Soft Proofing and Color Management doing odd things

[hunchi]

Hi all,

I'm new to the Affinity Suite and am a bit confused about color management (and noticable difference in color rendering compared to Adobe)... Maybe someone can help me out here?

I'll attach the relevant data if someone wants to try out him/herself. Also, I'll use Publisher below, but i also checked with Affinity Photo - same story. 

I have also already printed the shot attached, so i know how it finally did print and the Adobe softproof is nailing it while the Affinity softproof is quite off.

1. Start with opening the template*.idml attached. It's a print template to be used by the print store... First confusing thing here: They mention to do the full workflow in sRGB, but the document itself opens with CMYK in Publisher. 
2. Import the jpg attached (the jpg is sRGB). You'll immediately notice the color gamut problem when importing in this CMYK document. Save the file as "CMYK.afpup".
3. Now in document settings, change the color space in this publisher document to sRGB. Save the publisher file under a new name e.g. "sRGB.afpub". Open both publisher documents side by side
4. Now add a soft proof adjustment layer to both files, i have attached the relevant *.icc file, too. Use "Relative colormetric" as intend and "black point compensation". 

My findings are:

• When working with the document in CMYK, the color rendering is way off (the pinks are way too blown out. There is significantly more detail and texture in the final print). When adding the softproof layer it corrects the orange/reddish inner part of the blossoms correctly, but the pinks are still wrong. 
• When working with the document in sRGB, the color rendering is at least ok (well it's sRGB, so yes nothing changes). However when i apply the soft proof layer to this file, then the contrast of the full soft proof picture decreases rather significantly (mostly visible around the butterfly). Also something that I don't see in the final print. 

So either way, things are off. Am i doing something wrong? If i export from either of both files (of course with the softproof layer disabled) to a PDF in print quality, sRGB and import this PDF into Photoshop and use the Soft Proof Function there, it works just perfectly fine (and is extremely close to the final print). 

Disclaimers: I have of course checked to have the soft proof options aligned. 

With that, I'm a bit lost how to adapt my workflow... it seems maybe it's best to just use sRGB for the document and don't do any softproofing in Publisher itself but export to PDF and do it in Photoshop then? Rather tedious...

Thanks a lot for any help...

Best regards!

Christian

template_42x28.idml SaalDigital_Fotobuch_matt_10-15.icc

[lacerto]

On 1/7/2023 at 12:45 PM, hunchi said:

First confusing thing here: They mention to do the full workflow in sRGB, but the document itself opens with CMYK in Publisher.

Whether an IDML file will be opened in RGB or CMYK color mode in Affinity Publisher seems to depend on whether the IDML file was initially saved targeted for Web / Digital Publishing (in which case the color mode will be RGB), or whether it was saved targeted for Print (in which case the color mode would be CMYK). The actual profiles used as document profiles would probably also be dependent on whether this was done when saving the IDML file but in this case document profiles were not specified, so both in InDesign and Publisher the default working space profiles would be assigned.

As for working with photos and printing to a photo printer that expects the input to be in sRGB, you should choose RGB as the document color mode and then ensure that sRGB is the document color profile. When you subsequently set up soft proof using the ICC profile specific to the printer, you should see approximately identical simulations in both apps:

a) Soft proof off (both apps using sRGB document profile):

b) Soft proof turned on (in InDesign Preserve RGB Numbers unchecked; in Publisher using Relative Colorimetric rendering intent and Black point compensation checked) :

The simulated image is a bit more contrasted in InDesign (but I think it is a bit more contrasted already when having the soft proof turned off). But differences are not big. Are  you seeing bigger differences?

Did you turn off Soft Proof adjustment when printing or producing the print file? (You should.)

When exporting or printing, you would produce sRGB color space, which the printer expects to be able to make a proper conversion. This workflow would make it possible to place RGB images in diverse RGB color spaces and produce consistent results.

[hunchi]

On 1/10/2023 at 1:16 PM, lacerto said:

The simulated image is a bit more contrasted in InDesign (but I think it is a bit more contrasted already when having the soft proof turned off). But differences are not big. Are  you seeing bigger differences?

Did you turn off Soft Proof adjustment when printing or producing the print file? (You should.)

When exporting or printing, you would produce sRGB color space, which the printer expects to be able to make a proper conversion. This workflow would make it possible to place RGB images in diverse RGB color spaces and produce consistent results.

