ProPhoto RGB

[Potoroo]

Do the denizens in here use the ProPhoto RGB colour space?

[MATT]

I try to keep everything AdobeRGB all the way through. Then convert to sRGB for Web or as required.

MATT

[aim54x]

SRGB all the way for me....I guess I dont do much editing and dont have calibrated screens (should really get this fixed though) so I have never really bothered closely looking at the advantages and disadvantages of ProPhotoRGB and AdobeRGB

[Potoroo]

...I have never really bothered closely looking at the advantages and disadvantages of ProPhotoRGB and AdobeRGB

That's part of what I am curious about.

I stumbled on ProPhoto RGB when I was toying around with Lightroom (I'm still not convinced about it but that's a whole other thread) and noticed its default export colour space was ProPhoto RGB, which I had not heard of. So I Googled it.

Luminous Landscape has a nice article on ProPhoto RGB. In short, it's much wider gamut than Adobe RGB. Arguments for ProPhoto RGB (eg, The argument for a ProPhoto RGB Workflow) tend to run along similar lines to arguments for Raw itself, that by using it you are preserving information in your images throughout your workflow. Despite the Adobe RGB camera setting, apparently you can import the Raw image and convert its colour space to ProPhoto RGB (16-bit RGB), thus avoiding losing any IQ or dynamic range.

The argument then runs that you only convert to a narrower gamut like Adobe RGB or sRGB at the end of the workflow process when and if you need to. The advent of wide gamut monitors and printers, according to some discussions in the Luminous Landscape forums, increasingly makes that question if rather than when. IOW, if your monitor and printer exceed Adobe RGB why would you want to restrict yourself to it?

On the face of it, this all seems rather sensible, so I was curious how many people in here used it or had even thought about it.

[Murray Foote]

I doubt that any printers exceed aRGB and very few monitors do. I believe the reasons they advocate ProFoto RGB are because (1) though aRGB is likely to be a larger space than for any printer, printers can have small amounts of their gamut outside RGB which can then be encompassed by ProFoto RGB and (2) gamut of monitors and printers increase over time so if you work in ProFoto RGB then in 10 years time your images will still be able to take advantage of the full range of then-current monuitors and printers.

And if you have a workflow based on LightRoom, perhaps you might as well leave images in ProFotoRGB apart from softproofing and conversion to RGB for the web (though the latter will probably change before too long, too).

Regards,
Murray

[surenj]

I use sRGB. I can't be bothered to convert color spaces when posting to web.

Few Questions.

1. Can you actually see a difference on a screen with these profiles.
2. If so, which monitors can do that? I doubt any of the sub $1000 would.
3. Is there any advantage in printing?

If the above answers are 'no', then for most amateurs it may not be useful.

I assume pros already know the answers to these questions and acting accordingly.

[Matt. K]

If the largest colour space that can be captured by my camera is Adobe RGB ...then how can ProPhoto RGB give me even more colours than my camera is able to reproduce? Or am I missing something?

[Potoroo]

I doubt that any printers exceed aRGB

You'd be surprised. Higher end Epsons, for example, have exceeded aRGB for years. To be fair, your average amateur isn't going to run out and buy a 3800 in a hurry, but they certainly exist. I would be surprised if commercial grade printers at photoshops that (serious) amateurs might use did not exceed aRGB also.

very few monitors do.

Not your average $300 6-bit Acer, no. But my sub-$1000 (hi, Surenj!) Dell 2408FWP does 110% aRGB, and higher end monitors like Eizo go further still, so they do exist, they are coming down in price and will become more prevalent in future.

1. Can you actually see a difference on a screen with these profiles.

I can on my monitor. I made a Granger chart as per Reichmann's instructions on the article to which I linked earlier and the differences were easily visible (ProPhoto RGB and sRGB were chalk and cheese).

2. If so, which monitors can do that? I doubt any of the sub $1000 would.

My Dell 2408FWP, which is rated at 110% aRGB, cost me about $830, from memory. Not the cheapest but far from the most expensive too.

3. Is there any advantage in printing?

Not losing colours that are in the original seems like an advantage to me.

I assume pros already know the answers to these questions and acting accordingly.

I wonder, which is partly why I asked. I think the technology is becoming accessible enough to make this worth discussing.

[Potoroo]

If the largest colour space that can be captured by my camera is Adobe RGB ...then how can ProPhoto RGB give me even more colours than my camera is able to reproduce? Or am I missing something?

That's a good question. Apparently, most of today's sensors capture more than aRGB can handle but that data remains in the Raw image until it is processed. If I understand properly, colour spaces don't apply to Raw files until you actually convert them into usable formats like TIFF or JPEG. In ordinary circumstances the colour space tag from the camera that's embedded in the EXIF (aRGB or sRGB) is used by the processing software when you do the Save As, so to speak. However, if you can tell the software to not use the embedded tag but to save in ProPhoto RGB instead then you don't lose any of the colours or dynamic range that aRGB or sRGB can't handle. That would explain why Adobe would use ProPhoto RGB in Lightroom, so when you export the Raw to Photoshop or whatever the ProPhoto RGB space "goes with it".

I'm waiting for one of the resident pros to wade in soon and clarify this.

[Murray Foote]

If the largest colour space that can be captured by my camera is Adobe RGB ...then how can ProPhoto RGB give me even more colours than my camera is able to reproduce? Or am I missing something?

That's the largest colour space you can set your camera to so you get a JGP in aRGB. Thom Hogan advocates setting your camera to sRGB because that's what the LCD shows. The actual RAW image from your camera is in the colour space of your camera which will be a lot closer to profoto RGB than aRGB.

Regards,
Murray

[gstark]

one of the resident pros

That would be Matt K.

[Matt. K]

Thanks Murray
So you're saying that if I shoot in RAW, as I do quite often, then I'm capturing a wider space than Adobe RGB? Are there other references to this apart from Thom Hogan? It's something I need to get my head around.

