Colour space

[Harry Harrison]

27 minutes ago, MDM said:

I doubt that any buyer would be too concerned about which one to buy as the differences are hard to see

Hi MDM,

I appreciate that you're very busy but with respect I don't think that is the point. The visible difference on screen between an image with an embedded aRGB colour profile and the same image with an embedded sRGB version in a colour managed system would be very slight, you would see it on your wide gamut monitor but most wouldn't, and remember here we're only talking about the differences between the two images with an embedded colour profile, one sRGB and one aRGB, as they were uploaded.

 

The surprise for me is that the purchased images, although they have no profiles embedded seem to be in the sRGB colour space, in other words the underlying numbers of the colours only make sense in that space. If what you're saying is that no one needs aRGB images these days then that is perhaps the conclusion that we have to come to, because it's looking like you don't get them from Alamy. There's all sorts of ifs & buts that I'm leaving out for clarity but that's the gist of it I think.

[MDM]

17 minutes ago, Harry Harrison said:

Hi MDM,

I appreciate that you're very busy but with respect I don't think that is the point. The visible difference on screen between an image with an embedded aRGB colour profile and the same image with an embedded sRGB version in a colour managed system would be very slight, you would see it on your wide gamut monitor but most wouldn't, and remember here we're only talking about the differences between the two images with an embedded colour profile, one sRGB and one aRGB, as they were uploaded.

 

The surprise for me is that the purchased images, although they have no profiles embedded seem to be in the sRGB colour space, in other words the underlying numbers of the colours only make sense in that space. If what you're saying is that no one needs aRGB images these days then that is perhaps the conclusion that we have to come to, because it's looking like you don't get them from Alamy. There's all sorts of ifs & buts that I'm leaving out for clarity but that's the gist of it I think.

 

The busy bit is about how much time I can spend on this - I find it very interesting.

 

I didn't say nobody needs files in AdobeRGB at all - not sure where you saw that-  but it is probably not necessary for a lot of stock as it is probably irrelevant to the buyer. I supply in AdobeRGB and always have done in the knowledge that I am otherwise throwing away a mass of colour information which could certainly be important for printing (CMYK conversion) but a lot of buyers probably do not care or Alamy would be surely catering for that need.

 

I don't know how you can tell whether Alamy have converted to sRGB before removing the original tag or not. My knowledge is based on what I do in my own colour-managed workflow including printing - I am no expert in the science of colour management but have used it since the early 2000s. 

 

 

[wiskerke]

At the moment I'm traveling and only have a laptop with an uncalibrated TN screen available at the moment. Before I left I did upload something similar but just a general scene, no test patches. However I have included a whacked color profile image which will turn blue and green if handled in a certain way (let's not specify this any further at the moment). Could be that it will not clear QC because of the whacked profile. If it clears and is blue and green I will have some serious keywording to do. 🙂

I'm in Munich btw.

 

wim

[M.Chapman]

2 hours ago, MDM said:

I don't know how you can tell whether Alamy have converted to sRGB before removing the original tag or not. 

 

I think that's quite easy. In simplistic terms, relative to sRGB, to create an aRGB file I believe 2 things have to happen.

 

1) The colour data in the file has to be shifted (compressed) slightly to allow the extra colours (for the wider aRGB gamut) to be squeezed in.

2) The profile has to changed to aRGB. This profile describes how the data has to be processed to decompress the original colour data (to get it back to where it should be) and to define how the extra (wide gamut) colours have been encoded.

 

When I download what should be AdobeRGB image from Alamy it shows neither of these characteristics.

 

1) The image data has not been shifted (it looks like sRGB data)

2) The necessary aRGB profile is missing.

 

If it walks like a duck and quacks like a duck, I think it is a duck (ie. sRGB). Assuming I've not made a howling mistake, I think these images demonstrate  that Alamy is currently supplying sRGB images irrespective of the format uploaded. This is contrary to what one might expect from the Alamy statement quoted earlier.

 

...but we work with Adobe RGB and ignore any embedded ICC profiles when we process your images so before sending we recommend saving all images as Adobe RGB (1998).

 

NB. My test only looks at high res "Personal Use" image downloads, so I can't tell if it applies to other licence types. But, if it does it means 2 things

 

1) Producing aRGB images just for Alamy is a waste of time (unless that's what you produce anyway)

2) If customers believe they are receiving an aRGB file and therefore assign an aRGB profile they will be getting a significantly incorrect rendering of the image. 

