[BUG] wys IS NOT wyg - CM(?) problem

[andysalay]

Hello, I was in a positive mood, happy to provide a Show-off image, then I realized something weird.
ASP is not always showing the same colours as what you get in JPEG after conversion.

It seems problem is visible only in some circumstances, maybe on darker shots brightened more.
Above is RAW, under is JPEG from this raw, without any edits. (This JPEG is shown with the exactly same colours in other colormanaged viewers - XNview, XNviewMP)
I have EIZO FlexScan S2231W, with default factory S2231WD6.icm.
There are no problems with Bibble5.

With provided RAW, it is enoug to push
Exposure +1.0
Blacks + 2.0
Saturation +16
Vibrance + 10
and after JPEG conversion you get visibly different image than ASP RAW on screen rendition.
JPEG on screen rendition seems enough similair to XNview rendition.

I have turned off ASP sharpening, and have not used any noise reduction and plugins just in case.

Did someone encounter similair problem?
Can somebody confirm this on his HW/SW, with my files or even better with his?

Regards,
Andy Salay

Attachments:
RAW, XMP, JPEG, BatchQeue

[Gerard2]

just as a check: there are no "default settings" re running on the jpg ?

Did you try softproofing to srgb ?

[andysalay]

Thank you Gerard,
Softproof was off, of course, I know the effect of softproof (seems broken as well, generated extreme change, but thats for another investigation and possibly bug report).
As my EIZO has 92% of AdobeRGB gamut, the differences with soft proofing should be noticable regarding saturation, and they were exactly that in Bibble5.
But in ASP softproofing to SRGB significantly clip blacks. What is going on here?

Checked JPEG defaults once more right now, and they are empty.
DefaultRendered.xmp

Code: Select all

<x:xmpmeta xmlns:x="adobe:ns:meta/" x:xmptk="XMP Core 4.4.0">
 <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
  <rdf:Description rdf:about=""
    xmlns:bib="http://www.bibblelabs.com/BibbleToplevel/5.0/"
    xmlns:bset="http://www.bibblelabs.com/BibbleSettings/5.0/"
    xmlns:blay="http://www.bibblelabs.com/BibbleLayers/5.0/"
    xmlns:bopt="http://www.bibblelabs.com/BibbleOpt/5.0/">
   <bib:settings>
    <rdf:Description
     bset:settingsVersion="66"
     bset:curLayer="0">
    <bset:layers>
     <rdf:Seq>
      <rdf:li>
       <rdf:Description
        blay:layerId="0"
        blay:layerPos="0"
        blay:name=""
        blay:enabled="True">
       <blay:options
        bopt:autolevel="false"
        bopt:sharpenon="false"/>
       </rdf:Description>
      </rdf:li>
     </rdf:Seq>
    </bset:layers>
    </rdf:Description>
   </bib:settings>
  </rdf:Description>
 </rdf:RDF>
</x:xmpmeta>

Regards,
Andy Salay

[andreas]

Do not expect to see the same in ASP (16bit color managed in working color space) as you get in JPG (8bit sRGB?).

But if you do use softproofing to the same colour space you are exporting in your jpeg they should be the same.

Andreas

[andysalay]

Do not expect to see the same in ASP (16bit color managed in working color space) as you get in JPG (8bit sRGB?).

But if you do use softproofing to the same colour space you are exporting in your jpeg they should be the same.

Andreas

But while "dark violet" colour fits well inside sRGB gamut, and "dark cyan", too, shouldn't these two colours stay almost unchanged on colourmanaged display? (I know about colorimetric intents, I think on win is perceptual by default, isn't it? So when saturation not, at least the hue should be the very similair...)
But I am getting completely different hue on the sky. It does not seem to me like a so-rich colour that must be out of sRGB gamut.

Anyway, thank you. You provoked me to look even deeper, and to repeat the whole procedure on Bibble5. It seems I have overlooked that in Bibble the effect (difference) is similair, albeit a little bit weaker. So I am getting the above paragraph now.
Maybe the greater saturation - especially noise in the sky in the blue/green channel had caused significant out-of-sRGB-gamut shifts, which were afterwards moved back by CM calculations when exported into sRGB. So it seems the PICNIC error (Problem In Chair, Not In Computer)
Maybe I get used to my EIZO colours so much, that I already forgot how small sRGB is.

What scares me though, is the blacky-blacky sRGB softproofing in ASP. In Bibble5 everything stays the same (especially luminance), except the blue sky discussed above.

What do you think of it?

PS: funny, when widegamut of my monitor (0..255 means greater range than 0..255 in sRGB) was stuffed into the sRGB (profile change by mere assign, not convert - it was pure screenshot saved through mspaint and converted to JPG), the blacking looks not so strong in my colourmanaged Mozilla (it assumes that JPG screenshots are in sRGB).
Anyway, on my screen in ASP the shadow detail is completely blacked. Why? Does greater gamut of my monitor mean also darker blacks? Thus, does it mean that total black (luminance 0) in sRGB does have higher luminance than total black (the same pixel value of luminance 0) in Eizo gamut? So zero != zero? It seems total nonsense to me. But the observations are giving me impressions that maybe... I am getting confused.

Please, help me, if you can.

Regards,
Andy Salay

[andreas]

I downloaded your files and enabled soft proofing to sRGB. This is my hardcopy. They look very nearby. NEF is the lower image, JPEG the upper one. I see very little change in the JPEG when enabling sRGB-soft proof. So on my display it looks correct. But this is a samsung syncmaster and not an eizo. Did you try with different profile types? Really correct and actual profile selected in ASP? I do not know enough about CM to give you deeper insights.

Andreas