Jump to content
You must now use your email address to sign in [click for more info] ×

Soft Proofing does not work in the manner it would be helpful (Affinity Photo)


Recommended Posts

I don't know if it's a question or feedback. For the past 5+ hours I've been trying to make sense of how to check how my work will look like when printed, and here's what I found.

What I need

I need a quick way to check whether my colors are within the printable gamut range and also to see the out-of-gamut colors. They keyword here is quick.

What I get

The Soft Proof layer works not like that. Suppose I am in the sRGB color space. It seems like it takes the RGB coordinates of current image of the sRGB space and simply applies them to the Proof Profile that I select (without conversion). But that's not what I want, this naturally gives the wrong results. What happens is that once I apply soft proof with the "U.S. Web Coated (SWOP) v2" profile I suddenly get a brighter image, as if it tells me that this is how it will be printed. And if I click the "Gamut Check" button most of image becomes grey. But it's not the case, it's not how it will be printed. In fact, all the original colors will be printed (in my particular case). The step that this Soft Proof feature misses is the conversion of Color Spaces for this "check".

If instead of applying the Soft Proof layer I convert the color space (Document > Convert ICC Profile...) and select CMYK/8 + "U.S. Web Coated (SWOP) v2" then the image will remain exactly as is, it will not become brighter. Furthermore, if I now apply the Soft Proof layer and select "U.S. Web Coated (SWOP) v2" then the image will stay exactly the same, and if I click "Check Gamut" then no area will be grey. Because all colors are within the color space range.

Bottom line

It's very misleading what the Soft Proof shows if you don't convert your profile first. But if you have to convert your image profile to check the gamut then it makes the "Soft Proof" feature not helpful.

I am attaching an image, just for reference, which I used to play around with this.

Can you please help sort this out, why does it work like that and has that been done intentionally.

soft-proofing.png

rgb-cmyk-soft-proof-comparison.png

Edited by JOleg
appended comparison image
Link to comment
Share on other sites

I'd appreciate anyone commenting on this matter. It may be that I am missing something, or it's simply how it is by design, but if that's the case, then I would like to know the specific scenario when such behavior would be useful to me.

Again the idea that I am after is that no matter what color space I am currently in, I want to test it against any target color space of my choice provided that it shows me the remapped RGB values to the target color space. The "do not remap" RGB coordinates can be a separate check box that I can toggle on/off.

Link to comment
Share on other sites

Not an expert in the matter of CMYK, but I might provide some hints. Those who know better please correct where I’m wrong.

  1. The grey overlay (gamut check) shows what colors (RGB) cannot be perfectly mapped to CMYK. This is the main use case for the adjustment. It shows only that colors will differ, but not how much.
  2. In principle, with gamut check off, it should show the actual effect of the selected color profile conversion. This part seems to not work correctly, as you will get different results when actually converting the document. Could be a bug or by design, the help file doesn’t give information to decide about that.
  3. As we have other open questions about cmyk / rgb conversion of colors (color panel), I tend to believe  no. 2 is a bug.

the Tutorial might help:

https://affinity.serif.com/en-gb/tutorials/photo/desktop/video/316753313/

Edited by NotMyFault
Added tutorial link

Mac mini M1 A2348 | Windows 10 - AMD Ryzen 9 5900x - 32 GB RAM - Nvidia GTX 1080

LG34WK950U-W, calibrated to DCI-P3 with LG Calibration Studio / Spider 5

iPad Air Gen 5 (2022) A2589

Special interest into procedural texture filter, edit alpha channel, RGB/16 and RGB/32 color formats, stacking, finding root causes for misbehaving files, finding creative solutions for unsolvable tasks, finding bugs in Apps.

 

Link to comment
Share on other sites

This topic is not easy. From my experience, this is how I understand it:

You have switched the image both from RGB to CMYK and also from the space "sRGB" to  "U.S. Web Coated (SWOP) v2".

Going from one to another can be done in different way which will result in different looks. 