 

If there's no further input from the forum I'll contact Alamy Contributor Services with my findings to see what they have to say. It's not impossible a bug has crept in (it's happened before....)

 

Mark

Edited July 9, 2019 by M.Chapman

[MDM]

1 hour ago, M.Chapman said:

 

I think that's quite easy. In simplistic terms, relative to sRGB, to create an aRGB file I believe 2 things have to happen.

 

1) The colour data in the file has to be shifted (compressed) slightly to allow the extra colours (for the wider aRGB gamut) to be squeezed in.

2) The profile has to changed to aRGB. This profile describes how the data has to be processed to decompress the original colour data (to get it back to where it should be) and to define how the extra (wide gamut) colours have been encoded.

 

When I download what should be AdobeRGB image from Alamy it shows neither of these characteristics.

 

1) The image data has not been shifted (it looks like sRGB data)

2) The necessary aRGB profile is missing.

 

If it walks like a duck and quacks like a duck, I think it is a duck (ie. sRGB). Assuming I've not made a howling mistake, I think these images demonstrate  that Alamy is currently supplying sRGB images irrespective of the format uploaded. This is contrary to what one might expect from the Alamy statement quoted earlier.

 

...but we work with Adobe RGB and ignore any embedded ICC profiles when we process your images so before sending we recommend saving all images as Adobe RGB (1998).

 

NB. My test only looks at high res "Personal Use" image downloads, so I can't tell if it applies to other licence types. But, if it does it means 2 things

 

1) Producing aRGB images just for Alamy is a waste of time (unless that's what you produce anyway)

2) If customers believe they are receiving an aRGB file and therefore assign an aRGB profile they will be getting a significantly incorrect rendering of the image. 

 

If there's no further input from the forum I'll contact Alamy Contributor Services with my findings to see what they have to say. It's not impossible a bug has crept in (it's happened before....)

 

Mark

 

 

I don't think you are understanding what I was saying. This is academic but how can you tell the difference between a file that was originally Adobe RGB  (1) had the tag removed and left that way with (2) one that has been first converted to sRGB and then had the the tag removed? Perhaps if you could analyse the file with some software that gives a readout in the form of numbers or a graph but I can't see how you could do this by eye and make any assumptions about its provenance.

 

I can't see how using Adobe RGB as a working space is a waste of time. It takes a keyboard command to convert the profile to sRGB if you want to do that. 

 

 

 

[wiskerke]

OK my images have cleared. For the web version the procedure is clear: first convert to sRGB and then strip the profile.

I have not bought/downloaded my own images. I don't think I will keyword all three, so no live evaluation.

These  are from AIM:

 

whacked

 

sRGB

 

AdobeRGB.

 

wim

Edited July 10, 2019 by wiskerke
link removed

[M.Chapman]

2 hours ago, MDM said:

I don't think you are understanding what I was saying. This is academic but how can you tell the difference between a file that was originally Adobe RGB  (1) had the tag removed and left that way with (2) one that has been first converted to sRGB and then had the the tag removed? 

 

I'm not sure you're aware how much the data in the file is shifted when it's converted to aRGB. It's relatively easy to spot visually. What's difficult to spot is the difference between a correctly rendered sRGB and aRGB images. Try the following.

 

• Create a colourful Adobe RGB jpg (by whatever method you choose) and save it as Source.jpg.

To create Case 1

• Change Photoshop so working space is AdobeRGB (not sure this is needed -  but do it anyway to be sure)
• Re-open the Source.jpg in Photoshop
• Now select Edit>Assign Profile>Don't color manage this document
• Now save as a jpg with a new name (e.g. Case1.jpg). DO NOT tick the embed profile box
• This will produce a copy of the original file with the profile omitted.

To create case 2

• Leave Photoshop working space as AdobeRGB
• Re-open Source.jpg in Photoshop
• Now Edit>Convert to Profile>sRGB
• Now select Edit>Assign Profile>Don't color manage this document
• Now save as a jpg with a new name (e.g. Case2.jpg). DO NOT tick the embed profile box
• This will produce a copy of sRGB version of the original file with the profile omitted.

Now compare Case1.jpg and Case2.jpg using any tool you like (e.g. PS or Mac OS Preview) and toggle between them to see the difference.  NB. If using PS don't reassign a colour profile.