[Potoroo]

So you're saying that if I shoot in RAW, as I do quite often, then I'm capturing a wider space than Adobe RGB? Are there other references to this apart from Thom Hogan? It's something I need to get my head around.

Reichmann says the same thing in the article to which I linked earlier:

To see this look at Figure 2 above. The image shows the Canon 20D generic camera profile as a ghost, with the Adobe RGB 98 profile within it. What a difference! The camera's colour space is much much bigger than the Adobe space, especially in the deep reds and blues. Only in the yellows is the camera space smaller than Adobe RGB 98.

What does this mean? Simply, that if you are using the Adobe RGB colour space with a Canon 20D, for example, (and this applies to virtually every other DSLR on the market), you are not getting a lot of the deep saturated colours that the camera's sensor is capable of capturing.

Reichmann, BTW, is drawing on colour management expert Bruce Fraser, co-author of Real World Color Management. Google would no doubt throw up many more such references.

[Murray Foote]

Thanks Murray
So you're saying that if I shoot in RAW, as I do quite often, then I'm capturing a wider space than Adobe RGB? Are there other references to this apart from Thom Hogan? It's something I need to get my head around.

(Damn, I accidentally overwrote my response in grabbing links).

Here's a page from Dry Creek Photo showing gamut comparisons between various devices (cameras, printers, scanners) and various colour spaces. Unfortunately, Dry Creek has been out of action for three or four years so the examples are not that up to date. Still, the comparison between the Canon 1dsMk11 and aRGB shows the cameras gamut is hugely greater and in fact very close to Profoto RGB. I presume Nikon D3 and D3x would be even better. However, as you can see from that pages, cameras vary greatly.

Perhaps more useful is the German ICC View site where you can do comparisons with your own profiles. I'm not sure whether it would accept my Lightroom camera profiles but I don't have time at the moment to try and find them.

Regards,
Murray

[Murray Foote]

I doubt that any printers exceed aRGB

You'd be surprised. Higher end Epsons, for example, have exceeded aRGB for years. To be fair, your average amateur isn't going to run out and buy a 3800 in a hurry, but they certainly exist. I would be surprised if commercial grade printers at photoshops that (serious) amateurs might use did not exceed aRGB also.

I don't think that's right. You can see from the Dry Creek comparisons (post above) that the R1800 is in general less than sRGB and I doubt that the 3800 would be hugely different. Easy enough to test, too. just download a profile and compare on the ICC View site.

very few monitors do.

Not your average $300 6-bit Acer, no. But my sub-$1000 (hi, Surenj!) Dell 2408FWP does 110% aRGB, and higher end monitors like Eizo go further still, so they do exist, they are coming down in price and will become more prevalent in future.

I have an NEC2690 Mk1 which is 92% of aRGB I think and two years ago was state of the art and cost me $1,500. Things are obviously improving here. My point remains, though. You don't replace a monitor every day so most people's monitors are not state of the art even if they would be when they bought a new one. Gamut is also not the only indicator of monitor quality and PRAD reviews can be telling here (one way or another).

Regards,
Murray

[Potoroo]

I doubt that any printers exceed aRGB

You'd be surprised. Higher end Epsons, for example, have exceeded aRGB for years. To be fair, your average amateur isn't going to run out and buy a 3800 in a hurry, but they certainly exist. I would be surprised if commercial grade printers at photoshops that (serious) amateurs might use did not exceed aRGB also.

I don't think that's right. You can see from the Dry Creek comparisons (post above) that the R1800 is in general less than sRGB and I doubt that the 3800 would be hugely different. Easy enough to test, too. just download a profile and compare on the ICC View site.

See Epson 3800: Gamut Examples just as an example. The 3800 gets mentioned frequently in the LL forums as exceeding aRGB. The Epson x900 printers exceed ProPhoto RGB in some areas! (I have no idea what they cost but nevertheless they exist).

Things are obviously improving here. My point remains, though. You don't replace a monitor every day so most people's monitors are not state of the art even if they would be when they bought a new one. Gamut is also not the only indicator of monitor quality ...

Regardless of how frequently people replace monitors, wide gamut (beyond aRGB) monitors are becoming more common and cheaper, which adds to the argument that using ProPhoto RGB is becoming more feasible for more people and is therefore something people serious about IQ should be giving due consideration.

[sirhc55]

This site is worth a look.

http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml

And yes, I do use ProPhoto.

[Murray Foote]

I doubt that any printers exceed aRGB

You'd be surprised. Higher end Epsons, for example, have exceeded aRGB for years. To be fair, your average amateur isn't going to run out and buy a 3800 in a hurry, but they certainly exist. I would be surprised if commercial grade printers at photoshops that (serious) amateurs might use did not exceed aRGB also.

I don't think that's right. You can see from the Dry Creek comparisons (post above) that the R1800 is in general less than sRGB and I doubt that the 3800 would be hugely different. Easy enough to test, too. just download a profile and compare on the ICC View site.

See Epson 3800: Gamut Examples just as an example. The 3800 gets mentioned frequently in the LL forums as exceeding aRGB. The Epson x900 printers exceed ProPhoto RGB in some areas! (I have no idea what they cost but nevertheless they exist).

That's a useful page. However, I think we've been using the same words to mean different things. When you said the printers exceed aRGB, I took that to mean they had a larger colour space that encompassed aRGB. It seems to me clear from those graphs that there's a larger area in aRGB that the printer does not touch than the area of the printer's gamur outside aRGB. I'm not at all surprised that some of the printer gamut is outside aRBG and I agree that is an argument for using ProFoto as the colour space.

Of course it's all likely to be something of a moot point if you're not profiling your monitor with a good colorimeter and soft proofing. The gamut of the images you actually take is a factor too. Just because there's gamut available in a particular region may not mean you ever go there. Many images fall entirely within sRGB.

Clearly, though, the gamut of printers has increased and will continue to increase. I think that's the main likely area of improvements. I could end up getting a 3880 early next year. If so, it will be interesting to see to what extent I find an increase in print quality over my current R1800.

