The K3 III in the comparison tool

flektogon wrote:

Yes Jeremie, I agree with what you wrote:

To me it's very relevant what parameters a raw converter applies by default, exactly because as you said it's the reference you start from. If it's f**** because the raw converter applied crazy sharpening generating awful artifacts, then it's a problem and exactly what you want to avoid when you shoot RAW.

But at first let's assume that no photo editor opens a raw file such a way that it is already damaged.

So, it opens the raw file and will show you a picture. And yes, this is the starting point of your process. Once you have a likeable result, you will save it. If you save it as a jpeg file (what I am doing), all those parameter values you used (like the sharpening amount) are irrelevant, they won't be saved as a part of the jpeg data! Only if you save your processed image as another (or replacing original) raw file, those processing parameters will be saved, as the original data taken from the sensor won't be touched at all.

And that's it. So, the conclusion of this entire debate is, that it is irrelevant how your (renowned) editor opens the raw file, you should be able to process it to whatever result you want to get!

Ok for sure, but there are 2 contexts:

I'm a user / photograph

It's up to me to bother or not with the defaults from the raw converter. How I ended up with a jpeg and how it was done is my problem. If I don't care or see the sharpening, tone mapping etc, it's again my problem.

I can perfectly ignore all this and use tool X because for me, "tool X gives me better results", which might be just because I'm used to its default settings. I can also perfectly use sooc jpegs. I can do everything I want, I'm the God of my photos

The DPR comparison tool (or other pixel peeping / scientific use cases)

(which was the initial topic from the OP)

Apparently, they use ACR, that applies its own defaults, which may vary depending on the camera and/or brand and/or setting in-camera.

Apparently those settings (used to generate what you see in the comparison tool then, unless you download and open the RAW file), are stored back in the RAW file.

Side note: doing that is in no way "destructive" regarding the original data in the RAW file. A RAW file is just a container in which data can be embedded, along with metadata. Here they apparently just embed additional XMP data inside the RAW file, instead of storing it in a side file .XMP. So the RAW file is different, but the original captured data is the same.

The big difference is that the whole point of this tool is for pixel peepers. If you don't care about the detailed IQ/appearance of images taken by different cameras, you probably never consult this tool.

So my own conclusion would be:

- it's not a huge problem that different settings might have been applied, considering I can download the RAW files and compare myself on my computer if I want

- still, if confirmed it's a pity because it's much easier to use the tool online to perform these comparisons as it is well designed (and you don't have to download the files etc)

- does it matters for the amateurs among us ? To each his own.

- that anyone can take the most beautiful photo using an old disposable film camera and sending the film to a generic lab (for example) is true but does not mean that we can't pixel peep the pixel peeping tool if we like spending our precious time on useless debates of the internet

Ian Stuart Forsyth wrote:

flektogon wrote:

Yes Jeremie, I agree with what you wrote:

To me it's very relevant what parameters a raw converter applies by default, exactly because as you said it's the reference you start from. If it's f**** because the raw converter applied crazy sharpening generating awful artifacts, then it's a problem and exactly what you want to avoid when you shoot RAW.

Most raw converters do not apply a crazy amount of sharpening.

Of course, but my comment was in regards of the topic, which was pixel peeping images displayed by the DPR comparator.

Makes sense to know if there was +40 sharpening or 0, I believe, in this context.

You may have one converter that is set to 40 while the other is set to 10 and they sharpen the image similarly. Just as with a stereo from one unit it may be set to 10 and another set to 40 the output volume is similar

But at first let's assume that no photo editor opens a raw file such a way that it is already damaged

You are not destroying or damaging you are processing data to an output image

If you look at a raw file there is not much to look at what you see int the image using raw converter you are looking at only the processing done to that image

So, it opens the raw file and will show you a picture. And yes, this is the starting point of your process. Once you have a likeable result, you will save it. If you save it as a jpeg file (what I am doing), all those parameter values you used (like the sharpening amount) are irrelevant, they won't be saved as a part of the jpeg data!

Why would you want to save that data? you are processing that data into an image and that jpeg is the processed data

I can think of reasons why I would want that.

If I have the jpeg and lost the XMP, and if the jpeg looks nice, it could be nice for me to be able to check how it was generated.

It's like a build process in programming, or like any scientific study, being able to reproduce a result is important.

Of course it's highly theoritical, in practice I agree with you almost all the time we don't need this and there are other ways to obtain what I described.