 

Here are the files I just produced using this procedure https://postimg.cc/gallery/du9rktls/ if you want to download them to compare.

 

Alternatively here's a composite jpg I made of the coloured patches only. Case1 is on the left and Case2 is on the right. Note the shifts in red and green patches and others. I must say I find the differences are more noticeable when toggling between the two images (Case1 and Case2) than my composite, but they are significant, especially when looking at a colourful landscape image (for example) by customers who might be discerning enough to actually want an aRGB profile image. If I'd produced the same composite from correctly rendered aRGB and sRGB images the differences would be almost invisible.

 

 

PS. Another way to strip the profile tag is to use EXIFTOOL.exe

 

Mark

 

Edited July 9, 2019 by M.Chapman

[wiskerke]

2 hours ago, M.Chapman said:

 

<>

I must say I find the differences are more noticeable when toggling between the two images (Case1 and Case2) than my composite,

<>

PS. Another way to strip the profile tag is to use EXIFTOOL.exe

 

 

The more traditional way of comparing color using a Macbeth chart is to use smaller boxes inside the reference image like this:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The legacy Photoshop tool save for web also strips the profile, after converting to sRGB. So this may well be the tool that's in use at Alamy.

 

wim

[Marianne]

I wish I had just opened the correct image for my test (the unedited download) and saved myself and everyone else much of this long discussion, but many of the questions asked and observations are very helpful, so not a total waste, I hope.

 

After your observations, there is a part of me that is considering keeping AdobeRGB tiffs (with zip compression) and then sRGB jpegs for upload. It's weird, some of the micros provide AdobeRGB Tiffs (as extended licenses) and one site actually even provides RAW files, so it surprises me that Alamy, with so many publishing clients (who, in my experience, prefer Adobe RGB, often for conversion to CMYK for printing), would reduce the color space. This can wreak havoc on color crucial things such as skin tones (which often have a lot of both red and olive tones); it also is a nightmare for certain illustrations if you are using large blocks of color, especially in the green tones. 

 

I did not have the original of the full resolution image I tested for comparison, so I think that the image chosen here, with the large solid blocks of color where you can actually read the RGB numbers with an eyedropper - not to mention see the wide variation in the examples posted - was an ideal choice. It has left me more confused than ever. The sRGB images seem to have a richer color, and I am hoping Alamy can provide guidance. 

 

Meanwhile, I'm preparing images for a magazine and calendars, all AdobeRGB, and I'll check back here before uploading those that I can to Alamy to decide what to do next.  

[M.Chapman]

12 hours ago, wiskerke said:

 

The more traditional way of comparing color using a Macbeth chart is to use smaller boxes inside the reference image like this...

 

Hi Wim,

 

Just for you (Case 1 in background, Case 2 is centre areas). I also cleaned up the patches to remove the distracting noise. 🙂
 

Case 1 is an Adobe RGB image that has had the tag (profile) removed)

Case 2 is the same Adobe RGB image that was converted to sRGB before the tag (profile) was removed

There's clearly a difference. This is why I believe I can tell that Alamy have converted my AdobeRGB image to sRGB, then stripped the profile and provided for download as a lo-res preview and hi-res personal use licence.

 

 

I'll take a look at your whacked images when they go live.

 

Mark

Edited July 10, 2019 by M.Chapman

[Harry Harrison]

19 hours ago, MDM said:

I didn't say nobody needs files in AdobeRGB at all - not sure where you saw that-  but it is probably not necessary for a lot of stock as it is probably irrelevant to the buyer. I supply in AdobeRGB and always have done in the knowledge that I am otherwise throwing away a mass of colour information which could certainly be important for printing (CMYK conversion) but a lot of buyers probably do not care or Alamy would be surely catering for that need.

 

I just wanted to clarify what I was trying to say here, I was simply trying to extend the logic of what you were saying. It's not that you said that no one needs aRGB files at all but you did correctly suggest that most buyers probably wouldn't care and that sRGB would be fine for them. I'm pretty sure that Mark's Presentation Use image is in the sRGB colour space even though it has no profile so if (Big iF) all types of download turn out to be in the sRGB colour space then I'm wondering how those discerning buyers would get hold of the original aRGB images as they were uploaded.

