@platypus On DxPL Windows I don’t see any changes in the menu items between PL5.5.0 and PL5.7.0, you have to hover over the last entry to get the full details with DxPL(Win)
However when I was testing trying to get consistent results between tests was more of a nightmare than it should have been but I believe that the results are accurate so we certainly have differences between the outputs from PL5.5.0 and PL5.7.0 on Windows, in this case showing as different file sizes for the exported files for a start!?
There is a difference between my two systems, the PL5.7.0 release also has FilmPack installed! However, the only options used on both systems are ‘Color Rendering’ and ‘Distortion’.
FastRawViewer shows the following but I am not sure that helps!?
However, under no circumstances should DxO make any changes that affect the outputs from DxPL without at least some notification and between point releases?
Arguably it should always be possible to revert so that a user can continue to use their original edits and achieve identical results.
This is PL5 so no ‘Wide Gamut’ etc. so why the change!?
There’s been an identified difference/bug in applied noise reduction outcome between those versions on macOS - so it might be a cross platform code change but with different effects.
@Required thank you for the update, I suspected it had been found already and the sharpness comparisons from FRV seemed to show some difference but not as obvious as the BC comparisons.
But I had only ‘Colour Render’ and ‘Distortion’ set!?
the change that’s been done is just the naming: the one that was named “Neutral color, realistic tonality (Gamma 2.2)” / “Neutrale Farben, realistische Tonalität (Gamma 2,2)” is now named simply “Neutral color” / “Neutrale Farben”. Thisis the same rendering behind the scenes, and will have the same result.
@Wolfgang sorry it has been a long week, I actually started out thinking today was Thursday instead of Saturday (actually Friday but I was still a day out).
When you say all 4 versions give different results you are confirming differences between 5.7.0 and 5.6.0 (or 5.5.0 as I found because that was the version on my Test machine)
But
and
Which I do not appear to be seeing for whatever reason @Barbara-S.
Is that also your belief @Wolfgang i.e. that the output from an earlier release differs from PL5.7.0?
While rendering in DPL 5. seems to be mostly identical, DPL6 renders “Neutral Color” quite differently, even if WCS is set to “Legacy”. Sadly enough, this change is not mentioned in the update window or in the release notes.*
I’ve also been comparing PL 5.5.0 and PL 6.2.0 behavior regarding the various neutral renderings. Please try this: In PL6: Classic (Legacy) color space; optical corrections only; WB = As Shot; no DxO Smart Lighting; Color Rendering = Generic renderings – Neutral color.
In my set up, this produces seemingly identical images / histograms to that produced in PL 5.5.0 using the same options as above except choosing Neutral Color, Realistic tonality (gamma 2.2).
This may be what Barbara-S is trying to tell us.
My set up: Leica Q DNG raw files with in-camera aRGB jpeg setting; aRGB display.
"@eriepa I am sorry but unless I am doing my tests incorrectly that statement is incorrect PL5.7.0 simply does not produce the same image as PL5.5.0 for any of the “so called” “neutral” options.
1. The sizes of the exported JPGs differ 2. the bit maps are simply not the same"
Both statements are correct but the differences between the two releases is minute and not as bad as in my post.
The reasons for such a wide discrepancy with the original tests remained a mystery or rather it was until I output PL5.7.0 at different %ages when the problem became clear, somehow I had selected 96% for PL5.7.0 not 97% which is my normal amount and the amount I had used for the PL5.5 tests.
I must apologise to DxO and the forum members for my lapse in concentration, now back to the real world with 100%.
The following Tests are what prompted my re-investigation of the problems with my original test report.
and the PL5.7.0 images appear to correspond exactly to the PL6.2.0 images, both in size and image content, notwithstanding the name change.
@eriepa unless I have done my tests incorrectly (I only tested PL5.5 against PL5.7 for my earlier post) PL5.7 and PL6.2 produce exactly the same results, as I stated above.
Please supply snapshots or an image with the PL5.7.0 DOP and the PL6.2.0 DOP so we can test them.