Hi lacerto, many thanks for trying this out and the detailed and helpful answer - really appreciated and helps a lot - understood that sRGB as document color space makes more sense in this case and will go for that.

Indeed, I think i see some more differences - However I'm using Photoshop as the adobe solution as I don't have ID. Softproofing engine should be exactly the same for both adobe tools, though. Here's my view's:

a) without softproof, sRGB - identical from my perception. 

b) And then with softproof - contrast lacking in the publisher file (especially in the Bokeh and for the butterfly). Pinks are slightly different, but barely noticable i guess.

c) And zoomed in a bit while softproof remains on: 

d) here's the softproof settings for both

 

Exporting into the final print file (PDF) works fine - of course with the soft proof layer / view turned off. Also the PDF output when exporting into sRGB is 100% identical when exporting from Photoshop and exporting from Indesign. I'm not questioning the final output of Publisher which I believe is fine, I'm just wondering why there is such a delta in the proof view - shouldn't it be exactly the purpose of a softproof to have a (as) reliable (as possible) simulation of the final print result and shouldn't that come out the same in different tools with the same *.icc? 

Maybe the difference of the proof-views between adobe and affinity doesn't jump to the eye in the left/right comparison above. However, if i switch on/off the soft-proof in the tool itself, it is extremely obvious that

• in the adobe case apart from the blossoms nothing changes in perception. 
• in the affinity case, the contrast changes significantly (lower when soft proof is on). 

Thanks again. Will try one more thing later...

 

Edited January 15, 2023 by hunchi
edited for clarity

[thomaso]

1 hour ago, hunchi said:

d) here's the softproof settings

"Saal Digital" … compare this post + the Serif reply directly below:

[hunchi]

1 hour ago, thomaso said:

"Saal Digital" … compare this post + the Serif reply directly below:

Many thanks! Yes, i also called them and they immediately ask people to not use Affinity for softproofing - for the same reason as in the FAQ online, due to the inability to simulate paper. However, for the particular profile i am using (fotobuch_matt), they actual state, to *not* use paper simulation anyways, so I thought it's not hurting if Affinity is lacking this feature.

I tried one more thing: I checked the out of gamut warnings and they differ significantly. This might be the reason why the softproof rendered in Affinity appears flatter / lower contrast - it cuts off more of the darks as it says those are out of gamut. Screenshots below.

But now the surprising part (at least for me): I'm using SpyderX Elite, which comes with its own softproof tool and the simulation within this tool is actually closer to the Affinity render compared to the Photoshop render. This also becomes obvious when looking at the out of gamut warnings - very similar to affinity. 
To be honest, I have no strong knowledge about how these OOG warnings work and from the few articles I read online it seems this is not a feature recommended to be used nowadays anyways?

a) Photoshop - (almost) only the pink of the blossoms is OOG.

 

b) Affinity Publisher 2 - much more OOG, both in the blossoms but most notably also the butterfly and part of the bokeh dark olive green top right. 

 

c) Datacolor SpyderX Elite - very similar result compared to affinity (yet, not identical). (note: also when just looking at the target render, without OOG warnings, it's a bit flatter than adobe and closer to the affinity render). 

 

Well, that's the observation... now anyone who can draw a conclusion from that?

 

Thx again!

Edited January 15, 2023 by hunchi
typo

[lacerto]

6 hours ago, hunchi said:

Well, that's the observation... now anyone who can draw a conclusion from that?

IMO Affinity Out of gamut feature in context of Soft Proof adjustment is broken (has always been). If the feature is intended to be used in context of manual adjustment of off-gamut tones (e.g. using things like saturation control), it is clear that it should take into account different rendering intents in situations where these would be applied when the image is sent to printer and the printer takes care of color management (the procedure that is simulated when performing soft proofing). Manual adjustment could then be used to improve handling of single off-gamut color areas or tones so that they would still stand out (in case they would be muted by using standard rendering intents). On the other hand, if color management is done fully in software, then soft proofing would basically be unnecessary (other than just previewing targets in situations like checking adequate contrast in grayscale output, or determining similarity or difference of colors in RGB and CMYK color spaces).

The Soft Proof adjustment of Affinity apps has also a silly flaw as it would show the out-of-gamut warning color (instead of underlying image color values) when examining the image color values (e.g. using the Info panel of Photo).

I use X-Rite instead of Spyder so I cannot say why Spyder might be behaving similarly as Affinity apps, but IMO the way Photoshop behaves (rendering intent having effect on showing out-of-gamut color values) makes sense.