– You can "convert" a.k.a. change your original image data so that it results in the approx. same look in the new color space, or  
– you can "apply" a color space and leave the original image data the same. But the original image data (R value, G value, B value) will look differently in different color spaces.

(For example: How to deal with very saturated blue gradients from sRGB to SWOP v2? Shrink them down relative to each other or leave them all at the SWOP maximum saturation? There are good videos on youtube around that topic.)

If you send in your image data in Red Green Blue information, the printer will have deal with it somehow because his press will print in Cyan Magenta Yellow Black. So I would suggest to convert it your self because you have more control over the result.

spacer.png

Link to comment
Share on other sites

On 10/8/2021 at 6:22 PM, JOleg said:

Can you please help sort this out, why does it work like that and has that been done intentionally.

I get the feeling that there is some kind of paper simulation going on. When you soft proof with the same profile as the document it's seemingly doing nothing - even changing the rendering intent has no effect - and so no paper simulation also.

Link to comment
Share on other sites

In general, what you are describing sounds like soft-proofing is working as intended.

Soft-proofing is intended to use your RGB display to simulate an output device that is characterized by its ICC profile (eg, a printer) - simulate is the key descriptor here.  When your sRGB file is soft-proofed with the CMYK profile, the visual appearance of the displayed image changes according to the device profile you have chosen and the rendering intent that you have specified; however, your display is trying to give you a general idea of what a printed image is going to look like - clearly a transmissive display and a reflective paper print are not visually going to be identical, so soft-proofing is a visual approximation (an accurate soft proof requires a calibrated and profiled display that can simulate the output device you are specifying).  Also, the gamut check (the gray overlay) indicates to you which colors in the source color space (here, sRGB) are outside of the destination color space (here, CMYK).  That does not mean that those colors will not print on the destination device, it just means that the colors in the source space will have to be mapped to the gamut of the destination space to print and may not appear identical to the image in the source color space.  The way these transformations are handled is partially a function of the **rendering intent** that you select.

In terms of a shift in brightness or perceived appearance, this may be a function of the rendering intent you select and how the white/black points of the source and destination spaces might change - you need to investigate this by examining how the choice of rendering intent affects the visual simulation of your image on your display under soft-proofing, compared with viewing that image as an actual print under controlled lighting conditions.  This way you can choose which soft-proof rendering intent best simulates the appearance of the final print under the specified viewing conditions, and use that intent for soft-proofing.

Kirk

 

Link to comment
Share on other sites

Here is a visualization of your above image with sRGB as the source and SWOP2006 CMYK as the destination, with the ∆E values plotted in a false color image and some of those RGB patches sampled for their ∆E (in the list).  Changing the rendering intent will change the ∆E across the image, so you need to investigate how rendering intent affects your final printed output compared to the soft-proof.  As you can see, the deep purples in the image are particularly out of gamut and may present a problem with printing, especially with a relative colorimetric rendering intent (where OOG gamut colors get clipped to the destination gamut boundary) - you may need to adjust those colors locally or come up with a way to preserve them as well as possible prior to printing if they are important to the final image.  You can do this local adjustment with soft-proofing turned ON  so that you can see the simulated result while you make adjustments.

A tool like this is more helpful than just a gray overlay because it gives you an idea of how far OOG a color is in the destination space so that you can devise a specific strategy to target the worst offenders and leave the less intrusive problems alone, within acceptable error.

kirk

compare.jpg

Link to comment
Share on other sites

Here is a screenshot of a raw converter (Raw Photo Processor, for Mac) that includes a gamut diagram for inspection during raw conversion, so you can adjust your raw conversion while you take into account the ultimate destination color space if you choose.  Not as detailed as the Color Think application, but useful still.

kirk

RPP.jpg

Link to comment
Share on other sites

7 hours ago, kirkt said:

Here is a visualization of your above image with sRGB as the source and SWOP2006 CMYK as the destination, with the ∆E values plotted in a false color image and some of those RGB patches sampled for their ∆E (in the list).  Changing the rendering intent will change the ∆E across the image, so you need to investigate how rendering intent affects your final printed output compared to the soft-proof.  As you can see, the deep purples in the image are particularly out of gamut and may present a problem with printing, especially with a relative colorimetric rendering intent (where OOG gamut colors get clipped to the destination gamut boundary) - you may need to adjust those colors locally or come up with a way to preserve them as well as possible prior to printing if they are important to the final image.  You can do this local adjustment with soft-proofing turned ON  so that you can see the simulated result while you make adjustments.

