before starting processing

[sebbe]

I wanted to summarize all the stuff around processing. But as I am just a hobby photographer I wanted to challenge it with the forum here.
The idea behind it: See everything that has an influence on colours on one side. It should be an overview to do further explanations. target group: people with no or very few experiences in colour management.

Did I miss something? Is something wrong?
(If you see a spelling I'm happy if you correct me. English is not my mother tongue.)

Thanks in advance for your help.
Seb

Edit: First version removed. See reply #12 for version 2.

[Hening Bettermann]

A nice idea to visualise an overview over the work flow in one glance!

[Eric Myrvaagnes]

It looks like a very good overview, making it easy to see the links in the chain that one might be tempted to forget.

[daicehawk]

It all boils down to how close the color filters of the camera are to full conformity to the Luther-Ives condition and the quality of the light you shoot under. Monitor calubration and profiling are a must of course, custom printer profile probably not if you use original ink and media.

[sebbe]

Thank you all for the nice words. Each of you pointed exactly on one part of the content I wanted to have in the fact sheet. It looks like I made a good summary to support my purpose.

I will finalize it and share it here. Maybe others like to use it too.

Of course, I'm happy with further critics. Or ideas for an improvement.

[elliot_n]

There's a typo. You're missing an 'e' in 'Colour Temperatur', bottom left.

[Aram Hăvărneanu]

Also Tungsten.

[Eric Myrvaagnes]

Also Tungsten.

And "Fluorescend" should be "Fluorescent,"
just to be picky.

[smthopr]

It should be an overview to do further explanations. target group: people with no or very few experiences in colour management.

I'm afraid that people with little experience with color management will not understand any of your chart. It's a good chart for someone that needs to record a scene as accurately as possible for some scientific research.  But photographers don't need or want scientific accuracy.  For us it is more simple:  Calibrated display and calibrated printer = predictable results.  If one shoots RAW, and adjusts the processing controls by eye, one gets where they want to go.  No usual need to profile a camera itself.  Accurate color is not usually the most pleasing.

[sebbe]

I'm afraid that people with little experience with color management will not understand any of your chart. It's a good chart for someone that needs to record a scene as accurately as possible for some scientific research.  But photographers don't need or want scientific accuracy.  For us it is more simple:  Calibrated display and calibrated printer = predictable results.  If one shoots RAW, and adjusts the processing controls by eye, one gets where they want to go.  No usual need to profile a camera itself.  Accurate color is not usually the most pleasing.

So true. Thanks for this. It was not the intention that everything has to be done. That's why I will remove the processing sentence for my final version. Also about the profiling, Paper profiles and camera profiles are available and therefore there is no need to get into profiling there. I need to find a better title than 'create profile'.
Thanks to the others for the spell checking.

[GWGill]

For us it is more simple:  Calibrated display and calibrated printer = predictable results.  If one shoots RAW, and adjusts the processing controls by eye, one gets where they want to go.  No usual need to profile a camera itself.  Accurate color is not usually the most pleasing.

Except that this is doing things the hard way. Your "adjustment by eye" needs to simultaneously compensate for the characteristics of the camera, and the subjective adjustment you wish to achieve.

If instead you use a camera profile to transform colorimetrically into a standard colorspace, then your by eye adjustment is independent of the camera color behavior.
 

[sebbe]

Except that this is doing things the hard way. Your "adjustment by eye" needs to simultaneously compensate for the characteristics of the camera, and the subjective adjustment you wish to achieve.

If instead you use a camera profile to transform colorimetrically into a standard colorspace, then your by eye adjustment is independent of the camera color behavior.

Because profiling needs a quite long learning process at the beginning it's better to attribute the colour issues externally (blame a brand) or see them at least as "negligible". Or both.

Reminds me to add other RAW-converters in the sheet. Or at least make it visible that there are others.

[sebbe]

Here is the updated (and maybe final) version:
- added two visualisations for sensor readout
- added "monitor, workspace" to "choose"
- added two workflow-arrows: "capturing - processing - presenting" and "colour space / dynamic range"
- removed "processing" - sentence
- changed text in parentheses by "shooting light (colour temperature)" to "shooting light (control and measure)"
- other minor text changes
- spell checking
- layout

The addition of the sensor readouts does not add much but it may explain how RAW data occur.
I think it's on the limit of different information now.

Any additions, critics from your side?

[smthopr]

Except that this is doing things the hard way. Your "adjustment by eye" needs to simultaneously compensate for the characteristics of the camera, and the subjective adjustment you wish to achieve.

If instead you use a camera profile to transform colorimetrically into a standard colorspace, then your by eye adjustment is independent of the camera color behavior.