If the differences were on an earlier PL6.x.x release then please identify that release so that we can test it.
I just tested on PL6.1.0 and PL6.1.1 and the exported images do not agree with PL5.5.0, PL5.7.0 or PL6.2.0 (the latter two of which now appear to the identical).
So there appears to have been a change to restore the original renders in release PL6.2.0?
PS:-@Barbara-S I apologise for “maligning” DxO with respect to the PL5.5 versus PL5.7 tests, ther are differences where there should be none but they are tiny compared with my original report which was an error on my behalf!
BUT
Why was DxPL 6 released with the differences without
Tests to verify that no differences existed, much as I have done (hopefully correctly this time). Please purchase a copy of Beyond Compare for both platforms, download the various image extensions and use it to compare outputs when things are supposed to be the same (I do not get an commission from any software manufacturer and never have)
If any such changes exist then users should be warned and if DxO did not know about the changes they most certainly should have, i.e. we either have negligence in reporting to the user base or negligence in testing and there is no excuse on either account.
The reversion to the old render “scheme” should have been made clear in the PL6.2.0 release notes, if it was I will need to apologise
@Wolfgang The options give different “neutral” renders between them it is true but you have options as to which “neutral” you prefer. But they also gave different renders between different releases (until now, almost still a teensy weensy niggle between PL5.5 and PL5.7) as my report above indicates.
Bryan,
I was responding to @Barbara-S and the difference between Mac … and the Windows version 5.7.0 build 4820, but NOT going back to earlier PL5.x versions and also NOT comparing vs PL6.x
BHAY / Bryan - thanks, but you are responding to remarks by platypus, not eriepa. I did not compare different PL5.x versions, only PL5.5.0 vs. PL6.2.0.
@eriepa as a direct result of your remarks I revisited my PL5.5.0 to PL5.7.0 tests and found the errors in my testing and changed them and added PL5.7 (and PL5.5) compared to PL6.2.0 tests (and then PL6.1.0 and PL6.1.1 tests).
It was your remark that triggered my testing and it was to your remark I was responding while revising my remarks to @platypus
In my earlier post I did describe how, from specific initial settings, the Neutral Color, Realistic tonality (gamma 2.2) in PL5.5.0 seems identical to the PL6.2.0 Neutral color rendering. Thanks.
I think you continue to obfuscate the situation, but whatever. Let’s move on.
Yesterday in this topic, I elaborated on my methods and results comparing the various neutral color rendering options in PL5.5.0 vs. PL6.2.0. What I found was, from specified initial settings, the Neutral Color, Realistic tonality (gamma 2.2) in PL5.5.0 seems identical to the PL6.2.0 Neutral color rendering.
I posted additional details earlier today in the topic concerning camera-specific “linear” profiles. I paraphrase that post then add additional observations and comments below.
I’ve been comparing PL 5.5.0 and PL 6.2.0 behavior regarding the various neutral renderings. Please try this: In PL6: Classic (Legacy) color space; optical corrections only; WB = As Shot; no DxO Smart Lighting; Color Rendering = Generic renderings: Neutral color.
In my set up, this produces identical TIFF images / histograms to that produced in PL 5.5.0 using the same options as above except choosing Neutral Color, Realistic tonality (gamma 2.2).
The other three PL5.5.0 neutral renderings result in distinctly different images / histograms for their TIFF outputs. To date, I have found nothing in PL6.2.0 like these three other neutral settings, nor do I expect to find any.
Take note that the initial starting settings are critical in making these comparisons. The DxO wide gamut setting, one example, will produce different images / histograms that have no counterpart apparently in any of the PL5.5.0 neutral color renderings.
My set up: Leica Q DNG raw files; aRGB display; images exported to disc as 16-bit TIFFs with an embedded Adobe RGB 1998 profile, then examined in a number of different applications.
Regarding your methodology, I suggest you switch to full-size TIFFs and specify your exact initial settings. Thanks.