A tool like this is more helpful than just a gray overlay because it gives you an idea of how far OOG a color is in the destination space so that you can devise a specific strategy to target the worst offenders and leave the less intrusive problems alone, within acceptable error.

kirk

compare.jpg

It would be totally easy to provide this false color overlay if Affinity would provide an RGB2LAB conversion function for procedural texture filter.
Or even better, a deltaE function

Or even better, a delta E false color overlay for soft proof. The math is totally simple,  could be done in PT filter.

http://colormine.org/delta-e-calculator

 

A non-live workflow:

  • Add soft-proof (check gamut off)
  • merge visible
  • delete soft proof
  • convert to lab color mode
  • move original layer on top
  • set blend mode of upper layer to Substrat (difference not working in LAB)
  • add black & white adjustment 
  • add gradient map to create false color 

 

delta E mockup.afphoto

Mac mini M1 A2348 | Windows 10 - AMD Ryzen 9 5900x - 32 GB RAM - Nvidia GTX 1080

LG34WK950U-W, calibrated to DCI-P3 with LG Calibration Studio / Spider 5

iPad Air Gen 5 (2022) A2589

Special interest into procedural texture filter, edit alpha channel, RGB/16 and RGB/32 color formats, stacking, finding root causes for misbehaving files, finding creative solutions for unsolvable tasks, finding bugs in Apps.

 

Link to comment
Share on other sites

The main issue I have with soft-proof:

as @JOleg has shown, and I reproduce the same: soft-proof result differ severely from actual CMYK conversion results.

If you take Screenshots of both (who are RGB), and compare these screenshots, the soft-Proof is always too bright in shadows.

And this carefully uses identical settings (rendering intend, black point, profile).

Mac mini M1 A2348 | Windows 10 - AMD Ryzen 9 5900x - 32 GB RAM - Nvidia GTX 1080

LG34WK950U-W, calibrated to DCI-P3 with LG Calibration Studio / Spider 5

iPad Air Gen 5 (2022) A2589

Special interest into procedural texture filter, edit alpha channel, RGB/16 and RGB/32 color formats, stacking, finding root causes for misbehaving files, finding creative solutions for unsolvable tasks, finding bugs in Apps.

 

Link to comment
Share on other sites

45 minutes ago, NotMyFault said:

soft-proof result differ severely from actual CMYK conversion results.

Isn't that an apples-to-oranges comparison though? Really you need to take your printed material and compare it under correct lighting to an accurate monitor to see how accurate the soft proof is.

Link to comment
Share on other sites


Thank you all for the responses, I appreciate your input and time.
 
@kirkt, I understand what you are saying about the simulation, rendering intent, display calibration and so on. But, I would not be so confused if converting ICC profile from sRGB to CMYK + applying Soft Proof layer + same rendering intent would give me the same "bright up effect" (i.e. the simulation of colors on destination paper) that I see when I apply Soft Proof to sRGB. But this does not happen, with the same destination profile I get different results depending on which ICC profile is current for the image.
So I just don't understand how to make sense of Soft Proof in this case to help my workflow, and I don't know which of the results to trust (I'll be doing lots of designs which will ultimately be printed out). 

I even tried doing the following (not 100% sure this is accurate, but it did give the "mathematical" result of what I perceive with my eyes):

  1. Open my original image (sRGB), convert ICC profile to CMYK/8, copy image to Buffer
  2. Undo ICC profile conversion back to sRGB
  3. Convert ICC profile to Adobe RGB (it's big enough to fit sRGB and CMYK)
  4. Paste the image in the buffer and place it on top of the base layer
  5. Change Blend mode of the top layer to Difference.