 

Following on from that it's clear that a lot of images are just uploaded as sRGB, probably always have been but more now that Alamy no longer specify,  so how would a discerning buyer even know that an aRGB file was available?

 

Following on from that (sorry!), if it was discovered that they were supplying images in the aRGB colour space but with no colour profile then that would be very irresponsible imho.

 

I also upload as aRGB for the same reasons that you do, it's not particularly a workflow thing for me as I don't shoot in camera jpegs, I always shoot RAW and use Lightroom to export for Alamy so the colour profile is just a setting in the export panel, I could just as easily export as sRGB.

 

 

Edited July 10, 2019 by Harry Harrison

[wiskerke]

1 hour ago, M.Chapman said:

 

Hi Wim,

 

Just for you (Case 1 in background, Case 2 is centre areas). I also cleaned up the patches to remove the distracting noise. 🙂
 

 

I'll take a look at your whacked images when they go live.

 

Mark

 

Much clearer now! Don't you think?

I find that in general images it's the sky that suffers most from conversion (AdobeRGB > sRGB).

 

To fully appreciate the whacked profile, just download the images from the link. Last letter S=sRGB; A is AdobeRGB and WS is whacked sRGB.

I will not be keywording for at least a week. And with this outcome I will only keep 1 image and discard the other 2. No use to have 3 now almost identical images live.

I am still hoping that the files that go out to clients have the full proper profile. OTOH maybe all the clients that would have wanted AdobeRGB (=better paying discerning print clients) have left the building. Or retired 🚮. Knowledge > 🗑️.

Anyway it looks like I will have to brush up my knowledge of rendering intent. So looking forward to that. NOT.

 

wim

 

 

 

[MDM]

This is a really interesting thread and I would need to take some time to fully appreciate it but I can't do that right now as I am running to several deadlines. My comments may have missed some important facts/info as I have just been dipping into it rather than fully immersing so apologies if I have made inaccurate statements. I have been downloading the files that Mark and wim have made available but have only had a quick look so far. It will be interesting to see what will come out of it and if Alamy have any response. 

 

wim - the second and third jpegs on dropbox are corrupt - could you please re-upload.

 

 

[wiskerke]

20 minutes ago, MDM said:

This is a really interesting thread and I would need to take some time to fully appreciate it but I can't do that right now as I am running to several deadlines. My comments may have missed some important facts/info as I have just been dipping into it rather than fully immersing so apologies if I have made inaccurate statements. I have been downloading the files that Mark and wim have made available but have only had a quick look so far. It will be interesting to see what will come out of it and if Alamy have any response. 

 

wim - the second and third jpegs on dropbox are corrupt - could you please re-upload.

 

 

 

I did re-upload, and downloaded as well because of the message. and they're not corrupt. However there will be no preview available. Probably because of the profiles. If you download them with the download button on the page under the 3 dots in the upper left hand corner, you can play with them. Most useful: opening in a color managed program and a non color managed program side by side.

Further possibilities:  converting/stripping/assigningthe profiles in photoshop or any other editor that let you do this.

 

wim

[MDM]

45 minutes ago, wiskerke said:

 

I did re-upload, and downloaded as well because of the message. and they're not corrupt. However there will be no preview available. Probably because of the profiles. If you download them with the download button on the page under the 3 dots in the upper left hand corner, you can play with them. Most useful: opening in a color managed program and a non color managed program side by side.

Further possibilities:  converting/stripping/assigningthe profiles in photoshop or any other editor that let you do this.

 

wim

 

 

I'm still getting errors on the 2nd and 3rd files. In Dropbox the message is : ".jpg files are supported but something went wrong." and no file preview,

When I download using the buttons (which I always do anyway) and open in Photoshop, I get " This document may be damaged ( file may be tuncated or incomplete) Continue?"

 

If I continue and open in Photoshop the files are incomplete - horizontal areas are missing. Maybe it's a Mac/Windows thing. Unusual though,

 

 

 

[riccarbi]

I remember that some years ago I did many experiments on AdobeRGB vs sRGB. Starting from a 16-bit/channel Tiff photo with a broad range of colors, especially red nuances (which are notoriously difficult to print) I printed it again and again for two full days on my hexachromatic HP Print Jet 130NR changing color space, trying different semi-glossy, matt, and glossy photo papers, and so on. I didn't see any difference between converting my picture to 8-bit/channel aRGB or to sRGB. None, nada, zero. What really mattered was how good I was at calibrating the system for different types of paper.
Since then, despite the theoretical superiority of aRGB as color space; I'm always saving in sRGB.