Regards,
Murray

[Murray Foote]

This site is worth a look.
http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml
And yes, I do use ProPhoto.

That is a useful and clear article that I don't remember seeing before. Though, I might have just forgotten. It's not out of date but it's interesting that he is using a Sony Artisan Monitor and Bruce Fraser is still alive.

Regards,
Murray

[Potoroo]

This site is worth a look.
http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml
And yes, I do use ProPhoto.

That is a useful and clear article that I don't remember seeing before.

I've linked to it twice in this thread.

[sirhc55]

This site is worth a look.
http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml
And yes, I do use ProPhoto.

That is a useful and clear article that I don't remember seeing before.

I've linked to it twice in this thread.

Whoops! Looks like two of us missed it

But that aside, if you’re working in the graphic design industry then ProPhoto makes sense, hence my use of it.

[Matt. K]

Potoroo
Thanks for starting a very interesting thread! I've learnt something that I needed to know and thought I had a handle on. I don't know if it will change my workflow at this stage but it is something I'll keep an eye on and read up on. One day it's my intention to buy a large format Epson and produce fine-art prints so I better get good at colour profiling and I better get good at understanding the subtleties of colour gamuts, spaces and modes. It's the power of this forum and members such as yourself who take the time to post that help us all keep in touch with the technology.

[gstark]

I would like to take what Matt has said a little further, if I may.

I think colour space is something that not too many of us really understand. As a concept I think I know a little about it, and by "little", I really mean exactly that. I know that I can see different outcomes in my images if and when I use a different colour space, but the whole notion of why colour spaces exist, and how they, from a practical aspect, affect my images, is something that I am not altogether clear about.

And I strongly suspect that I'm not Robinson Crusoe in this matter.

Is there some way that we could prepare a forum document (we already have a number of tutorials and I would see this as being added to those) that can help demistify the concepts and usage of colour spaces?

I would like to have a simple, plain english explanation that even I could understand. If we can bring it down to that level of simplicity, I think we would all be winners.

[Matt. K]

Gary
Here's my somewhat limited understanding in a nut shell. The human eye can discern over 16 million colours....let's call that the 'human' colour space. However, the camera can't capture all of those colours...it captures a reduced 'camera'[/i] colour space. However, the computer monitor can't display all of the 'camera' colour space....nor can the printer print all of the 'camera' colour space. If your camera is set to produce JPEG files then it converts the RAW capture to either [i]Adobe RGB....a colour space that still can't be displayed or printed by most desktop setups...or Adobe sRGB which is a reduced colour space that can be displayed and printed by most desktop setups. You can set your camera to capture Adobe RGB or Adobe sRGB...it's your choice. Some photographers say....keep the work flow simple....shoot in sRGB....look at sRGB on your monitor and send it to your printer which will print all of the colours you can see. That way...no surprises. To many photographers this makes good sense. Other photographers say....shoot in Adobe RGB because you capture more data and then if you have to carry out any post processing you will be left with more data....even if you can't see it. This argument too has some merit. We have now learnt, thanks to Potoroo, that the camera is capable of capturing an even larger colour space when shooting in RAW and this is called, or approximates, Prophoto RGB. Your monitor still can't display all of these colour nor can your printer print them. However...technology is improving and if you spend a lot of money you can now buy a monitor that can display the entire Adobe RGB colour space....and some professional printers can print this expanded colour space. We are talking the top end of town here. So shooting in Prophoto RGB has some merit. So....where does all this leave us?
Here's my take....firstly, I have never met a photographer yet who could approach an exhibition quality digital print and say, "Oh! This was captured in the Adobe RGB colour space!" or "this was captured in Prophoto RGB!". In fact most photographers I know can't even tell if the image was captured and processed as a RAW file or a JPEG! However, if a photographer was shooting for ultimate quality then it makes sense that he captures a RAW file and perhaps works in the largest colour space possible providing he had the tools to utilise that space. The real question is, "Can we quantify the improvement in quality? Is it 10%?...or 3%?....or 14%?" The probable answer is that it is just one more element in the entire picture making process and if you've captured a great image then it counts for little. A bit like the difference between shooting on TRI-X or Ilford FP3. It's the image that counts.

[surenj]

Matt, that is a great explanation!!!!!

Thanks to this discussion I am way ahead on understanding color spaces.

From a amatuer/beginner point of view, I am quite happy to capture, see and occationally print colors in sRGB. I would hate for my images to change as the years go by!! I am more than happy with the screen and print quality that i am getting right now.

I raised a similiar point that Matt raises, is there a SIGNIFICANT difference. Potoroo suggests that there is. This then would warrant investigation as it's not difficult to capture and see in prophoto and output in sRBG... for now.

[Potoroo]

Adobe have a rather good technical paper on colour spaces: http://www.adobe.com/digitalimag/pdfs/phscs2ip_colspace.pdf (468k).

In essence, ask yourself this question: what does 255,0,0 really mean? In one sense it means "the most red possible". The problem is that "the most red possible" means different things on different devices if they cannot display the full gamut of human vision. Because devices such as monitors and printers can only display a subset of the gamut of human vision, different colour spaces (subsets) have been defined that cover specific parts of that gamut in the hope of bringing some regularity to the problem. sRGB and aRGB we know, ProPhoto RGB is newer and less well known, and there are others but we can pretty much ignore them.

So, if sRGB is a fairly small subset of the gamut of human vision then 255,0,0 will mean "the reddest possible in the sRGB colour space", which in terms of human vision is not very red at all.

Since aRGB is a larger subset than sRGB then 255,0,0 will mean "the reddest possible in the aRGB space", which in turn will mean "redder than 255,0,0 in the sRGB colour space" and "redder in the gamut of human vision but still not the reddest possible".

