographers that I know default to either Adobe RGB ProPhoto RGB when working and convert to sRGB when outputting to screen.

The original purpose of Adobe RGB was to emulate what commercial offset presses were capable of reproducing. Back in the day, most of the high end clients that Adobe served were the ones creating paper based magazines and catalogues with this process.

[xpatUSA]

I imagine that the sRGB color space contains about 96% of all the colors a photographer might encounter in everyday life. How wrong am I?

Hello Ed,

"All the colors a photographer might encounter in everyday life" is most often represented by 'Pointer's Gamut' - have a look here:

https://www.alamy.com/an-illustration-of-srgb-and-pointers-gamut-colour-spaces-overlaid-on-cie-1931-chromaticity-diagram-of-human-colour-perception-image262665446.html

Also here:

https://cinepedia.com/picture/color-gamut/

Both gamuts are shown in 2D. If we are thinking in terms of 2D, the illustration serves your purpose: just measure the area of each one and figure the percentage ratio or even just eyeball them. Looks to me like 96% is a bit optimistic.

Adobe RGB (1998) might be closer to 96% because it has more greens.

Others here will mention that "colors" are in a 3D space but I think that the 2D representation serves your question well enough ...

HTH.

[billtils]

If your default viewing mode is a computer screen and you share your images then sRGB is the way to go. sRGB was specifically develped by HP and Microsoft to provide a consistent reproduction of an acceptably wide gamut on PC monitors. If you are printing then perhaps the expanded colour space of Adobe or more recent colour spaces could interest you but any change due to selection of colour space will be small compared to the variety that comes with the various papers that you can use.

However, unless and until I can look at a colour image on its own and tell whether it is in a wider gamut colour space than sRGB I see little or no point in using anything else.

[pschlute]

However, unless and until I can look at a colour image on its own and tell whether it is in a wider gamut colour space than sRGB I see little or no point in using anything else.

But your conclusion here ignores the fact that editing an image in a wide gamut means smoother transitions, even when you eventually convert the image to sRGB for web publication or printing.

[xpatUSA]

... editing an image in a wide gamut means smoother transitions ...

Peter, is it possible to expand that statement because I don't quite follow it?

I do convert my 12-bit raws to 16-bit ProPhoto RGB and edit them with 32-bit floating point still in ProPhoto - so just wondering how that gets me "smoother transitions" for saving as 8-bit JPEGs.

[DanK]

But your conclusion here ignores the fact that editing an image in a wide gamut means smoother transitions

+1

However, unless and until I can look at a colour image on its own and tell whether it is in a wider gamut colour space than sRGB I see little or no point in using anything else.

You won't see it, of course, unless you use an output device that has a wider gamut than sRGB. If you are viewing on an sRGB device, the odds are that the device uses a perceptual rendering that preserves distances in the color space and looks quite natural--just not accurate. I usually use a wide-gamut monitor and print on a good printer using coated, wide gamut papers, and it's not all that uncommon for me to find colors that are outside of the sRGB gamut, particularly in the case of flowers, many of which have very intense colors.

sRGB was specifically develped by HP and Microsoft to provide a consistent reproduction of an acceptably wide gamut on PC monitors

Yes, just shy of 22 years ago. It was designed to accommodate CRTs, which no one has used for years. Display technology has improved a great deal since then. Even consumer goods are leaving the sRGB gamut behind, albeit slowly. E.g., the newer iPad displays can show the P3 gamut, which is larger than sRGB. Unfortunately, large wide-gamut computer monitors are still expensive, but that will eventually change.

[xpatUSA]

If you are viewing on an sRGB device, the odds are that the device uses a perceptual rendering that preserves distances in the color space and looks quite natural--just not accurate.

Unless things have changed recently, even if a profile of type 'mntr' has 'perpetual' selected, that will be ignored and the CMS will use Relative Colorimetric, whether you like it or not. That is because such profiles are normally of the matrix type and 'perpetual' rendition needs LUTs.

[Manfred M]

Unless things have changed recently, even if a profile of type 'mntr' has 'perpetual' selected, that will be ignored and the CMS will use Relative Colorimetric, whether you like it or not. That is because such profiles are normally of the matrix type and 'perpetual' rendition needs LUTs.

Profiling software and the drivers associated with it create LUTs during the profiling process. With higher end screens, the LUTs are stored in memory in the screen. Lower end ones have the LUTs stored on the system disk and this gets loaded during the computer start up.

I don't know what happens natively if the system has not been calibrated and profiled and those drivers are not installed; it is quite possible that the CMS does what you have written.

[pschlute]

Peter, is it possible to expand that statement because I don't quite follow it?

Aggghhh, I was hoping you were not going to ask that

My understanding is that when working on an image using say curves or colour sliders, any adjustments are going to be more gradual/smoother when working in a wider gamut..... ie less "banding" is one example.

My understanding is that even after finally converting to a narrower space like sRGB, any transitions between pixels arising from editing are likely to be smoother than if one started and finished with sRGB colour space.

[Manfred M]

But your conclusion here ignores the fact that editing an image in a wide gamut means smoother transitions, even when you eventually convert the image to sRGB for web publication or printing.

Peter, I'm afraid this is not correct. Bit depth and the size of the colour space are the key drivers of smooth transitions.