And I get a very small difference in shift of colors (attaching PNG). The max RGB value which I sampled in the darks is #030303, min #000000, which means very close to black, which means (in context of Difference blend mode) very close to not being altered at all from the original.

So point being, the "bright up" effect might only has to do with the paper profile as @BofG pointed out. But then in this case, why doesn't the same bright up appear when I Soft Proof the image in CMYK profile (I select the same paper after all). It would just make sense (at least for me).
Unfortunately I don't currently have a printer to test the actual printing. But when I do get a chance and if I get interesting results, I'll post back here.

sRGB-CMYK-difference.png

Edited by JOleg
typos, rephrasing
Link to comment
Share on other sites

Color Think will provide an accurate color preview on the display for the various rendering intents and will provide the corresponding change in the ∆E false color map (see screenshots for differences based on specified rendering intent), as well as permitting the user to sample from the image to see the specific ∆E value for that pixel.

As far as quick or easy, one still has to verify that the soft-proof is useful by comparing it to a print under specified lighting.  Trying to control gamut between the source and destination color spaces does not require a print, by verifying that any adjustment one makes to the file to produce the desired printed appearance will require an accurate soft proof, otherwise one would inspect the print and make changes based on it.  I  do not come from a CMYK workflow, so please take my comments with whatever grain of salt might apply specifically to a CMYK production workflow.

I suppose that differences between a conversion into the CMYK profile and a soft-proof of that profile may be related to some soft-proof simulation parameter (like simulating paper white, or whatever) that is not exposed to the user.

kirk

CTcompare.jpg

Link to comment
Share on other sites

10 hours ago, BofG said:

Isn't that an apples-to-oranges comparison though? Really you need to take your printed material and compare it under correct lighting to an accurate monitor to see how accurate the soft proof is.

Lacking a printing device and correct paper, or a printout i can’t.

For me it seems soft proof combines too many things, at the same time is lacking essential functionality:

  • Provide preview of CMYK conversion (leaving aside paper etc) - missing
  • Provide preview of printout on paper - maybe?
  • Show areas where no 1:1 mapping is possible - grey overlay 
  • Show Delta E - missing

i would like to see these functions selectable (any combination which makes sense)

  • cmyk conversion 
  • Paper / printout effect preview
  • grey overlay of non-mappable colors
  • delta E preview 

If soft proof provides a preview of printout, i do not understand why this does not work if document is already cmyk.

One lasting issue is that Photo provides functions indistinguishable from Photoshop in name, but sometimes implements them slightly different. The help file and tutorial videos often do not clearly explain the crucial difference in functionality, leaving users (especially long time PS users) in the dark. Its lots of experiments, reverse engineering, waiting for James posting a tutorial, sweating and swearing. 

The next issue is Photo offers lots of UI functions completely useless in some combinatins, so you never know if it’s  a bug, user error, by design, etc. (layer fx and blend modes on masks)

Mac mini M1 A2348 | Windows 10 - AMD Ryzen 9 5900x - 32 GB RAM - Nvidia GTX 1080

LG34WK950U-W, calibrated to DCI-P3 with LG Calibration Studio / Spider 5

iPad Air Gen 5 (2022) A2589

Special interest into procedural texture filter, edit alpha channel, RGB/16 and RGB/32 color formats, stacking, finding root causes for misbehaving files, finding creative solutions for unsolvable tasks, finding bugs in Apps.

 

Link to comment
Share on other sites

1 hour ago, NotMyFault said:

If soft proof provides a preview of printout, i do not understand why this does not work if document is already cmyk.

I think either:

A) the whole thing is a mess.

Or

B) there is some 'paper simulation' going on, and there's a bug when the proof profile matches the document which means nothing is done to the colours, which means the 'paper simulation' doesn't get processed.

