I have been using ProPhotoRGB for a long time, believing like many that if any data falls outside of my working space I will compromise that data. From my limited non-scientific perspective, I have always thought the only issue with a working space was making sure it was large enough to contain all of your original (and perhaps modified) data, and a working space that was quite a bit larger than the image data, while maybe unnecessary, had no issues at all.

In the video, Ray indicates that in some images this may compromise smoother transitions. On the surface this seems logical, but I can't wrap my head around why this would actually occur, or at least be significant enough to be concerned about.

I would appreciate any thoughts on this ... is this a common idea expressed by most color scientists? Are there any tips on the practical application of this ... ie what types of images I may wish to limit to aRGB because of this. Obviously from the comments any that have important transitions ... but to be honest most images have some pretty critical transition areas in them.

I asked Michael this very question in person when he was in Melbourne on his Phase one sponsored visit.

In a nutshell - his answer was always use Pro-Photo as your main color space - then if you need to do a dumbed down version for the web or whatever you can. But start with the biggest possible color space to preserve the data. Even if you cant print all of it today - you may well be able to on the next generation printers.
Makes perfect sense to me.

Have you seen the video? Based on your reply, perhaps you haven't?

Yes, this is a philosophy that many photographers have had for quite some time. I have personally used ProPhoto for these reasons for many years, well before I ever discovered this website and started participating.

Michael stated this very thing to Ray as the introduction to the concept ... and the answer surprised me. That's the reason for the post.

Yes, I have seen and own the video. And I referenced the video in my question to Michael so he new where I was coming from.

Rays answer also surprised me - hence my follow up question to Michael in relation to this.

Bottom line from Michael in his answer:[although I am sure he can reply to this thread to state the same]. Use Pro Photo.

Edit - Forgot to add - when Ray says that he beleives Adobe RGB98 is 'ok' for the vast majority - I beleive he is reffering to the vast majority of point and shoot photographers - not professionals or serious amateurs who take their craft very seriously. Thats my take on it anyway after watching it again.

Hi!
Some people have suggested using the smallest space that does not clip any data. This might depend on what kind of images we are dealing with and where they are destined to. For example a fairly unsaturated caucasian skin might be well within the gamut of sRGB, colormatchRGB or eciRGB. On the other hand brightly lit sunflowers most certainly will need prophotoRGB whether we can print all the yellow and green hues or not (today) if we do not desire to clip anything in the early stages of PP. In most cases though I guess prophoto will cause the minimum of hassle.

Maybe Digitaldog might lend his hand on this topic.

Cheers,
J

I really don't see the issue, either in practical or theoretical terms.

The analogy that I use is one of transferring water between buckets. Your camera produces a lot of water (colour). It fills a very large bucket, lets say 10 gallons.

You now need to transfer it to a holding tank where the contents will be stored and mixed. If that tank only can hold 7 gallons some of it will fall on the floor and be lost forever. Your final destination may only hold 5 gallons, but by putting it in a tank that can hold less than 10 gallons you're losing some of it forever.

What happens in a few years when there are printers that can reproduce wider gamuts? Sorry, you lose. The data is gone, unless you return to the original raw file and redo all of your work.

Just stay in 16 bit mode and there's no harm in working in Prophoto and lots of benefit. (16 bit Prophoto is the native working space of Lightroom and Camera Raw, so there's nothing to be done until and if you export to Photoshop or another application).

Michael

Let's take Michael's analogy one step farther. ProPhotoRGB is a 55 gallon drum and aRGB is a 10 gal bucket (diff added for drama!). Let's say a photo has 7 gallons of "water." What Ray is pointing out is that if you (for some reason) want to work on the fourth gallon from the bottom, that would be easier to do using the 10 gallon bucket than the 55 gallon drum, at least theoretically.

However, like Michael I think there are two over-arching issues: First, I would be hard pressed to find a situation where I would actually see a difference in some adjustment I was making that would "expose" this issue. Second, the disadvantage of not having the ability to use the expanded space with future technology far outweighs anything else.

Dave Chew

Hi Wayne

One point you mention that hasn't really been picked up concerns smooth transitions, e.g. sky gradients. Ideally you want a working space that fully contains all the colours captured in your image so that nothing gets clipped at the edge of the working space gamut. However, if the working space is very large (like prophoto) then the gaps between the discrete colours that you can represent with your finite number of rgb combinations is bigger than for a small smace with the same bit depth used in processing.

