[BruceBanner]

I would like to share something with you all that happened to me recently. Rather than (overly) repeat myself I shall instead link you to the site where I sought specific help with this issue; Export As Different Colour from Main | Photoshop Gurus Forum

In summary (if we are to trust John Wheeler);

- What I am experiencing is not a bug, but an image that shows noticeable difference in what we lose when moving from an editing process of starting off with a 14bit RAW DNG file, editing in a 16 bit ProPhoto environment (both in PS and LR) and then finally witnessing what it will look like in its final form as a 8bit sRGB Jpg (the differences are fairly pronounced).

- I have been working in a workflow of this manner for years and have never noticed an export jpg on my calibrated 43 inch 4k photographer monitor to showing something that immediately gave me feedback that something was off. I first noticed something was off when Exporting the image in LR as a sRGB Jpg (standard for me) and then minimising LR and using FastStone Image Viewer to look at the export. Even without side by side comparisons like I show in the OP (with the file in PS environment and its Export As preview window) I could tell something was off. First time ever that has happened to me. Why is this? Just a really good flukey example of a sRGB export struggling with the reds/magentas?

- Speaking for myself here, but 99% of what I shoot with a camera becomes a 8bit Jpg. Sites like Flickr, Facebook, Instagram and of course images for clients are always passed over and accepted as Jpgs. I am not questioning the validity of shooting RAW and the additional benefits it brings, but I am questioning at what point during the editing process do I 'swap over' to using 8bit Jpg, because up till now it has been only at the end (export), but look at the 'nasty' surprise I got

- I cannot work on an edit thinking it will look like A when in fact comes out looking like B. So I need a solution, perhaps working in a Soft Proofing mode is it, or perhaps I do some initial editing in a 16bit ProPhoto environment to take advantage of the basic exposure recovery or manipulation, but when it comes to starting to work with artistic editing choices I should toggle over to 8bit sRGB working environments...

It just strikes me as being a little odd, it seems ProPhoto and 16bit is really a temporary thing, its something the photographer sitting in front of their lovely professional calibrated monitor gets to appreciate when running software that can handle those environments (LR/PS etc). But what is the point in any of that if the end article cannot accurately reflect that...

Just thought I'd share some of this with you all, curious to know if also any of you have had similar experiences, or whether my current mindset is flawed etc. Does the changes in my image seen in the photoshopguru post actually suggest more is going wrong than simple ProPhoto>sRGB loss?


TIA

BB

[clackers]

You do all your operations in 16bit ProPhoto, right, getting all the advantages, then it's only the final output down to the JPG, right?

[BruceBanner]

You do all your operations in 16bit ProPhoto, right, getting all the advantages, then it's only the final output down to the JPG, right?

Correct clackers.

It seems in this case I am either;

a) experiencing something greater than a shift from 16bit ProPhoto into sRGB 8bit (a bug or a box ticked somewhere deep in the menus that should not be on etc)
b) found just a flukey image that really highlights the deficiencies when moving between the two Colour Spaces

Either way, this thing has gotten me thinking about my PP workflow.

[pschlute]

Bruce....I cannot see your examples, but my understanding has always been:

edit in the widest possible working space (ProPhoto or Adobe RGB) then downsize for output and convert to profile. If your posting to web then sRGB profile, if printing then whatever the printer requires ( usually sRGB or Adobe RGB) Editing in the wider CS and in 16bit vs 8 bit means less artifacts.

When you convert to a smaller profile (to sRGB from ProPhoto) you may lose colours, yes you need to watch for that.Most of the time you will not notice it, but if you have colours that are not capable of display in sRGB space then yes you will have an issue.

This also assumes you have a wide gamut monitor correctly profiled.

[stevebrot]

Is 16bit ProPhoto worth it if 99% of everything ends up 8bit sRGB Jpg?

Yes...because it gives you some control over how gamut is coerced in the bit and gamut reduction. If you are not using a tool that allows "soft proof", it may be a good idea to do so to allow edits specific to atone for what the bit-reduction and gamut narrowing will do. (Lightroom virtual copies are amazing for this task.)

The working rules are capture deep and broad, PP as deep and broad as possible, and publish to the limitations of the target medium.