ProPhoto RGB covers a gamut fairly close to that of human vision, so 255,0,0 in that colour space will mean not only "the reddest possible in the ppRGB colour space" but also "a red that is pretty close to being the reddest possible in the gamut of human vision and way redder than 255,0,0 in either the aRGB or sRGB colour spaces".

IOW, 255,0,0 in the sRGB colour space may only correspond to (I'm making up numbers here only to illustrate the point) 198,0,0 in aRGB and 123,0,0 in ppRGB.

Therefore, if today's camera sensors are capturing colours in a "camera gamut" that approximates ppRGB/human vision why would you want to throw half the colours in your images away by restricting yourself to a smaller gamut when you don't have to?

[Murray Foote]

the camera is capable of capturing an even larger colour space when shooting in RAW and this is called, or approximates, Prophoto RGB.

Matt, I don't think this is necessarily true. The Fuji S2, for example, shows in the Dry Creek page as having a gamut less than sRGB. The best cameras have a gamut close to ProPhoto RGB but I expect there is a fair range of gamut in cameras currently on the market. The gamut of ProPhoto RGB will not change but the gamut of cameras might continue to expand, especially if there could come to be software or firmware controls to restrict the gamut to the visible or expand into IR or UV.

The other factor as I understand it is that gamut does not directly correlate to image quality. If this were so, it would be impossible to improve on cameras that already approximate ProPhoto RGB and this is clearly not the case. Dynamic range, for example, can vary considerably between cameras, as shown in this graph by Robert Clark. I expect the same would apply to monitors and printers.

Regards,
Murray

[Murray Foote]

See Epson 3800: Gamut Examples just as an example. The 3800 gets mentioned frequently in the LL forums as exceeding aRGB. The Epson x900 printers exceed ProPhoto RGB in some areas! (I have no idea what they cost but nevertheless they exist).

There is a review here from Luminous Landscape of the Epson 7900 and 9900. He shows a gamut chart comparing 7900 to aRGB. He also cautions "Don't interpret this to mean that the 7900 has a greater overall gamut volume than Adobe RGB". There is no question that the new HDR inkset is a great improvement over the K3 inkset and he also shows a chart that appears to depict the 9900 as exceeding the gamut of the 3800 by a significant amount (ie entirely encompassing it at all points). Therefore the 3800 does not exceed aRGB (though it might at some points).

It is possible that the 7900/9900 might exceed ProPhoto RGB in some places but it is clear that even the 7900/9900 comes nowhere near to filling the ProPhoto RGB colour space and that the development of printers has a long way to go before this can happen.

Regards,
Murray

[Murray Foote]

(Arrgh! Just lost a post!)

From an amatuer/beginner point of view, I am quite happy to capture, see and occasionally print colors in sRGB. I would hate for my images to change as the years go by!! I am more than happy with the screen and print quality that i am getting right now.
I raised a similiar point that Matt raises, is there a SIGNIFICANT difference. Potoroo suggests that there is. This then would warrant investigation as it's not difficult to capture and see in prophoto and output in sRBG... for now.

If you’re shooting and keeping RAW images, then at least you can go back and adjust them if for example you need to output to a wide-gamut printer at some time in the future.

The difference is huge in theory. Whether there is a difference in practice depends on the kind of images you shoot. Many landscapes with muted colours will fall entirely within sRGB so it won’t matter. However, especially if you shoot flowers or live music then it becomes a significant issue because you’re never going to be able to output the full gamut of the subject (well, at least not for the foreseeable future).

However, I think a more significant issue than what colour space you’re using is whether you’re calibrating and profiling your monitor with a good colorimeter. A wide gamut monitor no doubt helps. However, if you use a wide gamut colour space with an uncalibrated monitor you’re probably just increasing the likelihood that the colour you select is not really what you think it is.

Regards,
Murray

[gstark]

technical paper

Please don't take this the wrong way, but the last thing that I'm wanting to see is another technical paper.

Rather, I want to avoid as much of the jargon as possible, and have something that can be understood by ... my mother .... your grandmother ....

Or even by a used car salesman, a politician, or maybe even by a telemarketer.

In all honesty, while one of the purposes of the use of jargon s to simplify and clarify meanings and intents, in actual usage, it's quite the opposite: to close doors and keep outsiders out. I see it used in this manner quite frequently in the IT industry, and in photography as well. As a technical educator, that really pisses me off.

So, I think that the first word - the first piece of jargon - that needs to go is "gamut". I think that it's, in this case, quite an intimidatory piece of jargon in this context.

I think that Matt's explanation comes close to what I see as a starting point. I don't think that he used "gamut" at all in it. Let's keep it that way.

what does 255,0,0 really mean? In one sense it means "the most red possible". The problem is that "the most red possible" means different things on different devices if they cannot display the full gamut of human vision.

I want to pull back from where you're starting from here ... what, exactly, is "red"? What, exactly, is "the most red possible"?

You've correctly pointed out that this might mean different things on different devices, but how does that manifest itself in a practical sense?

Our perception of a colour (red, in this case) may change as we change settings on a monitor. Making a monitor brighter may make the red that we're looking at appear brighter. Even if the value might be 255,0,0 the colour that we're seeing is now different. Is this affected by a colour space, or by different colour space settings?

What about our viewing conditions? Those too can affect how we perceive these colours that are being displayed, and while they might not be changing, the outcomes that we may be producing as a result of what we're seeing may well be different.

And so while I think I understand the concepts that you've stated, I'm not sure that you've attained the clarity that I believe is needed here.

And I think that we also need to keep this at a practical level too. We need to keep in mind that the human eye is an amazing device, that permits us to see a range of colours (much better than "gamut" IMHO) that's in excess of 16 million different colours.

The reality is that we can see an infinite number of colours, but any digital device, using current and currently envisaged technologies, can only see a finite number of colours. The actual number varies from device to device, and is governed by a number of factors inherent in those devices' designs. As well as the number of colours though, the range of those colours - the variations in hues, and way that the primary colours and their associated shades - is also governed by those designs, and as such, will also vary as we traverse a range of devices.

