I have no idea what you want to say with this.
that shall be a subject for a different topic … you state FOR THE RECORD that with with “Export as DNG (all corrections applied )” you do not have white balance baked in demosaicked data in linear DNG… do I understand your statement correctly ?
PS: so if I will change WB only ( like 2 very different WBs and still the same demosaicked data in RGB channels ) in DxO PL6 I will be getting the same demosaicked data in linear DNG 
- because we can illustrate if this is the case -OR- not very easily … shall we go ?
No, you don’t. I say (in other words) that, no matter if I export with “all” or “oco”, metadata looks the same.
1 Like
In case you keep forgetting, you are writing in a public forum and you are expected to ‘act’ accordingly.
If you cant’t, please go elsewhere !
2 Likes
ok, temporary peace then
have something to contribute on the subject of the bug you ?
Sorry to see that you think of these discussions of something else than civil exchange.
Btw: It would be quite polite to unhide your profile.
1 Like
that is why I am using the word “baked”… metadata the same - but demosaicked data is DIFFERENT… so the end result when using in the 3rd party raw converter is not the same ( 1 + 1 = 2, but 10 + 1 = 11 )
@platypus
The pixel values in the demosaiced dng have already a lower value, on the export. The metadata is ok.
I don’t know where the gamma correction comes in. It’s a non-linear edit that should be visible in the dng compared with the ‘normal’ image.
George
with people who do understand the problem and the tools and how to apply them , but when “George” barge in - that is an insult to any “common sense” …
things like this is what is insulting :
“Who is doing the scaling and where is that based on?”
“I downloaded an app “raw image viewer”, it only shows the raw image, no editing posible. It shows the same image as in LR, well I think.”
I mean seriously ?!
that’s it… here he comes again… with “gamma correction” my $$$$$$
If one applies different edits to an image, exported image data will be different.
Starting with RAW, WB is always applied in order to get an RGB image…and measuring the WB in RAW might be difficult, but maybe RawDigger (et al.) can do it.
of course not - that depends on the software that you are using… Iridient raw converter or RPP can let you apply UniWB = that is exactly NO WB applied
also you are not precise what is RGB image ? is it demosaicked data ? then of course NOT … WB ( if applied ) can be applied either before or after demosaick - it all depends on the software ( RPP for example does apply before demosaick, Adobe ACR/LR do apply after demosaick, etc, etc )… is it color transformed data ? that is “RGB” are coordinates in some formal RGB color space ( as opposed to various non-RGB color spaces for example ) ? again - WB application is not necessary ( the mere fact that it is customary does not mean WB is always applied )
Apply UniWB = exactly NO WB? sounds like an oxymoron.
Anyways, I’m off for today.
multiply raw channels by [1,1,1] = means leave them unchanged ( there is no material difference, as if you completely by pass the code that applies WB ) … you disappointed me 
PS: by UniWB I do not mean what you set in camera in this case - I mean that those converters let you directly set WB as multipliers, so you dial in exactly 1.0, 1.0, 1.0 ( and in case or RPP there is even a 4th parameter to account for green channels disbalance = screenshot https://i.ytimg.com/vi/9wEcb3lXDRw/maxresdefault.jpg )
we do not need to measure it - we can simply illustrate that changing WB ( WB and WB only - leaving everything else the same ) in DxO PL when you export with “Export as DNG (all corrections applied )” results in different image data … that is alone a good reason why some users want only “Export as DNG (Denoise and Optical Corrections only)” - not to mention the rest of the stuff that gets baked in
saw the light in the tunnel you now
I think you DO, but if I am wrong then / in the spirit of civilized conversation / it means that you saw magenta tint in the skies …
PS: what I do not understand still is - you do not really need to see it… you just need ( and I posted about this before ) to do a simple math : equal RGB values in demosaicked linear DNG, you pull brightness and you do apply WB … “daylight-ish” WB means “Blue” and “Red” will be multiplied more than “Green” channel … this will result in magenta tint - UNLESS the raw converter does detect clipping and use a technique to proportionally block WB application in the affected areas … you do not need to see it, you can try to understand the principle … because 2x2 = 4 , you do not need to use fingers or stones or whatever past a certain age 
PS2: you can ask why then it is not immediately magenta w/o pulling brighness down - that is because a normal raw converter when it deals with a non clipped data ( and what DxO PL6 produces is a non clipped data for ACR / LR ) it ALSO checks if there is a clipping AFTER WB applied… so if you do not pull brightness low enough then ACR/LR detect clipping after WB application and counteracts it to a great extent ( plus it is all bright enough to start with with all stuff like hidden expocorrections, tone curves, etc applied ) … but when you pull down at some point ACR/LR see unclipped data not being clipped after WB application and stops counteracting that… that is why magenta tint is appearing when you pull low enough … in you just go by numbers in that TL2 exampe in ACR magenta tint ( miniscule difference ) starts to appear around -0.7 “stops” if not earlier… by numbers granted, visually you still see all white … on my monitor w/ my eyesight I start to see magenta tint @ around -2.5 “stops” pull ( but it is technically there well before that )
I have just checked the metadata in an exported DNG with all corrections, but with only optical corrections plus a tone curve, and everything to do with colour temperature, white levels, etc was identical.
yes, found that too. Nevertheless, DPL’s DNG creates a “coloured white” from “neutral white” as shown in the RawDigger screens wayyyy up. I did not expect such an effect with a “no correction” export.
But isn’t that “coloured white” just an indicator for over-exposure? Like the indicators in PL?