Edited July 10, 2019 by riccarbi

[spacecadet]

6 hours ago, MDM said:

 

 

 

 

If I continue and open in Photoshop the files are incomplete - horizontal areas are missing. Maybe it's a Mac/Windows thing. Unusual though,

 

 

 

Same here on a   PC.

[M.Chapman]

9 hours ago, MDM said:

 

 

I'm still getting errors on the 2nd and 3rd files. In Dropbox the message is : ".jpg files are supported but something went wrong." and no file preview,

When I download using the buttons (which I always do anyway) and open in Photoshop, I get " This document may be damaged ( file may be tuncated or incomplete) Continue?"

 

If I continue and open in Photoshop the files are incomplete - horizontal areas are missing. Maybe it's a Mac/Windows thing. Unusual though,

 

 

 

 

Same here. I did a direct download of all 3 which arrived in a zip file which I unzipped, but the 2nd and 3rd files show corruption (horizontal strips of missing data).

 

Mark

[M.Chapman]

10 hours ago, wiskerke said:

To fully appreciate the whacked profile, just download the images from the link. Last letter S=sRGB; A is AdobeRGB and WS is whacked sRGB.

I will not be keywording for at least a week. And with this outcome I will only keep 1 image and discard the other 2. No use to have 3 now almost identical images live.

 

It's probably worth keywording all 3 so they can be downloaded as previews to see how they arrive, and maybe even purchase the Adobe RGB version as a hi-res download to see if you agree with my observation that Alamy have converted to sRGB and stripped the profile? I'd value your confirmation. You could leave the WS version on sale as example of the effect of a whacked profile? You never know what might sell !

 

Mark

 

[M.Chapman]

21 hours ago, Marianne said:

The sRGB images seem to have a richer color, and I am hoping Alamy can provide guidance. 

 

Meanwhile, I'm preparing images for a magazine and calendars, all AdobeRGB, and I'll check back here before uploading those that I can to Alamy to decide what to do next.  

 

sRGB only has "richer" colours than aRGB if the aRGB file has been incorrectly rendered. The colour image data in an aRGB file has been compressed slightly to make the colours less rich in order to make room for to fit the extra (out of sRGB gamut) colours in. When an aRGB file is correctly rendered by a system using an aRGB profile the colours are expanded back out again to where they should be, along with displaying the extra out of sRGB gamut colours. But, if an aRGB file is rendered using an sRGB (or default) profile, the compressed colours aren't expanded back out, so the image appears dull. That's my understanding anyway.

 

If you want another perspective (some would say controversial) on the whole aRGB vs sRGB debate then see Ken Rockwell's site https://www.kenrockwell.com/tech/adobe-rgb.htm. It's somewhat out of date now as many systems now handle aRGB correctly so the situation is nowhere near as bad. However, it appears that even Alamy are currently having some difficulty with proper colour management. They really shouldn't be providing files from which the profile has been stripped. The profile tells the client application how the image data inside the file is to be interpreted. 

 

Mark   

[Richard Tadman]

Given that I started this thread and have been incredibly intrigued by  the complexity of analysis and discussion that it has generated, would it be disingenuous to conclude that while there is clearly a 'real' if minimal (to the user) difference  between the colour spaces under discussion, that to all intents and purpose the reality is that this is fairly insignificant to all but the most discerning or pedantic? I don't mean this to sound in any way disparaging, just suggesting that in the real world it may not be of the magnitude we are affording it. There is much fascinating and technical debate which in truth has been very educational but goes largely above my head, although I have tried quite hard to follow it. Many of you are obviously far more proficient in the technicalities than I am. That said, standing back, I conclude that as photographers we could get obsessively engrossed in the minutiae of colour space, but with the best will in the world Alamy has been selling perfectly acceptable images to the world at large for a long time and my aRGB submission have apparently resulted in sales, with no customer detriment that I can detect.
Please don't think I am in any way being critical of the comments from very knowledgeable contributors, it has been a revelation, but with the best will in the world, I am still struggling to get clear in my mind if this is a real issue to customers or an esoteric topic that in the cold light of day is very interesting but of little practical significance.