And so, would it be correct to suggest that concept of colour spaces is perhaps a method of mapping the relationship of similar colours across a range of devices, where each of those devices presents a different range and breadth of colours that each of them is capable of supporting?

[Potoroo]

the camera is capable of capturing an even larger colour space when shooting in RAW and this is called, or approximates, Prophoto RGB.

Matt, I don't think this is necessarily true. The Fuji S2, for example, shows in the Dry Creek page as having a gamut less than sRGB. The best cameras have a gamut close to ProPhoto RGB but I expect there is a fair range of gamut in cameras currently on the market.

If an EOS 20D so far exceeds aRGB then it's a fair bet all the later Canons and Nikons do too.

There is a review here from Luminous Landscape of the Epson 7900 and 9900. He shows a gamut chart comparing 7900 to aRGB. He also cautions "Don't interpret this to mean that the 7900 has a greater overall gamut volume than Adobe RGB".

He also says:

... it makes little sense to have a state-of-the-art printer such as the 7900 and feed it anything but the highest quality files. GIGO.

This means, for example, ensuring that files are in the widest possible colour space, ie: ProPhoto RGB. If you've been following the advice of some self-proclaimed gurus who preach that Adobe RGB is good enough, ask them what happens to the colours that can now be reproduced by a wide gamut printer and which can not be contained in a space such as Adobe RGB. The answer is, that they fall on the floor, never to be reproduced.

it is clear that even the 7900/9900 comes nowhere near to filling the ProPhoto RGB colour space and that the development of printers has a long way to go before this can happen.

The fact they exceed aRGB (partially or wholly) is sufficient to warrant using ppRGB if you're going to use printers of that quality. You won't get the full ppRGB gamut but you will get more colours than if you restricted yourself to aRGB.

However, I think a more significant issue than what colour space you’re using is whether you’re calibrating and profiling your monitor with a good colorimeter. A wide gamut monitor no doubt helps. However, if you use a wide gamut colour space with an uncalibrated monitor you’re probably just increasing the likelihood that the colour you select is not really what you think it is.

While colour calibration is necessary (I finally got my Spyder Pro 3 recently), I cannot agree that throwing away colour information by choosing an inappropriately small colour space can ever be insignificant.

[Potoroo]

technical paper

Please don't take this the wrong way, but the last thing that I'm wanting to see is another technical paper.

Don't judge a book by its cover. I think that paper is straight forward and not unduly difficult. Also, the last section contains a useful guide to using Adobe Camera Raw's histogram to help select the most appropriate colour space for a given image, which should please Murray at least.

As to the rest, it bears thinking about but more on that later.

[Murray Foote]

the camera is capable of capturing an even larger colour space when shooting in RAW and this is called, or approximates, Prophoto RGB.

Matt, I don't think this is necessarily true. The Fuji S2, for example, shows in the Dry Creek page as having a gamut less than sRGB. The best cameras have a gamut close to ProPhoto RGB but I expect there is a fair range of gamut in cameras currently on the market.

If an EOS 20D so far exceeds aRGB then it's a fair bet all the later Canons and Nikons do too.

But it doesn't. According to Dry Creek, the 20D is about the same as aRGB, probably a little less and the Nikon D70 is about the same. Therefore there may be many current Nikons and Canons with similar specifications while some Sonys, Pentaxes, Olympuses or digicams might perhaps even be less.

The fact they exceed aRGB (partially or wholly) is sufficient to warrant using ppRGB if you're going to use printers of that quality. You won't get the full ppRGB gamut but you will get more colours than if you restricted yourself to aRGB.

I agree with that. I think it's important though not to give the impression that printers may have the gamut of digital cameras.

While colour calibration is necessary (I finally got my Spyder Pro 3 recently), I cannot agree that throwing away colour information by choosing an inappropriately small colour space can ever be insignificant.

You're not throwing away colour information as long as you retain the RAW file, but you may be right about this one, I’m not sure. If you are not using colour management you might perhaps get a more reliable result that requires less trial-and-error test printing by muting the colours (and probably reducing quality thereby).

On the other hand, I recently printed some images taken in very low blue light that looked OK on the screen but were mainly out of printer gamut and I had to whack them repeatedly with a virtual sledgehammer to soft proof them into some kind of shape. Without a colour managed workflow that could be a daunting task indeed, and probably less likely to succeed the wider the editing workspace.

I think that paper is straight forward and not unduly difficult. Also, the last section contains a useful guide to using Adobe Camera Raw's histogram to help select the most appropriate colour space for a given image, which should please Murray at least.

You got me to go back and think about that again. The paper is probably a summarised extract from Andrew Rodney’s book on colour management. It dates to 2006 and is essentially before Lightroom. Setting the colour space for each image strikes me as a curious idea that might be more trouble than it’s worth. If you’re going to use ProPhoto RGB, then why not use it for all images? Setting an image that is inside sRGB to a wider gamut isn’t going to remap the pixels of the image to a gamut outside sRGB. It implies individual image processing in Camera RAW. I find Lightroom too useful and powerful for me to want to do that and AFAIK you can’t even set a colour space in Lightroom other than their customised version of ProPhotoRGB.

This thread has been useful to make me go back and rethink things I take for granted.

One thing it has made me realise that I hadn’t really thought about before is that there may be a quality advantage in getting some images commercially printed on say a 7900 simply because it will have a wider gamut and can produce more colours.

Regards,
Murray

[Potoroo]

If an EOS 20D so far exceeds aRGB then it's a fair bet all the later Canons and Nikons do too.

But it doesn't. According to Dry Creek, the 20D is about the same as aRGB, probably a little less and the Nikon D70 is about the same.

I haven't seen the Dry Creek profiles because I don't have that VRML plugin or whatever it is. But this is the Canon EOS 20D generic camera profile as a ghost with the Adobe RGB 98 profile within it from Luminous Landscape:

What a difference! The camera's colour space is much much bigger than the Adobe space, especially in the deep reds and blues. Only in the yellows is the camera space smaller than Adobe RGB 98.