@eriepa I am sorry that you feel I am trying to “fudge” the issue in some way that is the furthest thing from my intentions.
From my perspective there is no difference between PL5.7 and PL6.2 for the 4 ‘Neutral options’, the JPGs are the same size and Beyond Compare bit by bit comparisons shows no difference between images, i.e.
The first 3 items match in title between the two releases and the fourth is ‘Neutral color, realistic tonality (gamma 2.2)’ in PL5.7.0 and just ‘Neutral color’ in PL6.2.0 and in spite of the title I have tested these in the “hope” that they will be the same, i.e. that there is continuity for the user in spite of a change in name within the product.
So, after discovering that I had managed to set some options on in the PL5.7.0 copies (when they should have been off !?) I finally got what I had before, all JPGs are identical between the PL5.7.0 exports and the PL6.2.0 exports and then exporting to TIFFs gives even bigger files but they are identical in size and in contents, I haven’t used my eyes or histograms just a bit by bit software comparison.
@eriepa I am sorry but I don’t see any difference between PL5.7.0 exports and PL6.2.0 exports, on my Windows system, which was not the case with previous PL5.x.x exports and PL6.0.x and 6.1.x releases.
I had trouble with the re-runs because options had been turned on in the PL5.7.0 product and I am not sure I understand how or why that happened but when I returned all the settings to the settings I used (the ones that you suggested should be used) I got back to my original conclusion and repeating with TIFF just confirmed it
Hence I believe that DxO have restored the settings to their original intent, i.e. exports from PL6 should now match those from PL5 with respect to the ‘Neutral’ options with the ‘Legacy’ gamut scheme set in PL6.
I have never said anything about “fudge”. In fact, I have just completed additional tests and agree completely with your principal conclusion that the 4 neutral color renderings are directly comparable in PL5.70 vs. PL6.20. What I have objected to is your using @eriepa quotes from platypus as if they are my personal quotes. You have done this again in your latest post. Maybe just a misunderstanding about who said what. All is OK. Hats off to DxO for normalizing the neutral color rendering between PL5 and PL6. Thank-you for providing additional information regarding your testing.
@eriepa Thank you for confirming my conclusions, it took me longer than it should to get to them because of simple errors by me along the way!
I am sorry I saw and used you quoting @platypus as if they were your quotes and that your protestations fell on deaf ears.
I am used to those quotes looking like this (above and below) and failed to look properly at the first line of your post (and failed to recognise those lines as coming from the first post) as a consequence I created even more confusion between us which was entirely unnecessary and counter productive.
Once again I apologise, I must look harder and see and understand exactly what is there to avoid confusion (my confusion).
In the meantime there are differences between PL5.5.0 and PL6.2.0, i.e. the differences I found between PL5.5.0 and PL5.7.0 and reported in a post earlier in this topic.
These are tiny differences between the images @Barbara-S but it would be good to know what they are?
While I am glad that this item is fixed it would also have been good to have this item listed amongst the fixes in the PL6.2.0 to alert users, enabling them to verify that it is fixed in their case and to return to using those features.
Had a look at the settings again in DPL versions 5.5, 5.6 , 5.7 and all available versions 6 on Mac.
The verification shows that, no matter what the settings’ name is, it’s reported to be the “Fidelity” setting in the .dop sidecars. Assuming that an “internal” setting stays the same, all versions should render the image in the same way…but some slight changes persist between DPL versions 5 and 6, and we’ve already mentioned that in other posts too.
For the tests, I only used colour rendering in DPL5 and saved the .dop file, which was then read by the next version. All other settings (tools) were switched off. As reported by others, the setting in question was not renamed in DPL for Windows.
Thank-you, and I can see how I may have contributed to the misunderstanding of who was saying what when. No problem, really. Am glad we got to quick agreement regarding the neutral color rendering (Classic aRGB) in PL5.7 and PL6.2, the latest versions. At least in Win PCs, that is. Can’t speak to what may going on with the Mac versions though (see recent platypus post). Cheers!