Jpeg usually is the final step in the processing, this is saving how you want the image to appear. does it matter what processing is done to a film negative to achieve a working print? no.

many raw converters allow the user to save the processing settings done while retaining the raw data.

Only if you save your processed image as another (or replacing original) raw file, those processing parameters will be saved, as the original data taken from the sensor won't be touched at all.

very few converters allow the user to replace the raw data, and again why would you want too, what is fascination with what setting are being done to the raw file the need to store it in the jpeg file ?

You would not want to make changes to a film negative to alter it only to save it as another negative. With digital raw it is no different we have a negative we apply processing to the negative, we record what those processing setting ( with film we would write down what was done and store it with the negative or use a default without heavy Developmemt ) with digital we store that data using a sidecar or even exporting those setting into a note doc if we every want to go back process the raw data again

And that's it. So, the conclusion of this entire debate is, that it is irrelevant how your (renowned) editor opens the raw file, you should be able to process it to whatever result you want to get!

You can you and that is the whole reason for raw converters, you are making a mountain out of a molehill.

Jpeg is your final image as to how you want it processed

True, but speaking for myself I seldom have just 1 final jpeg per RAW. Usually I have several depending on target media, and then I also have several with different PP across the time (as tools evolve), and also several different depending on my taste for the PP and different trials.

So the ability to trace back a jpeg to the RAW+XMP used, I find it interesting in fact ! so thanks to this discussion I will look if with DT I can add some metadata in the jpegs during export, thanks everyone !

flektogon wrote:

Hello KPM2,

Well, your thread actually revealed one, maybe a really problematic side of the raw file processing. As you found, two different photo editors opened the same raw file applying different parameter settings. Like one used the sharpening amount of 45, while the other 0. And, as few responders replied, maybe both editors applied the same sharpening amount, just they use different "measures". And this might be a source of worryings. Imagine, you want to sharpen your image a little bit more, so you apply a value of 10. When you (re)save such processed raw file (using the first editor), that sharpening parameter will be saved as a value of 55. And now, when you open it with the other editor, you will end up with over-sharpen image!

These parameters, or metadata, are stored in different groups in the file.  The Pentax parameters for sharpening, contrast, etc are stored in a Pentax group.  When you set parameters in LR they are stored in an Adobe group (perhaps in the file, perhaps in a seperate XML file, perhaps in the LR catalog).

I am not any expert in the Adobe raw data spec, but I thought that every photo editor would use exactly the same measures for the same image parameters.

Do not assume that they are using the same algorithm to o sharpening - there are many.  So the parameters are not necessarily interchangeable.

Of course, these problems will become evident only if you open/process raw file (actually you don't even need to modify any image parameter!) in one editor and save it as the original raw file replacement.

KPM2 wrote:

Hello JeremieB

JeremieB wrote:

With exiftool -xmp -b, I see the same things as you on the Z9 and K3III RAW files from the DPR studio scene.

Darktable can't open the .NEF of Z9 at all (not supported yet).

RawTherapee can but I couldn't see anything related to those ACR settings.

But if I run same exiftool command on one of my DNG files from K3III, there is nothing output - so I think it means the XMP information from ACR really is embedded in the RAW files from the studio scene.

Yes, in the DNG files we can download from DP for the Pentax cameras we can see the embedded XMP files. We even can see what programs DP do use and how often the file was opened.

But for the Z9, we see not the XPM from DP, we see the XMP file made from our PC's and on this way, so long we use not our own settings, the Z9 XMP show default setting.

The problem is now:

I know that DP use a sharpness of 0 for the Pentax cameras, do to the embedded XMP files in the DNG file, but what I do not know:

1.) is this ACR's default ? I got answers: yes it use 0 - and in an another answer: it's 25 like with your LR.

2.) do use DP only default settings, or do set DP for all camera brands the sharpness to 0 for a fair compare, independently what default setting the ACR would use for the cameras.

3.) same for the color noise: do set DP this setting to 0 (like it do for the Pentax cameras) for all cameras for a fair compare, or use it default settings (and here, for the Z9 ACR would use a little color noise NR as a default - as we could see in the XPM of our PC)

I don't have the answers for these questions, but even with this color NR I think there's a lot of color noise in Z9 files at high ISO ...

At last: LR is based on ACR (how much I do not know), therefore I think it is for me easier to compare a LR and ACR JPG of this downloaded files. For example the first picture, with the ACR XMP from DP the picture in the comparison tool looks very similar to that, I made with my LR with their ACR XMP settings.

Yes but ideally you would have to use the same version of ACR that was used by DPR for each RAW.

best regards. KPM2