Basically, for a given sized space the possible colours are closer together for higher bit depths (as there are more colours that it is possible to represent) and so smooth gradients appear smoother. Hence the recommendation to work in 16 bit (rgb) for large spaces like prophoto. A practical soloution is to work in 16 bit at all times.

A further complication is that most print drivers and (I think but am open to correction) all monitor drivers work in only 8 bits per colour channel. However, at present printers and monitors all have considerably smaller gamuts than prophoto, which reduces the problem of big distances between adjacent colours, although as printer gamuts and some monitor gamuts are now looking similar to Adobe rgb (between Srgb and prophoto in size) it is possible that there will be some visible posterisation even where it would not have been visible if the original file had been able to be displayed/printed accurately.

Not a complete answer, but hopefully I've written clearly enough to help a bit. I usually have more difficulties with out of gamut colours when printing than with smooth gradients, although these do occasionally create problems - probably due to the limited bit depth of the output channel. FWIW, if I use LR to convert RAWs I leave them in prophoto 16bit and if I use DXO I tend to use either prophoto or Jo Holmes (smaller) Dcam3 colour space. From DPP I get Adobe rgb, everything in 16 bit.

Out of interest does anyone know comparable delta E values for adjacent colours in prophoto 16 bit and adobe rgb 8 bit?

Cheers

Mike.

Thanks for the comments. All pretty much echo my viewpoint for many years. Perhaps I'm hearing something different on the video, or am curious about something that is really pretty insignificant. Of course, if so I wonder why Ray makes a point of it. Is he just playing devil's advocate and bringing up something that is theoretical but not very relevant?

Here's what I'm hearing him say ...

If you have an image that will fit, or almost fit, in the aRGB space (and there are a great many that will), using the ProPhoto space will be disadvantageous because by using only a part of the space you reduce quantization points. If I want to measure accurately 13.5 ounces of liquid to put in a "bucket" but I only have a 1 gallon container marked in 1/16 gallon increments, I will have a hard time quantifying that measurement. If on the other hand I choose a 1 pint container, again marked in 1/16th increments, I can measure my 13.5 ounces far more accurately.

By using the smaller color space, again as long as the image fits, I gain more quantization points within the data. This implies the number of "points" in any space is the same, so the larger the space, the greater the difference between points. According to the video, subtle transitions would benefit.

Anyway, thanks to all that gave input. Sounds like I'll tuck this one away as a curiosity that isn't significant. As they say, if it ain't broke, don't fix it.

Actually, the newer Epsons in Lightroom in Leopard can indeed print using a full 16 bits and in Photoshop and Lightroom, Adobe ACE (the Adobe CMM) is doing 16 bit to 8 bits display profile transforms in 20 bit precision so the increased precision of a 16 bit image can still be used on a display.

But this doesn't really have anything to do with PP RGB vs any other color space...

What Ray is talking about is that's it's theoretically more efficient to use the smallest color space that an image needs for both capture and printing...and this is technically correct. However, he doesn't offer any solution to the extreme difficulty in coming up with a workflow that would make this easy (or useful).

You can pop a TIFF image into ColorThink and see the image colors plotted against various color spaces and determine which color space holds all the colors without clipping. Many images will be contained by Adobe RGB, many even by sRGB...but nowhere near enough a high enough percent of images can be contained in those smaller color spaces to avoid clipping if you simply decide to quit using PP RGB and adopt some other color space. Many images will have some color clipping.

Then you can plot an image and compare it to the gamut of your printer...unfortunately, the printer gamuts have grown at a rate that indicates that many colors you can capture in digital cameras now can be printed by current printers (let along future wider gamut printers). If you had chosen A RGB or sRGB for all your image, well, you are throwing potentially useful colors out by working in those smaller spaces...and the problem will only grow over time because newer printers will not have _SMALLER_ gamuts but bigger gamuts.

So, while Ray is technically correct–that one should ideally try to use a color space that efficiently encompasses all the colors and only those colors you NEED–it's simply not practical to try to use a variety of color spaces and keep bouncing between them based upon what will and won't clip.

In actual use and experience, working in 16 bit in ProPhoto RGB is not more prone to banding and inefficient use of the colors that your cameras can capture and your printers can print...and it has the upside of NOT EVER clipping any potential captured or printable colors...either now or in the future.

I use Pro Photo RGB. So does Michael (and so does Thomas Knoll who knows a little something about digital imaging). So do most of my friends who know about digital capture and digital printing. So, using Pro Photo RGB will indeed put you into some good company.