Assuming the same bit depth, the smaller the colour space, the closer the gap between individual colours and the smoother the transitions. If you edit in 16-bit sRGB, you will get the smoothest transitions.

One of the main arguments against using ProPhoto RGB is that the colour space is so large that individual colours are quite far apart for adjacent colour values. Push ProPhoto hard in 16 bit and you will see banding (i.e. poor transitions).

[Manfred M]

If your default viewing mode is a computer screen and you share your images then sRGB is the way to go. sRGB was specifically develped by HP and Microsoft to provide a consistent reproduction of an acceptably wide gamut on PC monitors. If you are printing then perhaps the expanded colour space of Adobe or more recent colour spaces could interest you but any change due to selection of colour space will be small compared to the variety that comes with the various papers that you can use.

However, unless and until I can look at a colour image on its own and tell whether it is in a wider gamut colour space than sRGB I see little or no point in using anything else.

Not quite true Bill.

sRGB was developed in 1996 and was based on the CRT screens that were available at the time (as Dan has pointed out). I don't recall having any discussions on the colour spaces at the time I was using those, but certain higher end screens seemed to have a wider gamut.

Adobe RGB was developed in 1998 in an attempt to accurately show what offset press output would look like on a standard computer screen. Offset presses with their CMYK process on high quality paper had a wider gamut than what was visible on a computer screen. I know some photographers who worked in the commercial field at that time and what one saw on the screen and what came off the press were often quite different. Adobe developed the Adobe RGB colour space in an attempt to show what press output looked like on the screen that post-processing was done on. These issues are what drove the need for a colour managed work flow; otherwise the skill of the pre-press operator was the key to getting the colours to look right in the magazines and catalogues of the day.

The general view I have been taught is to use a wider colour space than what the final product will use as the actual data captured by the camera (with appropriate raw conversion), which protects out of gamut colours and allows for more accurate work, even though one is still reliant on the gamut of the computer screen. Most knowledgeable photographers that I know suggest working in a wider colour space than sRGB to produce more accurate colours, even when working for images to be posted on the internet. The final output should be an sRGB conversion.

Things get more complicated when working on images to be printed on high end ink jet photo printers. Depending on the paper used, these devices are capable of producing output that exceeds the Adode RGB colour space. This is especially true for scenes with very saturated or vibrant colours. There is an argument that can be made for working in ProPhoto RGB, unless the image is pushed too hard (and poor colour transitions occur), when Adobe RGB is more appropriate. The P3 colour space (the default on many Apple products) is wider than sRGB but narrower than Adobe RGB and parts of the gamut (yellows and reds) exceed what colour printers can produce. I have been told, but cannot verify, that paper manufacturers base the profile data on the Adobe RGB colour space.

[Abitconfused]

OK! I get it. I have ironed my fingers. Ouch!! One more thing? Why not use the entire RGB color mode as a color space? I suppose that too many out of gamut colors would then be forwarded to the internet or printers...yes?

[pschlute]

Peter, I'm afraid this is not correct. Bit depth and the size of the colour space are the key drivers of smooth transitions.

Assuming the same bit depth, the smaller the colour space, the closer the gap between individual colours and the smoother the transitions. If you edit in 16-bit sRGB, you will get the smoothest transitions.

One of the main arguments against using ProPhoto RGB is that the colour space is so large that individual colours are quite far apart for adjacent colour values. Push ProPhoto hard in 16 bit and you will see banding (i.e. poor transitions).

Aha, thanks for the clarification. I have mis-understood this.

[billtils]

But your conclusion here ignores the fact that editing an image in a wide gamut means smoother transitions, even when you eventually convert the image to sRGB for web publication or printing.

Not ignored, just another instance of does it make a discernable difference? I don't think it does ...

[billtils]

+1



You won't see it, of course, unless you use an output device that has a wider gamut than sRGB. If you are viewing on an sRGB device, the odds are that the device uses a perceptual rendering that preserves distances in the color space and looks quite natural--just not accurate. I usually use a wide-gamut monitor and print on a good printer using coated, wide gamut papers, and it's not all that uncommon for me to find colors that are outside of the sRGB gamut, particularly in the case of flowers, many of which have very intense colors.



Yes, just shy of 22 years ago. It was designed to accommodate CRTs, which no one has used for years. Display technology has improved a great deal since then. Even consumer goods are leaving the sRGB gamut behind, albeit slowly. E.g., the newer iPad displays can show the P3 gamut, which is larger than sRGB. Unfortunately, large wide-gamut computer monitors are still expensive, but that will eventually change.

All correct Dan, and I'm not disputing any of that, just applying Occam's Razor. Sure it can be fun, and absolutely it's normal to not want to be left behind by 'progress', but at the end of the day we are making pictures and if you can't tell the dfference without a side-by-side (and often microscopical) scrutiny, does it matter?

[DanK]

I don’t work in a wide gamut because it’s more fun or because I worry about being left behind. I do it because sometimes—not always, by any means—it gets me better-looking prints. It’s the same reason I have spent a lot of time carefully picking the papers I use.

If you mostly post on the web and print only casually, it may not make an appreciable difference for you.


Sent from my iPhone using Tapatalk