PLv6: Color Rendering observations

I’ve noticed some subtle color rendering changes with PLv6;

  • Pre PLv6, when the Color Rendering tool was deactivated - the result would default to the Generic Rendering for the DxO Camera Profile … That is, DxO’s rendering for the camera body.

  • Now, with PLv6, it defaults to Generic Rendering - Neutral Color
    – This can be easily confirmed by assigning these settings and toggling Color Rendering activation


  • I know that quite a few (experienced) PL users apply Generic Rendering - Neutral Color as their starting-point for Color Rendering;
    – However, with the Wide Gamut WCS, I’m noticing that this rendering causes some colour-shifts in saturated colours … for example, saturated orange-colours shift to bright-reds … etc.

  • As an alternative to consider, I’ve found Camera Body = DxO One to be very natural … an excellent starting point - and without the colour-shift issues noted above.
    – I’m guessing that DxO put solid work into this, to reflect well on their own “One” camera.


And, of course, to be sure that what you’re seeing on-screen/within-PL is exactly what you’ll get when you Export-to-Disk and view the result on the same monitor … always have Soft Proofing activated.

John M

Neutral presets and DPL6 don’t agree with each other currently. Until DxO fixes this, DPL6 is not ready for action, unless one is willing to start with DPL6 as if it were some other app: learn from scratch. For all who have used non-neutral presets, the upgrade to DPL6 is possibly less annoying…

That’s how I had set up PL5 – and PL6 simply overtook these settings (DxO Standard …). :slight_smile:

This is the main reason that I am sticking with PL5 for the work I want to get right, without having to work out what does and doesn’t work.

Genuine question: What’s the benefit of using a Neutral preset as one’s starting point ? – Doesn’t that just mean that more work is required to get to the desired end result ?

John

@John-M I agree with your question and was about to ask @platypus what a “Neutral preset” was and
I looked at my second monitor where I have been testing situations with respect to @Hugo’s post In V5 processed images look different in V6 and discovered that “somehow” (!?) I had the following

I am not sure what you are saying or whether my testing was wrong!

Starting point:-

Switching to ‘Wide Gamut’:-

Switching has left the setting the same but DxPL(Win) has severely restricted my options!

@John-M if I deactivate the option how do I know what will be used (sorry if I am being “thick”)?

Switching between the two options with ‘Soft Proofing’ on and off produced a screen reaction regardless?

This is not an area I normally “fiddle” with so what am I not understanding and what should I be seeing?

The benefit of using a neutral preset is about seeing the potential, it’s not about “better”.
It’s about having a canvas and a palette instead of some “happy meal”.

  • I like to see what’s in a photo (before I start editing) and neutral presets provided
    that capability so far, while presets like DxO Standard e.g. often hide what’s in the lights
  • Coming from film photography, I prefer natural colours to what we often call the “digital” look

Yes and no. What’s the difference between a) having to tone down an image to only then discover what has been hiding behind a preset’s “magic” agains b) tuning something that shows its potential right from the start? A flat image might look unexciting at first, but I prefer to make it shine myself.

Highlight recovery? Many apps brag about their ability to recover highlights…which is easy to prove if you first blow them by whatever setting is applied by default: A few of my old images always lit the “blown highlight” warning in earlier DxO apps, but DxO has corrected this issue over the last few years. Nevertheless, I often find DxO’s processing too heavy handed.

Example: Monument Valley with overcast sky as seen with different presets, Can you see the differences in your browser?


From left to right

  • No Correction Preset
  • DxO Standard Preset
  • DxO Standard Preset + Neutral Preset
  • DxO Standard Preset + Neutral color, realistic tonality (gamma 2.2) rendering

Anything is possible from any starting point:


Note: Minimal editing used (1-3 settings each)

Thanks for the detailed explanation … I’ve oft wondered about the purpose of a “neutral” starting point.

Yes - - I can see what you mean.

John

Just saying that;

  • With PLv5, if you assign Color Rendering to Generic/Camera Profile - then toggle CR On/Off, there would be no change to the resulting color rendering
    – That is, the DxO profile for your camera (E-M5) was the PLv5 default setting for Color Rendering

  • With PLv6, if you assign Color Rendering to Generic/Neutral color - then toggle CR On/Off, there will be no change to the resulting color rendering
    – That is, for PLv6, Generic/Neutral color is now the default setting for Color Rendering.

Nothing to do with the working color space setting.

John

…looks like we have to check and change what we set as our default presets, including settings of the WCS and SP tools. :man_shrugging:

@John-M I am not sure that my computer agrees with your statement, i.e. I don’t think that is what I see with respect to PL6, presuming that I understand what you say, and I am starting with the correct settings and going where you suggest I go!? The camera has changed in this case because I am using one of my images instead of an @Hugo image!?