Steve

[pschlute]

" In PS I went to the Colour Settings and changed from ProPhoto to sRGB, then hit ok and exited.
In LR I went to the settings where it mentions External Editors and set PS to being sRGB there as well
I then did the LR>Edit In>PS with the file, which then seemed to look relatively no different in PS even though now we're in a different colour space
When bringing the Export As preview window up again (with these different changes) a significant and possibly the exact same tone was seen. "

I would change your working space in PS back to ProPhoto the same as LR.

I generally edit in PS... resize...convert to profile...then save as jpeg. Dont have an issue...but you must watch what happens to colours and watch the histogram when your convert profiles. The out of gamut tools can also help.

[BruceBanner]

Bruce....I cannot see your examples, but my understanding has always been:

edit in the widest possible working space (ProPhoto or Adobe RGB) then downsize for output and convert to profile. If your posting to web then sRGB profile, if printing then whatever the printer requires ( usually sRGB or Adobe RGB) Editing in the wider CS and in 16bit vs 8 bit means less artifacts.

When you convert to a smaller profile (to sRGB from ProPhoto) you may lose colours, yes you need to watch for that.Most of the time you will not notice it, but if you have colours that are not capable of display in sRGB space then yes you will have an issue.

This also assumes you have a wide gamut monitor correctly profiled.

The example is not on this site but in the OP on the Photoshopguru site where I posted help initially. It is a far deeper longer thread, but the OP has the example at least (Export As Different Colour from Main | Photoshop Gurus Forum)

Correct my monitor is relatively modern, calibrated and supporting wide gamut.

My kinda point is that why do we even bother editing in a higher 'space' if the 99% (for all of us really) the final version of all our work is 8bit sRGB Jpgs. Sites like Flickr (which are soon to support 6k images natively) will still only accept Jpgs, and I think they are pretty much the top of the field when it comes to online phoro web display. This forum, FB, Instagram, Flickr, websites, it's always Jpgs. My point was, we don't usually pass over tiffs, pngs or DNGs with LR catalogs to clients, typically its Jpgs they want.

So what is the point in even the manoeuvrability in PP with 16bit in terms of data and colour if both things are lost in the final version... Should we do ourselves a favour and just start editing in 8bit SRGB environments? At least that way WYSIWYG? (playing devils advocate here).

I'm starting to think the 14 bit DNG file is a bit like a pyramid. Initially wide at the base (more information) we can use in PP to recover colours, highlights and all that stuff better, but as the journey proceeds to the pinnacle we lose that recovery, or potential too if its stuff that is hard for the sRGB 8bit to hold onto. Depending on what happens with the beginning edits/tweaks of the file, where it finishes off may not drastically show much loss of colours, or quality etc, in essence had you began the editing process with a 8bit Jpg and never needed those additional bits of information then the same end result would be the same. If that is true, then why do we get so excited about RAW and 14bits of data if it cannot be held onto in its final form. Is it true to say that any 14bit RAW image that ends up as a 8bit sRGB jpg would look exactly the same as a 8bit Jpg counterpart that started off that way?

For example;

I take a landscape shot in RAW+, I have the same image as a 8bit Jpg and a 14bit RAW. I then take the 14bit raw off to LR and edit the file, recover shadows and highlights, make some colour changes, stuff like that. I then SYNC those changes to the Jpg version. At this time, side by side the images look a bit different, the RAW might look better as I'm viewing the images in a 16bit ProPhoto space, but then I export both images as 8bit sRGB Jpgs, now how do they compare? Has a lot of the good stuff that the RAW landscape shot had going for it been lost in translation? How do we know what information from the 14bit file is lost and what is kept.

Yes...because it gives you some control over how gamut is coerced in the bit and gamut reduction. If you are not using a tool that allows "soft proof", it may be a good idea to do so to allow edits specific to atone for what the bit-reduction and gamut narrowing will do. (Lightroom virtual copies are amazing for this task.)

The working rules are capture deep and broad, PP as deep and broad as possible, and publish to the limitations of the target medium.


Steve

Thus far I have only soft proofed in LR when using the Print module, so will have to investigate that from an editing perspective.

My issue is how do I know the publishing limitations? How do I know that one image will rarely change from moving between the two formats, and another will suffer. My head is currently thinking 'better the devil I know' and thus edit in a more confined space because then at least what I am seeing and the changes I am making are going to stay the same when I publish...

[pschlute]

So what is the point in even the manoeuvrability in PP with 16bit in terms of data and colour if both things are lost in the final version

Data and colour is not lost when converting from 16 to 8 bit. It is just rendered differently. Doing all your editing in sRGB may not allow you to utilise the full dynamic range of your cameras sensor. It may also introduce artifacts that would not be present had you edited in ProPhoto.
These are big "mays", and probably only a pro will notice the difference.