It boils down to what colours do you currently have and what do you want to do with them.

Skin, for example, pretty much fits within sRGB. If you have say 99% of your colours within sRGB and want to play with them and you're not really worried about that 1% that's outside and are happy to bring it into sRGB, then doing so means you have more levels between the extremes of any given colour which means smoother transitions/gradations etc. This can make editing easier/better (subjective obviously).

It doesn't matter if future printers might expand current gamuts, because we already have printers that can exceed Adobe RGB in places and complete encompass and exceed sRGB but if your colours don't exist outside of sRGB then extended printer gamuts are not an issue - you might as well make sure you have the best progression of levels in the colours you have.

If, however, you have say skin tones (sitting virtually entirely within sRGB) but a large amount of say blues and reds sitting outside sRGB and even outside Adobe RGB then you wouldn't want to compress your working space if you wanted to keep those colours relative (or as relative as possible given your output gamut) to the skin tones.

There's nothing inherently wrong in working in a large colour space, but you *might* find an advantage through working in smaller space to give you more steps/levels/tones/gradations/etc during editing.

Yes, 16 bit is cool, but open up your image and go into a levels adjustment. You have 256 (0-255) discreet levels you can adjust and if you go to curves, it's the same. That covers the entire available range so if that range is larger, the steps must be bigger between each other.

Want to use the water analogy? Get a really large bucket so that only 1/32 of an inch of water covers the bottom (1/32" deep) and another that is say 4" deep with the same volume of water. If you splash some of the water from the 1/32" deep bucket with a finger and it goes out of the bucket, you may or may not have water run in to cover the dry spot. The bucket may not be perfectly flat and the water tension might be enough that it just starts to pool, leaving dry spots all over.

To do that with the 4" deep bucket you'd have to throw out a LOT of water - almost all of it until it gets down to being 1/32". Of course that's an extreme example and please don't anyone come and tell me that water doesn't bring in surface tension effects until 1/64" or something, 'cause that's not the point.

The point is, if you spread yourself too thin you may end up with areas with no water - no data.

It's not that every single image should be plotted against all available colour spaces and then grab the one that just fits - but that *might* be appropriate depending on the image and what you intend to do with it.

EDIT: For the record, I use ProPhoto almost exclusively, but I don't process large volumes of images so if I was having problems then I'd consider going to a smaller space.

Schewe posted while I was and did a better job of explaining things anyway :-) My only contention with what Jeff posted is that in the future Pro Photo might *not* be proof against clipping from output spaces since it doesn't encompass all of human standard viewer colours (let alone the possible colours of tetrachromats, but that's a whole other discussion and pretty damn interesting :-). However, he is probably going to be right for a significant part of my life time so close enough!

Actually, it would be interesting to get into a discussion about which color space has the largest total volume of color, but if humans can't SEE a color, is it a color? Technically, no...and in fact, Pro Photo RGB does indeed contain colors outside of human vision (by a little bit) and while some may then point at L*a*b as a potential color space, unfortunately, no...the main reason I dislike Lab is that the color coordinates don't really line up with the color coordinates of RGB color spaces and since sooner or later a Lab image HAS to go into a different color space (there sure as heck ain't _ANY_ cameras or printers that can use Lab) you end up with hue torquing that cause it's own set of problems (think purple blues). That and Adobe's implementation of L*a*b leaves a little something to be desired.

There is no such thing as a "perfect color space" which is why we have so many...but the one color space that seems to cause the _LEAST_ amount of problems for me is ProPhoto RGB in 16 bit.

And the point about skin tones and sRGB isn't actually correct...one of the problems with sRGB and skin colors (at least Caucasian skin color) is that deep dark skin color (pinkish yellow in the shadows) is an area of color where your camera can capture it and your printer may indeed be able to print it but it will be clipped in sRGB. The skin color clipping in sRGB can be a real problem that leads to color blotchiness in the skin shadows...

Well, I did say "virtually", but certainly a fair point to clarify. As we've both already said, it would require checking a rendered image plot against a particular colour space with something like ColorThink to see if there was clipping and whether it would be an issue for a given image. And I use Pro Photo anyway most of the time.

As to Pro Photo having colours outside of normal human vision - absolutely, but then so do some films (well, they are sensitive to some spectrum of electromagnetic radiation that human eyes are not) and that hasn't really ever been a problem. So long as we don't go deliberately moving something into that portion of Pro Photo (or any other colour space) then it should be no more of an issue than it was on film (ie none :-) This doesn't mean that capturing data outside of human vision has no point - being able to see, through translation into a colour we can see, might be quite interesting (and certainly useful for some technical applications).