I have selected ‘Camera profile’ and the adjustment is ‘active’:-

I have deselected ‘Color rendering’ and

Can you see any difference between the two screen snapshots? They are a different size?

and there is indeed a slight difference

Color Rendering ON, Generic Neutral ON:-

Color Rendering OFF, Generic neutral ON:-

So, if I conducted the test correctly we have (for PL6)

  1. No ‘Color Rendering’ variant ON is identical to the OFF setting (the default as you describe it)

  2. ‘Camera profile’ is nearly identical but nearly is not exact

  3. ‘Neutral Color’ is way off (by eye let alone by software comparison)

What have I done wrong!? In all cases the Beyond Compare comparisons are between exported JPGs not between the screen grabs (although that could also be done but my grabs are not identical in shape, haven’t mastered that bit yet but could have just grabbed the whole screen).

I was aware that it had nothing to do with the WCS, set to ‘legacy’ these tests.

I am sorry if I appear “puzzled” by some of this but during my testing I have encountered situations where the drop down options for ‘Rendering’ are severely curtailed and have not been able to identify the exact sequence necessary to reproduce the “restricted” list and whether it is legitimate or …

With respect to your observations and my testing please correct any errors in my process, I will repeat the tests with PL5 when I have more time!

Regards

Bryan

EDIT 1:-

Test repeated in test directory (not original file directory). [M]aster contains image with ‘Neutral Color’ and VC[1] contains ‘Camera Profile’. First entry with CR=ON, second entry with CR=OFF.

Summary of comparisons

Row 1:- NC (CR=ON) versus NC (CR=OFF) & CP (CR=ON) versus CP (CR=OFF)
Row 2:- NC versus CP (CR=ON) & NC versus CP (CR=OFF)

EDIT 2:- The promised (threatened) PL5 test summary

Comparing PL5 with PL6 currently has a complication! I did PL6 tests first and there is a ‘Tone Curve’ in use and replicating that exactly in PL5 is … The tests should have been done without the tone curve or a PL5 preset created and imported into PL6, presets don’t go the other way!

Until a few days ago, I never noticed a difference with Color Rendering on or off as specified above. But a few days ago I did and simply attributed it to the “Protect saturated colors” slider not being zero. You have it set to “1” in your first comparison, Bryan. Is there still a difference if it’s zero?

You didn’t set the Working Color Space to DxO Wide Gamut. That’s the only case where neutral color should be the same as disabling the Color Rendering palette. Otherwise, with the Legacy/Classic WCS, DxO camera rendering is more or less the same as having Color Rendering off.

2 Likes

Hi Bryan - -

It seems to me you’re “over-thinking” this. There’s no need to Export-to-Disk - that’s not what I’m referring to.

  • The difference in behaviour is evident directly on your PL screen.
  • The difference relates to PLv5 versus PLv6 … and, within PLv6; for Wide-Gamut versus Classis

To confirm, just do this, and nothing more (or less).

Open any image in PLv6 … and set WCS = Wide Gamut

  • Select Color Rendering, Category = DxO Camera Profile (DXO’s interpretation for THIS camera)
  • Toggle Color Rendering between activated and deactivated.
  • You should see changes in the rendering of the image on-screen/within-PL

  • Select Color Rendering, Category = Neutral color
  • Toggle Color Rendering between activated and deactivated.
  • You should NO changes in the rendering of the image on-screen/within-PL

  • That is; PLv6’s default setting for Color Rendering (with WCS = W-G) is Neutral color

Repeat, with WCS = Classic/Legacy … or, repeat using PLv5.

  • Select Color Rendering, Category = DxO Camera Profile (DXO’s interpretation for THIS camera)
  • Toggle Color Rendering between activated and deactivated.
  • You should NO changes in the rendering of the image on-screen/within-PL

  • Select Color Rendering, Category = Neutral color
  • Toggle Color Rendering between activated and deactivated.
  • You should see changes in the rendering of the image on-screen/within-PL

  • That is; the default setting for Color Rendering (with WCS = Classic) is DxO Camera Profile


Previously, (using WCS = Classic) I would leverage this behaviour to enable a quick comparison between the rendering for/of my camera versus some alternative rendering.

Now, (using WCS = W-G) it enables a quick comparison between Neutral Color rendering versus some alternative rendering … … Personally, I prefer it the way it is when WCS=Classic.

John M

1 Like

John, this is a wonderful post that, I think, makes things clearer.

Unfortunately, for me, it also highlights how difficult it makes using PL6 for the uninitiated who have never before had top make such “life changing” decisions, as to choosing default renderings, etc.

I still find myself putting off making the jump to PL6 and am continuing with PL5 for work I need to do without all this extra consideration.