The fact they exceed aRGB (partially or wholly) is sufficient to warrant using ppRGB if you're going to use printers of that quality. You won't get the full ppRGB gamut but you will get more colours than if you restricted yourself to aRGB.

I agree with that. I think it's important though not to give the impression that printers may have the gamut of digital cameras.

Fair enough. It does beg the question though, of how you deal with colours in the image in the ppRGB space that the printer can't handle, along the lines of your "virtual sledgehammer".

While colour calibration is necessary (I finally got my Spyder Pro 3 recently), I cannot agree that throwing away colour information by choosing an inappropriately small colour space can ever be insignificant.

You're not throwing away colour information as long as you retain the RAW file, but you may be right about this one, I’m not sure.

Assume your Raw image has colour data that exceeds aRGB. Assume out of habit you convert it into a TIFF with an aRGB color space. Surely you've thrown the "excess" colour information away in your working image. Yes, you can "recover" it by generating a new working image from the Raw - if you think of it - but in the context of a standardised workflow a default use of aRGB may well be effectively losing colour information from lots of pictures.

If you know you're only going to publish online and you convert to sRGB then fair enough, but if you know you're going to print on a high quality printer then aRGB in this case will be inferior to what you could have achieved. OK, at the end of the day it's about horses for courses but in the context of this discussion, which is about the merits or otherwise of ppRGB, a colour space that until very recently virtually no-one in here was taking seriously let alone using, surely we need to nut these things out.

I think that paper is straight forward and not unduly difficult. Also, the last section contains a useful guide to using Adobe Camera Raw's histogram to help select the most appropriate colour space for a given image, which should please Murray at least.

Setting the colour space for each image strikes me as a curious idea that might be more trouble than it’s worth. If you’re going to use ProPhoto RGB, then why not use it for all images? Setting an image that is inside sRGB to a wider gamut isn’t going to remap the pixels of the image to a gamut outside sRGB.

On the face of it I would agree. You might be "wasting" a biggish chunk of the ppRGB colour space but whether that has any detrimental effect anywhere down the track I don't know.

It implies individual image processing in Camera RAW. I find Lightroom too useful and powerful for me to want to do that and AFAIK you can’t even set a colour space in Lightroom other than their customised version of ProPhotoRGB.

If you check Lightroom 2.5 Edit > Preferences > External Editing you can select ppRGB, aRGB or sRGB. It defaults to ppRGB (16-bit), which is what prompted my curiosity in the first place. Within Lightroom itself, according to Michael Clarke over at O'Reilly, the default color space is "basically a ProPhoto RGB color space with an sRGB tone curve, nicknamed Melissa RGB by the folks at Adobe."

This thread has been useful to make me go back and rethink things I take for granted.

One thing it has made me realise that I hadn’t really thought about before is that there may be a quality advantage in getting some images commercially printed on say a 7900 simply because it will have a wider gamut and can produce more colours.

One of the advantages to being a late convert to digital is I'm not habituated to a particular workflow. No doubt in a few years I'll be as settled in my ways as anyone else.

[ATJ]

Maybe I don't understand this correctly... but aren't colour spaces somewhat of a moot point if I always shoot raw and I'm only preparing images for displaying on the web? And even if I'm printing, the colour space only comes into it when exporting the file (TIFF, JPG, etc.) for printing?

It was my understanding that colour spaces mean nothing in the raw image and only become relevant when converting to other file formats, which would included displaying on the screen. I use Lightroom which uses Adobe Camera Raw and ProPhoto RGB by default so I don't actually think I have any say it it. When I export the images for web, I always choose sRGB because I have to assume people aren't using a colour managed browser. If everyone was using a colour managed browser there might be some value inn using a different colour space.

I also thought there was a trade off with these colour spaces, especially if you are working in 8 bits. There is a limit to the total number of colours and the different colour spaces just choose a different subset of colours. While one may give a wider range (i.e. biggest difference from the "brightest" to the "dullest" of a colour) it would lose in the sublties - or do I have that wrong?

[Murray Foote]

Hi Andrew

What you say is essentially correct.

Your camera has an implicit colour space and when you work on an image in a RAW processor or an image converter then you will be using another colour space to do so. In your case or mine, using LightRoom means editing in ProPhoto RGB which is probably close to the colour space of the camera.

However, your monitor will be somewhere between sRGB and aRGB which means that the image may contain many colours that the monitor cannot accurately display. As you say, this only becomes an issue when generating output for print or web.

Assuming your monitor is profiled with a good colorimeter/software package, when you come to print you can soft proof to determine what colours are out of gamut and usually get pretty close to what the final print will look like. Well, at least that's true if you have a recent monitor with an aRGB gamut because even the best printers are not far from aRGB. If you have a monitor with sRGB gamut (new or old) and the printer is closer to sRGB than aRGB then that gets a bit more complex. You will be able to see what colours are out of gamut and bring them back but many aRGB colours may not display well on your screen so you may not get an accurate softproof. This could mean much more time and expense in test prints.

With preparing an image for web display it's a bit more nebulous. A few years ago all monitors were sRGB so you prepared for sRGB although many monitors are unprofiled and different browsers can show colours differently so it was still anyone's guess how your images would appear on other monitors. Nowadays there is an increasing minority of aRGB monitors and an image in aRGB viewed on a aRGB monitor can look much more stunning than an image in sRGB on a sRGB monitor. There is no standard but since most people have sRGB monitors, preparing web images for sRGB is still the usual practice.