Editing in a wide gamut 16bit environment and converting to sRGB is not the same as editing solely in sRGB 8 bit. This is the crucial difference.

---------- Post added 10-14-19 at 02:58 AM ----------

My head is currently thinking 'better the devil I know' and thus edit in a more confined space because then at least what I am seeing and the changes I am making are going to stay the same when I publish...

You can see that when you click "convert to profile"

[bwgv001]

When you see differences, where is that? Monitors are most likely the limiting factor in the first place. So what shift is happening?

[stevebrot]

My issue is how do I know the publishing limitations? How do I know that one image will rarely change from moving between the two formats, and another will suffer. My head is currently thinking 'better the devil I know' and thus edit in a more confined space because then at least what I am seeing and the changes I am making are going to stay the same when I publish...

Is it bothersome to think that the image we see of the Mona Lisa on the Web is less than what we see in person or even that available on a fine art print? Three versions, all of which derived from the master's vision.


From Wikipedia | Mona Lisa


Steve

[Kiwizinho]

I can give a pretty emphatic reason for editing at 16 bits and using a wide gamut colour space. I started off with Photoshop Elements, which only supports editing at 8 bits, and also used Gimp a bit, back when it only supported 8 bits per channel, and found that once I started using multiple layers, and things like gradients, the results were absolutely hideous.
Sure, I might want the finished results as an 8 bit per channel jpg, but to be able to do complex edits without nasty editing artefacts, I need 16 bits per channel. Whether I need a wider colour space than sRGB, I'm not sure, but for the same reasons as needing 16 bits per channel, it's probably a good idea.
To use an analogy, money is only measured to the nearest cent, or $0.01, but plenty of financial calculations use far more than two decimal places, even if the final result is rounded off to just two decimals. For small calculations, those extra decimal places might not make much difference, but for larger calculations, those extra decimal places can actually add up to quite considerable differences.
Basically the same applies to images. If you don't do major edits, then 8 bit sRGB can be fine, but the more complexity in terms of editing steps and layers, the more significant the rounding errors become until they make a noticeable difference to your image.

[PDL]

Just ask yourself, am I only putting out JPEG's or am I doing something else, like printing.

If you are only showing JPEG's then you will learn to live with that format's limitations. Do a little research and see what those limitations are.

If you are printing, then using only JPEG to print, then just take out a gun and shoot your foot completely off. I have heard, and discussed, that you can't print from RAW. Well if that was the case, then give up LR, Capture One, ON1 and other RAW converters and really check to see if you are not saving the RAW as a JPEG and then printing only the JPEG. DUH.

I know I have said it before, here and other places, that digital imaging is all about the data. The more data the better. From my perspective shooting and developing (post processing) in wide gamut color spaces is just the most logical thing to do. Adobe LR uses ProPhoto as it's default color space. If you think it is not logical, then stop using Adobe products and go to an environment that only uses JPEG's. You will regret doing that though.

[BruceBanner]

Okok... stop everyone and just watch this

Youtu.be

Youtu.be

[Class A]

It just strikes me as being a little odd, it seems ProPhoto and 16bit is really a temporary thing, its something the photographer sitting in front of their lovely professional calibrated monitor gets to appreciate when running software that can handle those environments (LR/PS etc).

Note that your monitor most likely only supports 8bit output.

Only advanced monitors support 10bit and you'd have to have a special (expensive) graphics card with the right connection to the monitor for a 10bit setup to work.

This may give you an indication as to why 8bit output isn't as bad as you think.

As others have pointed out, you need higher bit representations to keep the data malleable. If you start with 8bit and edit with 8bit you'll get all sorts of artefact (e.g., banding) once you start pushing the data. With the higher bit representation you get a lot more headroom to play with. Once all hues and tones have been pushed and pulled into the right place, they can be represented using 8bit without losing that much information.

[stevebrot]

As others have pointed out, you need higher bit representations to keep the data malleable. If you start with 8bit and edit with 8bit you'll get all sorts of artefact (e.g., banding) once you start pushing the data. With the higher bit representation you get a lot more headroom to play with. Once all hues and tones have been pushed and pulled into the right place, they can be represented using 8bit without losing that much information.

Mannnn....I wish I had said that! Thanks for the concise and accurate explanation.


Steve