Anyway, I am in the midst of preparing a small exhibition and it is all B&W Jazz photos, so I could really care less about colour gamut :sunglasses:

1 Like

@John-M I am not overthinking this at all, I am attempting to verify your statement in a reproducible manner and gathering the evidence so that I can evaluate whether I understand what you have said and agree with your statement in full or am still “lost”.

One of the “problems” with a topic like this one is that we are talking about visual aspects that are being perceived by users, in their homes with their settings! The “need” for exporting is to be able to undertake a reproducible, quantitative assessment rather than a subjective, qualitative assessment particularly when I have typically used ‘Generic Rendering’/‘DxO camera profile …’ rendering and not “bothered” with anything else!

Are you taking into account whether ‘Protected saturated colors’ is set to ‘Auto’ or being left at ‘0’ as @Egregius indicated that the snapshot showed a setting of 1 rather than 0 , given that the ‘Auto’ setting does nothing with ‘Wide Gamut’, i.e. it doesn’t “work” (and it doesn’t make “much” difference)!

Which of the parameters makes any difference at all is immediately obvious when an image is exported and compared, albeit using a JPG which is itself subject to algorithms to compress the image!

Incidentally, when you made your original statement about the change you did not explicitly state that you were comparing PL6 with ‘DxO Wide Gamut’ to PL5, i.e. not a “like with like” comparison but a “new with old” comparison!

My tests above were deliberately comparing PL5 with PL6(Legacy) to verify, or otherwise, DxO maintaining the “status quo”. I exported because I haven’t evaluated the accuracy of ‘Soft Proofing’!?

So back to your challenge but quantitatively assessed and part of a larger test that I had already started but may not complete depending on the outcome of this evaluation!

  1. Copy the RAW image to 2 directories PL5 and PL6. Then add PL6(WG) as a result of @Egregius responding to my earlier post! This grew somewhat until I wound up with

image

(which I added to for this post - being a pedant has its price!!)

  1. Add 4 VCs to each [M]aster which is assigned the ‘5 - No correction’ preset as set in the ‘Preferences’.

The [M]aster is to remain unchanged but the VCs are set up as follows (and this is a change of order compared with the previous test, i.e. it is now Neutral on and off, followed by Camera on and off)

VC1 = Neutral colour, realistic tonality (gamma…) on PL5 and ‘Neutral Color’ on PL6 CR=ON
VC2 = VC1 settings but CR=OFF
VC3 = ‘DxO camera profile (DC-G9)’ on both releases
VC4 = VC3 settings but CR-OFF)

  1. Export all Images from [M] to VC[4] in 2 directories using PL5 and PL6 ‘Classic (Legacy)’ etc.

Now for the @John-M “challenge” using data already created for my other tests but verified for this test!!

This is VC1 versus VC2 ('Neutral …, ON & OFF)

This is VC3 versus VC4 (‘Camera…’ ON & OFF)

PL5 VC1 versus VC2:-

PL5 VC3 versus VC4:-

With ‘Protect saturated colors’ = 0 this becomes the following

I have had problems getting consistent results because of errors in settings but these appear to fit with your analysis with the “slight” caveat that the ‘Psc’ option has a slight impact and you do need to export and compare and provide snapshots if you intend to prove something to a doubting “Thomas”!

PS:-

Previous tests have shown PL5 (Psc=Auto) versus PL6(L)(Psc=Auto) and PL5(Psc=0) versus PL6(L)(Psc=0) produce identical results in the above tests scenario.

PPS:-

  1. If the change is documented then it is a “feature” if not then it is a "bug and should be reported as such @DxO_Support-Team.
  2. If the lack of a working ‘Auto’ feature with PL6(WG) is documented with a good reason then it is a “feature” if not then it is a “bug” and I should report it!!

Thank you for a very clear explanation.

The PL6 manual from October 4 states following regarding Neutral Colors.

Neutral colors are identical to DxO Standard, except that the colors are less saturated and the contrast is less pronounced.

1 Like

@Required presumably relating to PL6(WG) the comparison software I am using can show differences but not in the way described so

PL5(Psc=Auto) V PL6(WG):-

VC1 v VC2 ‘Neutral…’ ON & OFF
VC3 & VC4 ‘Camera…’ ON & OFF

                 ON                                        OFF

1 Like

@Required ‘1 - DxO Standard’ is one of the “standard presets” provided by DxO, as is ‘3 - Neutral colors’ they “happen” to invoke one or other of the ‘Color rendering’ options that @John-M is referencing, i.e. ‘1 - DxO Standard’ uses ‘DxO camera profile…’ and ‘3 - Neutral colors’ uses ‘Neutral color’ on PL6 and ‘Neutral color, neutral tonality v2’ on PL5 which is actually different to the option I had chosen which was the next option (the last option) in the list, so I need to repeat the tests or not, life is too short!