In theory one should soft proof for sRGB for web images because there can be a considerable difference from the image in ProPhoto RGB. Personally, I don’t usually bother and just convert them because how they will appear on unprofiled screens is a mystery – but this is probably not good practice, especially if your image may have a large gamut. There is another complication here too, if you have an aRGB monitor. Mine will simulate sRGB (as a setting) but you can’t profile for sRGB and switch profiles. I think many aRGB monitors will not even simulate sRGB. I do have a second monitor which is sRGB but I don’t bother profiling it very often so that the colours are usually noticeably out – but that is a possibly desirable setup – two monitors, one aRGB and the other sRGB (both profiled).

What you say about processing in 8 bit is also a good point and is to some extent an argument for processing in 16 bit. However, some Photoshop filters only operate in 8 bit so some images you may need to convert to 8 bit so you can use the filter. In that case, it’s probably a good idea to try varying the colour space first and see whether you can use a smaller one without clipping the histogram – especially if intended output is for print.

Regards,
Murray

[ATJ]

What you say is essentially correct.

Thanks, Murray. I guess I do understand it then.

Assuming your monitor is profiled with a good colorimeter/software package

Mine is. Eye One Display (or I1 Display or whatever combination they use now).

With preparing an image for web display it's a bit more nebulous. A few years ago all monitors were sRGB so you prepared for sRGB although many monitors are unprofiled and different browsers can show colours differently so it was still anyone's guess how your images would appear on other monitors. Nowadays there is an increasing minority of aRGB monitors and an image in aRGB viewed on a aRGB monitor can look much more stunning than an image in sRGB on a sRGB monitor. There is no standard but since most people have sRGB monitors, preparing web images for sRGB is still the usual practice.

In theory one should soft proof for sRGB for web images because there can be a considerable difference from the image in ProPhoto RGB. Personally, I don’t usually bother and just convert them because how they will appear on unprofiled screens is a mystery – but this is probably not good practice, especially if your image may have a large gamut. There is another complication here too, if you have an aRGB monitor. Mine will simulate sRGB (as a setting) but you can’t profile for sRGB and switch profiles. I think many aRGB monitors will not even simulate sRGB. I do have a second monitor which is sRGB but I don’t bother profiling it very often so that the colours are usually noticeably out – but that is a possibly desirable setup – two monitors, one aRGB and the other sRGB (both profiled).

The other thing to consider is the browsers themselves.

As far as I am aware, Internet Explorer is not colour managed at all - but perhaps the latest version is. So, even if the monitor is profiled, the images may not look "correct".

Firefox 3 and later supports colour management, but it wasn't turned on by default in 3.0 (not sure about 3.5) so as above, if colour management isn't turned on, who knows what goes on?

Safari is supposedly colour managed, but from what I have read it does a poor job under certain circumstances - which I can't remember at the moment (I think it is where the colour profile is not included with the image).

I know I have compared my images (created with sRGB) with IE 6.0, Firefox 3 (with colour management turned on) and Safari. Firefox and Safari were close, and close to what the image looks like in Lightroom, but IE 6 is mostly way off.

[Potoroo]

I also thought there was a trade off with these colour spaces, especially if you are working in 8 bits. There is a limit to the total number of colours and the different colour spaces just choose a different subset of colours. While one may give a wider range (i.e. biggest difference from the "brightest" to the "dullest" of a colour) it would lose in the sublties - or do I have that wrong?

That is correct. If the gamut in your Raw image was sufficiently wide and you assigned ppRGB 8-bit then that could lead to banding, which is why Lightroom's default export settings are ppRGB 16-bit.

On a different point completely, I wonder why Lightroom also defaults to 240ppi. It seems like an odd choice given 300ppi is the traditional default for print quality.

[ATJ]

On a different point completely, I wonder why Lightroom also defaults to 240ppi. It seems like an odd choice given 300ppi is the traditional default for print quality.

Probably because it will make no difference at all. The "resolution" number stuffed into a file is just that, a number. It has no influence on anything other that the few dumb programs that refer to it. Most printing software let you choose how the image will be printed my means other than the "resolution".

To prove the point, create 3 different JPEG files with different resolutions, one with one with 240, one with 300 and one with 900 dpi. The resulting files will be identical other than 1 or two bytes (which is the resoltion).

[Murray Foote]

What the theoretical resolution is depends on the printer. For Canon and HP it is 300, for it 360 - or multiples, for example such as 720 or 1440. There has always been some unresolved discussion about this. Fraser and Schewe a few years ago were saying that what you might gain by resizing to an "even" resolution would be outweighed by what you would lose by resampling. Qimage, conversely, advocated the advantages of printing at higher resolution such as 720dpi. I'm not sure whether there might be a difference there between resizing for print in QImage and resampling in Photoshop. More recently, Jeff Schewe has been saying that the way LightRoom works, there is actually some advantage in resizing to say 360dpi or 720dpi provided you don't push it too far.

The other general rule of thumb is that you probably won't notice the difference unless you go below 180dpi. Of course, it all depends on the viewing distance of the prints.

Regards,
Murray

[Matt. K]

So what's the dynamic range of ink jet printing paper and how does that relate to the colour space/mode/gamut that the papers are able to exhibit?

[Murray Foote]

Dynamic range or at least DMax is often cited in paper reviews (brief summary). I think the dynamic range is just part of the gamut, a pole of the volumes in 3D-diagrams. It's just that we are more used to seeing wafers that show the colour range. In the diagram at the top of this page, the dynamic range appears to be a vertical pole in the middle, in this case showing premium lustre on the Epson 2400 to be considerably smaller in gamut than aRGB as well as somewhat smaller in DMax.

Regards,
Murray

[gstark]

So what's the dynamic range of ink jet printing paper and how does that relate to the colour space/mode/gamut that the papers are able to exhibit?

Matt,

I need to remind you that you should always open your inkjet papers in a quiet, darkened room. If you don't, you may run the risk of fogging the paper.

[Murray Foote]

... and the room has to be quiet so you don't get noise in the prints?

[Murray Foote]

(This goes back 11 posts in this thread)

I haven't seen the Dry Creek profiles because I don't have that VRML plugin or whatever it is. But this is the Canon EOS 20D generic camera profile as a ghost with the Adobe RGB 98 profile within it from Luminous Landscape.