Whichever it is, there is certainly room for a lot of improvement.

Link to comment
Share on other sites

On 10/13/2021 at 11:09 AM, JOleg said:

...converting ICC profile from sRGB to CMYK + applying Soft Proof layer...

Perhaps it is helpful to keep in mind that colour spaces (format) and colour profiles are two different things.

Spaces are RGB and CMYK. Profiles are AdobeRGB, sRGB, U.S. Web Coated (SWOP) v2, PSO uncoated V3, ...
 

RGB SPACE is for Monitors/Displays adding light to mix colours.
=> Adding  Red + Green + Blue = white on your display.

Within  the RGB space, RGB PROFILES are for example

  • sRGB,
  • AdobeRGB,
  • ROMM RGB etc...

 

CMYK SPACE is for printing ink on paper to mix colours. 
=> Adding Cyan + Magenta + Yellow = Black (or close to black) + K (black) to make things even darker.

Within the CMYK space,  CMYK PROFILES are for example

  • PSO uncoated V3 FOGRA 52 ...
  •  U.S. Web Coated (SWOP) v2 ...

 

So, switching from one space to the other, you will always have to change the profile as well, or not? Correct me if I'm wrong here...
At least in Photo I can't select a RGB profile when I'm in CMYK and vice versa.

format profile.jpg

Link to comment
Share on other sites

By chance i found an article on Eizo explaining how to use soft proof in Photoshop.

It seems to confirm most of our observations, e.g. its complex, and Affinity lacks a clear definition and some functionality in that area. (Paper simulation on/off).

https://www.eizo.eu/knowledge/colour-management-calibration/colour-management-for-photographers/12-soft-proofing-in-photoshop/

Edited by NotMyFault

Mac mini M1 A2348 | Windows 10 - AMD Ryzen 9 5900x - 32 GB RAM - Nvidia GTX 1080

LG34WK950U-W, calibrated to DCI-P3 with LG Calibration Studio / Spider 5

iPad Air Gen 5 (2022) A2589

Special interest into procedural texture filter, edit alpha channel, RGB/16 and RGB/32 color formats, stacking, finding root causes for misbehaving files, finding creative solutions for unsolvable tasks, finding bugs in Apps.

 

Link to comment
Share on other sites

I guess it boils down to properly documenting how these features work in particular software. Usually we can live with how certain features are implemented, provided that we know exactly how they work, so that we can control them (e.g. we benefit from Blend Modes because we know how they mathematically operate, we don't guess). But in this case of Soft Proof, we do a lot of speculation about what happens under the hood. It would just be nice if exact process was documented better, even if this process is not exactly what we want from it, but at least knowing what it's doing would help achieve desired results with some extra steps.

Affinity team, if you read this, do you think you can update "Soft Proof adjustment" documentation with details required to answer questions raised in this thread?

Link to comment
Share on other sites

RGB and CMYK are "color models" - how color is represented.  sRGB, AdobeRGB, ISO, SWOP etc. are color spaces within their respective color models - think gamut volume as the space in which a color in a particular model (RGB) can live for that particular space (sRGB).  A color profile is a file on a computer that defines the color space (its "tristimulus values" for R, G and B) and its white/black point, among other characteristics.

On a Mac, you can use the ColorSync utility to view the contents of your color profiles (screenshot).

kirk

color.jpg

Link to comment
Share on other sites

On 10/15/2021 at 10:59 AM, Lagarto said:

They work so badly that CMYK profiles should not even be included in the list (not even when the document color mode is CMYK).

I totally agree. That's why I still have to fall back on Photoshop for most of my work.

Windows 10 Home 64 / AMD Ryzen 1700 @ 3.6 GHz / Asus Prime B350-Plus / 32 GB RAM / GeForce RTX 3060 12GB with latest studio drivers
Affinity Suite V2 latest official versions

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...

Important Information

Terms of Use | Privacy Policy | Guidelines | We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.