Some animals already have a larger or different chromacy to humans - we just need to translate it in an way that we can then make sense of (which is not an unusual thing - we do it everyday with radar and x-ray imaging, for example).

The most likely advancement to deal with this would seem to be capturing in 32bit so that the amount of data at any stop would be more than enough to give the illusion of analogue because the steps would be too small to be ever be distinguished. Every move along the way would help, of course.

While I haven't seen the above cited video, Joseph Homes, a fine art photographer with a long history working with refined printing processes, has a number of in-depth articles well worth reading. He's made extensive research & development efforts into working spaces for digital cameras.

It isn't so much about finding a space which best matches the printer rather...

My understanding of Joseph's research, to put it a simplified manner, is to use an appropriately sized colour space based on the particular subject matter contained in a RAW file. Essentially, that each image should have a working space just large enough to hold the source colours of the particular subject without clipping, plus some room for editing. In order to avoid quantization damage at the printing stage.

He also has made development work into using RGB working spaces, Chroma Variants, vs. using RGB saturation tools to control colour. Which helps to avoid clipping and minimizes the shifting of tones.

•Here he discusses RGB working spaces.
http://www.josephholmes.com/propages/AboutRGBSpaces.html

•This is his main profiles page with links to his other articles.
http://www.josephholmes.com/profiles.html

-Anthony

So, you would first convert the image to, say sRGB, and then re-edit it?

Joseph Holmes does raise some interesting points in the referenced articles. While the quantization errors of large spaces are minimized at 48 bits/pixel rather than 24, a large space can introduce unreal colors with injudicious editing, where a real color can be pushed into an imaginary color. These imaginary colors, often involving the blues, won't be displayed on current monitors but can cause very broken output even with good printer profiles. Such imaginary colors are less likely to occur with a smaller space.

While 48 bit/pixel ProPhotoRGB is a very usable workflow option, one should be aware of the above problem. Mr. Holmes' profiles also use a custom 1024 point tone curve rather than gamma encoding, and he claims the former is more perceptually uniform and gives better results with printer profiles.

In his discussion of BetaRGB, Bruce Lindbloom also criticizes the 1.8 gamma used in ProPhotoRGB. According to his calculations, a gamma 2.12 results a smaller RMS ΔE, but I'm not certain that this difference would be significant with 48 bit images. Similarly, the luminosity function of CIE L*a*b (L*) rather than a gamma curve is becoming popular in Europe. Bruce also determined the colors likely to be present in real world images and chose chromaticities somewhat smaller than those of ProPhotoRGB.

Thanks. This helps quite a bit. I can imagine a few possible methods that Adobe might be able to use to overcome the workflow problem, but it appears the benefit doesn't justify the effort.

One possiblel application of this might be an occasional image with some extremely subtle detail and not a lot of saturation? It sounds like by using a smaller working space you may be able to retain some of that detail?

No...the differences (and where Ray is talking) is really about color that is near but not exceeding the gamut of the printer being more efficiently adjusted in a smaller rather than larger color space. What it really boils down to is that the tools we are using (primarily Photoshop) were designed for life in a smaller gamut 8 bit world, so the color adjustments can be pretty heavy handed when used in PP RGB in 16 bit. Some people claim you can't adjust PP RGB accurately enough with say, the Curve adjustment in PS because the 0-255 range is too limiting. A one level change (say from 127 to 128 RGB) will be more accurate and thus smoother in Adobe RGB vs PP RGB.

Of course, there are a ton of way to alter the impact of Adjustment layers...simply lowering the opacity will have the effect of producing smaller moves...and there's also the issue of color samplers and histograms...a lot of it really is whether you are good at doing color adjustments...

But if you are using Camera Raw or Lightroom on raw images, those tools _WERE_ designed to output optimized 16 bit ProPhoto RGB images. And since PP RGB is the internal processing space the only modification upon output to PP RGB from those processors is to adjust the gamma since the chomaticities are already in PP RGB.

Unless Pro Photo RGB has been causing real and repeatable problems for you and you WANT to change color spaces, I would suggest listening to what Ray says, take what's useful to you and move on. There is zero reason to change from PP RGB if it's working for you now.

Thanks Jeff. That's very helpful.