It's curious that the Dry Creek view is so different from Michael Reichmann's one using the Apple ColorSynch utility.

I had a look at the German ICCView Utility I linked to earlier but it only accepts ICC and ICM profiles so I couldn't use the LightRoom Camera Profiles. However I did make comparisons between aRGB and my monitor and between my monitor and a custom profile I made for Ilford Gold Fibre Silk on my R1800. ICCView generates 3D diagrams that you can rotate and that really brings home how misleading a 2D representation of that might be.

My monitor closely follows the boundaries of aRGB, more inside than outside and outside mostly for the reds. The comparison of the paper profile against the monitor profile was interesting though, because they were quite different shapes. The paper profile was in general much smaller than the monitor profile but considerably larger in some areas of yellows, greens and cyans.

This is not so much an issue for the colours my monitor shows that the printer cannot depict. That's just a fact of life that I pick up in soft proofing. The colours the printer can print but the monitor can't display are more of a challenge. Perhaps I should look for those colours when soft proofing and assess whether the image might improve if any of them were significantly brighter.

If the answer were yes, it might be useful to use hue/saturation to increase saturation and perhaps to change lightness for the specific colours, identifying them with the sliders at the bottom of the dialogue box. ... then to make some test prints to explore the differences.

The fact they exceed aRGB (partially or wholly) is sufficient to warrant using ppRGB if you're going to use printers of that quality. You won't get the full ppRGB gamut but you will get more colours than if you restricted yourself to aRGB.

I agree with that. I think it's important though not to give the impression that printers may have the gamut of digital cameras.

Fair enough. It does beg the question though, of how you deal with colours in the image in the ppRGB space that the printer can't handle, along the lines of your "virtual sledgehammer".

I had a look at one of the images. Usually my adjustments for softproofing are quite subtle, usually a slight increase of contrast, sometimes slight tweaks with hue-saturation. This time there were pretty radical changes with both curves and levels, a slight adjustment in hue-saturation and use of a black-and-white adjustment layer at less than 100% opacity.

One of the advantages to being a late convert to digital is I'm not habituated to a particular workflow. No doubt in a few years I'll be as settled in my ways as anyone else.

I guess I’m a late convert to digital though in my case I got bored with the fume room and largely dropped out for some years. I started using Photoshop (initially Elements) in about 2004 and got my first DSLR last year. For me it is a two-stage process. First you learn how to do everything properly however long it takes (not that there’s ever an end to that). Then you learn how to do things efficiently (there’s probably never an end to that either). Then when many things become more automatic, it’s easier to focus experimentation.

Regards,
Murray

[ATJ]

Here's another good explanation of the use of colour spaces: Digital-Image Color Spaces, particularly with reference to browsers.

A little over a year ago I experimented with one of my photos where I created JPEGs using different colour spaces. I'm pretty sure I actually embedded the profile into the image so it was available for the browser to reference (if the browser could reference it).

These are all images exported from Lightroom using the different colour profiles/spaces.

sRGB


ProPhoto RGB


Adobe RGB (1998)


sRGB - I can't remember what was different about this one and the first one.


With Firefox (with colour management turned on), The ProPhoto RGB looks closest to the image in Lightroom but the Adobe RGB is not far behind.

If you look at them with Internet Explorer, the ProPhoto RGB one looks the worst (as IE isn't colour managed and uses the default sRGB mapping).

They appear slightly different again with Safari (which is mostly colour managed).

[Murray Foote]

Andrew

That's a useful link and telling images. I've read about this several times before and ignored it all - but you've convinced me to fire up Firefox (and if I remember correctly, in 3.5 Colour Management is turned on by default).

I found another useful link, much less comprehensive than the one you cite above but also much shorter. It helped me to finally understand what it means that IE is not colour-managed.

Essentially, IE delivers images in sRGB, irrespective of their profile. At first glance you might think that should be OK for images saved to sRGB, even if it doesn't work well with other profiles. As we know, unprofiled monitors vary wildly in their depiction of colours so images on them will also vary widly, whether the browser is colour managed or not. The difference here with IE is that the browser is not adjusting the image to the profile of your calibrated and profiled monitor, so that even if the image is sRGB it's exactly the same as if you were viewing it on an unprofiled monitor.

Regards,
Murray

[Potoroo]

Here's another good explanation of the use of colour spaces: Digital-Image Color Spaces, particularly with reference to browsers.

There's a lot of useful information there. It's scary how many of the "pro" image web sites get colour management wrong though.

[Murray Foote]

Colour Management in Mozilla Firefox

[BullcreekBob]

Maybe I don't understand this correctly... but aren't colour spaces somewhat of a moot point if I always shoot raw and I'm only preparing images for displaying on the web? And even if I'm printing, the colour space only comes into it when exporting the file (TIFF, JPG, etc.) for printing?

I'd like to offer an alternate opinion on this. Colour spaces ARE important if you shoot and edit raw files. When doing editing, I've found that doing the editing in a wider colour space DOES make a noticeable difference in the final result even though that invariably is seen in a narrower gamut like sRGB.

For the first image, I told CaptureNX2 to use Adobe RGB as a working colour space, then
- opened the raw file,
- applied my "standard" default processing step as per Jason Odell's book (Set the picture control to standard, set the sharpening to 0, applied a USM step, applied a minor default curve adjustment)
- cropped to 4:3 ratio
- resized to 800*600 pixels
- changed colour space to sRGB
- saved as a jpg.



For the second image, I told CaptureNX2 to use Prophoto RGB as a colour space then did EXACTLY as above with the result


I prefer the colours, especially the reds in the second image. Since doing this "experiment" a while ago, I changed to Prophoto RGB for all editing and I do believe I can tell a difference even though I only ever view the final result in a sRGB colour space. I repeated this test with quite a few other images, portraits, wildlife and landscapes and I was equally pleased with the result.