[Marianne]

44 minutes ago, M.Chapman said:

 

sRGB only has "richer" colours than aRGB if the aRGB file has been incorrectly rendered. The colour image data in an aRGB file has been compressed slightly to make the colours less rich in order to make room for to fit the extra (out of sRGB gamut) colours in. When an aRGB file is correctly rendered by a system using an aRGB profile the colours are expanded back out again to where they should be, along with displaying the extra out of sRGB gamut colours. But, if an aRGB file is rendered using an sRGB (or default) profile, the compressed colours aren't expanded back out, so the image appears dull. That's my understanding anyway.

 

If you want another perspective (some would say controversial) on the whole aRGB vs sRGB debate then see Ken Rockwell's site https://www.kenrockwell.com/tech/adobe-rgb.htm. It's somewhat out of date now as many systems now handle aRGB correctly so the situation is nowhere near as bad. However, it appears that even Alamy are currently having some difficulty with proper colour management. They really shouldn't be providing files from which the profile has been stripped. The profile tells the client application how the image data inside the file is to be interpreted. 

 

Mark   

 

I know that AdobeRGB has a larger & richer color space in actuality, but I was commenting on how, after the AdobeRGB is rendered as sRGB, the original sRGB appears to have richer colors. Because of that, I'm wondering if it would be better to upload sRGB. 

Edited July 10, 2019 by Marianne

[M.Chapman]

3 minutes ago, Marianne said:

 

I know that AdobeRGB has a larger & richer color space in actuality, but I was commenting on how, after the AdobeRGB is rendered as sRGB, the original sRGB appears to have richer colors. 

 

Yes, that's exactly right. 🙂 

Mark

[M.Chapman]

24 minutes ago, Richard Tadman said:

Given that I started this thread and have been incredibly intrigued by  the complexity of analysis and discussion that it has generated, would it be disingenuous to conclude that while there is clearly a 'real' if minimal (to the user) difference  between the colour spaces under discussion, that to all intents and purpose the reality is that this is fairly insignificant to all but the most discerning or pedantic? I don't mean this to sound in any way disparaging, just suggesting that in the real world it may not be of the magnitude we are affording it. There is much fascinating and technical debate which in truth has been very educational but goes largely above my head, although I have tried quite hard to follow it. Many of you are obviously far more proficient in the technicalities than I am. That said, standing back, I conclude that as photographers we could get obsessively engrossed in the minutiae of colour space, but with the best will in the world Alamy has been selling perfectly acceptable images to the world at large for a long time and my aRGB submission have apparently resulted in sales, with no customer detriment that I can detect.
Please don't think I am in any way being critical of the comments from very knowledgeable contributors, it has been a revelation, but with the best will in the world, I am still struggling to get clear in my mind if this is a real issue to customers or an esoteric topic that in the cold light of day is very interesting but of little practical significance.

 

Yes the difference between correctly rendered aRGB and sRGB images is minimal and many customers wouldn't notice the difference.

 

But, the reason that it's getting quite some attention from me (and perhaps others?) is that some of us have spent a great deal of extra time preparing both aRGB images (because Alamy said that's what they wanted) as well as regular sRGB images required for our own use, web-use and when submitting to other libraries. If it turns out that Alamy is now converting aRGB to sRGB and stripping the profiles before supplying to customers it's rather disappointing (an understatement!!), and confirming this could save further work in future (although I've already swapped to an sRGB only workflow).

 

So I intend to inform Contributor Services of what I'm seeing to get their response, but I'm hoping that others could confirm my findings first, in case I've made a mistake or misunderstood something.

 

Mark  

Edited July 10, 2019 by M.Chapman

[MDM]

6 minutes ago, M.Chapman said:

 

 

But, the reason that it's getting quite some attention from me (and perhaps others?) is that some of us have spent a great deal of extra time preparing both aRGB images (because Alamy said that's what they wanted) as well as regular sRGB images required for our own use, web-use and when submitting to other libraries. If it turns out that Alamy is now converting aRGB to sRGB and stripping the profiles before supplying to customers it's rather disappointing (an understatement!!), and confirming this could save further work in future (although I've already swapped to an sRGB only workflow).

 

 

I don't think this is the issue as all it takes to convert to sRGB from AdobeRGB is a Lightroom export or Photoshop action and let it run. The real issue is the conversion of all images to sRGB and the subsequent removal of profiles as you are suggesting.