Colour Management in PL6

Do you mean that even work done in PL<6 is being rendered differently in PL6? That is, just open an image that was edited in a previous version and get a slightly different rendering in an export and/or on the display?

That’s a real line in the sand for me since it invalidates editing done in previous versions.

1 Like

I’ve seen slight differences when I stacked two exports (DPL5 and DPL6 at legacy WCS) in Photoshop and amplified the difference as shown in a post or two. Differences can exist but are really subtle in a way that they are well within any tolerance of involved gear/software/processing etc. Moreover, differences show in images with saturated colours mostly.

I’m not that obsessed with tiny differences, after all, we’re not feeling the same every day…

1 Like

Interesting, and I agree. I know color rendering has changed a bit across previous versions of PL and its predecessor, too - sometimes significantly (and IMO for the better since OpticsPro became PhotoLab). Thinking about this, though… If the differences between PL5 and PL6 (Legacy WCS) are only measurable in export or with soft proofing, they can probably be attributed to the new algorithm meant to preserve detail in saturated colors (the PSCA). Otherwise, perhaps there’s more complex tweaking of the pipeline in PL6 or some difference in the ICC profiles used for export? Or even differences in the noise reduction algorithm or another part of the demosaicing process?

The default setting of Smart Lighting seems less reliable to me in PL6, for what it’s worth. Might just be how it handles the DxO Wide Gamut WCS, but I wonder if it could have been adjusted in a way that impacts results with the legacy WCS, too. One other thing that’s bitten me a couple of times now is that some lens correction profiles aren’t working properly in PL6. (Too much vignette correction or not enough.) And the FilmPack features keep being affected by new PhotoLab releases.

So there’s a lot to test if one wants to isolate the reason for rendering differences in PL6 vs. PL5. But I am guessing this really isn’t a worry.

Sure, but small differences over many versions will add up over several upgrades. I’m not keen on having to see PL as a tool that I have to continually export from to retain a rendering I’ve arrived at.

I remember Skylum Luminar 2018 (which I otherwise quite liked) being unusable across even minor updates.

While I can absolutely agree with your reluctance, I nevertheless have to say that processing in DPL is an ever evolving thing, unlike a TIFF or JPEG file, both of which are fairly old and tried formats.

Therefore, it might be wise to keep processed files of the few extraordinary shots we took, in order to preserve what we thought optimal at the time.

2 Likes

PL6 Wide Gamut vs. PL6 Classic-Legacy

Complementing my findings from post #31, I compared DxO’s Wide Gamut WCS with the Classic-Legacy and also included an export from the former PL5.

like before → DxO Standard, sRGB screen
_DSC0311.NEF (15,9 MB)
but new
_DSC0311.xmp (8,0 KB)
_DSC0311.NEF.dop (41,0 KB)

.
1.


pic #1 = WG

Wide Gamut = orange
Classic-Legacy = blue
softproof & export with sRGB IEC61996-2.1 profile

PL6
pic #1 = Wide Gamut
pic #2 = WG / softproof from pic #1
pic #3 = Classic-Legacy
pic #4 = CL / softproof from pic #3
.
pic #5 + 6 = WG / export from pic #1 + 2
pic #7 + 8 = CL / export from pic #3 + 4

PL5
pic #9 = export from PL5 with sRGB profile

.
2.


pic #2 = WG / softproof from pic #1 → Monitor gamut warning

.
3.


pic #4 = CL / softproof from pic #3 → Monitor gamut warning

.
4.


pic #2 = WG / softproof from pic #1 → Destination gamut warning

.
5.


pic #4 = CL / softproof from pic #3 → Destination gamut warning

.
comparing the screenshots

  • 2 vs. 3 → the Monitor gamut warning area in screenshot 3 is slightly reduced
  • 4 vs. 5 → the Destination gamut warning area in screenshot 5 is slightly reduced

which means …

  • The PL6 Classic-Legacy mode reduces some of the out-of-gamut colours
    and the export output from Wide Gamut is richer than from Classic-Legacy !
    → export from pic #1 or 2 (WG) and from pic #3 or 4 (CL) to compare them
    → add an export from PL5, which is identical to pic #7 and 8 (CL)

  • Make use of the softproof to get a “true” WYSIWYG → PL6.

Of course, but if PL becomes a moving target from version to version then there are other applications that do a better job. Both Capture One and Exposure inspire more confidence with their process versions than does PL’s non-existent version handling and best effort approach to backwards compatibility.

FYI: I’ve just made a very concerning “discovery” …

image

The new Wide-Gamut option for working color space IS included with the Essential version …

But, Soft Proofing is not !!

image

Wow ! - That really is NOT a good thing … because, without Soft Proofing, a WYS-is-WYG result is NOT guaranteed.

John

2 Likes

Its baffling what they do at times and how they arrive at what they do is beyond me as well. Like lack of support for phone DNGs, the fastest growing sector but to not do soft proofing when its clear for many this is the best way of arriving at an accurate output with wide gamut is beyond me.

1 Like

They have also messed up the ability to use DCP. I opened a few old images and many had distortion which was only removed by changing from my DCP to the default PL one. I find I now have a large number of edits made usless in 6 by these changes unless I remove the DCP used,. Clear areas such as the sky in this one are where it shows most.

DCP used

PL camera used
DCP turned off and PL defult Camera used

Put it in as a bug report, makes use DCP use impossable as you can’t prdict what stuped things they will do in the future

Does that weird haloing show up in export or just in the viewer? I’ve seen it in the viewer without using dcp profiles and a restart of PL6 fixed it.

This brings me back to the option to select a display profile which was removed from PL6 and I believe would resolve a lot of issues if it was brought back. Simply select a display profile such as sRGB and PL6 should work just like PL5.

1 Like

I’m under the impression that it doesn’t need to be brought back. For one thing, it might be possible for PL6 to get the display profile from the OS. But more immediately, Soft Proofing replaces the option in PL6 and ensures WYSIWYG. It’s an all-in-one tool for representing the output accurately, which I don’t mind even if the idea of using it for your primary monitor is very weird. But that makes Soft Proofing an ESSENTIAL tool, right? :smile: So rather than rewrite a redundancy into PL6, I think PL should rethink what the Essentials version of PL should be, not to mention the relationship between PL and ViewPoint (which is also strange).

That makes no change at all and its there in exports.
Export of original V5 imige2 from V6

Screen clip from Soft Proofing

Its more fundermatal I fear, they have messed up DCP with V6

Even if it went in export it would result any editing being totally unreliable.
PL has been restarted a number of times as well.

The original file, but I have the same on many others as well.
_DSC2815.ARW (23.9 MB)

DCP
3_CF A6400 Medium V2.zip (221.7 KB)
dop is usless as with out DCP being installed in the same location as I have the imige will not open (another bug wating to be corrected)

What does DCP mean?

George

1 Like

https://rawpedia.rawtherapee.com/How_to_create_DCP_color_profiles

Its a problem opening imiges edited in 5 (or earlier). I copied imige to clean folder and opened in 6 used all same adjustmentss no problem. Deleted database and dops in the originl folder and again no problem. Put the original dops back and the problem is back.

I just downloaded your image and loaded it in both PL5 and PL6 with no apparent problem like you are demonstrating.

What am I meant to have to do to reproduce it?

I think it’s a problem from 5 into 6. I redid it clean in 6 and no problems. Replaced with the original dops and back.
f that’s so it could be the problem is a possably of how DCP in 5 worked and 6 not dealing with this diffrence in earlier versions? But the result is any edits with DCP may have to be redone if used in 6.