I think most photographers will find an accurate reproduction (even as a starting point for adjustment) doesn't look "photographic".  And I'm not convinced that the controls in photoshop offer any intuitive way to re-matrix the transform.  To do this, one would need to adjust hue and saturation to the image in LAB space or YUV space.  Each control effects another and it's not a straight forward way to color correct an image.  The camera manufacturers understand this and have developed their own matrix transforms to get to what they think is a good starting point for "Photographic" reproduction.  And, I think that for most photographers, this is the best place to start.  And of course, Adobe has their own versions as well in their RAW converter.

I guess what I'm saying is that, for most photographers, camera profiling is a rabbit hole best left for those who have found they have a special need for it.

I also say this as I've been doing some motion picture color correction and I've been exploring the "camera manufacturer's" standard matrix transforms, vs creating my own.  To conduct my experiments, I've got a shot of a standardized DSC color chart that I'm working on, but, to really see what I'm doing, I need to view the image on a waveform and vectorscope.  These tools are not available in photoshop or lightroom, and working on normal photograph with only a histogram, would be like throwing darts in the dark.

[sebbe]

I think most photographers will find an accurate reproduction (even as a starting point for adjustment) doesn't look "photographic".  And I'm not convinced that the controls in photoshop offer any intuitive way to re-matrix the transform.  To do this, one would need to adjust hue and saturation to the image in LAB space or YUV space.  Each control effects another and it's not a straight forward way to color correct an image.  The camera manufacturers understand this and have developed their own matrix transforms to get to what they think is a good starting point for "Photographic" reproduction.  And, I think that for most photographers, this is the best place to start.  And of course, Adobe has their own versions as well in their RAW converter.

I guess what I'm saying is that, for most photographers, camera profiling is a rabbit hole best left for those who have found they have a special need for it.

I also say this as I've been doing some motion picture color correction and I've been exploring the "camera manufacturer's" standard matrix transforms, vs creating my own.  To conduct my experiments, I've got a shot of a standardized DSC color chart that I'm working on, but, to really see what I'm doing, I need to view the image on a waveform and vectorscope.  These tools are not available in photoshop or lightroom, and working on normal photograph with only a histogram, would be like throwing darts in the dark.

A constant starting point for adjustment is not about perfect colour reproduction. The idea of a profile is to get the colours exactly to the point you want to have them. But to get a good constant starting point your profiling has to be close to perfect and repeatable. Thus it can be said that your first goal within profiling is to get perfect colour reproduction. If you achieved that you customize the "perfect" colour space to your desired standard colour space (for processing). Of course you can have as much standard colour spaces as you like. E.g. I have profiles that transfer RAW-Data of RGB-sensor-cameras into a B/W-colour space or others that adds an S-shaped tone curve with a slight desaturation for darker mid-tones.

The clue is that profiles for different hardware with the same standard colour space will transfer the different RAW-data into identical pictures colourwise.

[smthopr]

A constant starting point for adjustment is not about perfect colour reproduction. The idea of a profile is to get the colours exactly to the point you want to have them. But to get a good constant starting point your profiling has to be close to perfect and repeatable. Thus it can be said that your first goal within profiling is to get perfect colour reproduction. If you achieved that you customize the "perfect" colour space to your desired standard colour space (for processing). Of course you can have as much standard colour spaces as you like. E.g. I have profiles that transfer RAW-Data of RGB-sensor-cameras into a B/W-colour space or others that adds an S-shaped tone curve with a slight desaturation for darker mid-tones.

The clue is that profiles for different hardware with the same standard colour space will transfer the different RAW-data into identical pictures colourwise.

My point is that there are "look" elements contained in the manufacturer's transforms that one may well undo by profiling to a color chart.  And these transforms can not be easily replicated using normal color correction controls.  These transforms are certainly not limited to output curves that one would normally apply in lightroom or photoshop, but, they do contain built-in curves as well, just to make things complicated.  In photoshop, when we perform a color space transform, the output curves are not included in the transform (aside from a basic gamma 1.8 to 2.2 for example).  This is not the case with original RAW to working space conversion in the RAW processing software where the color space transform (which is not "accurate" and the output curves are mixed together into a "look".)  When you profile to a color chart, you'll defeat the "look" part.  Maybe you'll like it better, that's entirely possible, but it will not be the same look.

[sebbe]

My point is that there are "look" elements contained in the manufacturer's transforms that one may well undo by profiling to a color chart.  And these transforms can not be easily replicated using normal color correction controls.  These transforms are certainly not limited to output curves that one would normally apply in lightroom or photoshop, but, they do contain built-in curves as well, just to make things complicated.  In photoshop, when we perform a color space transform, the output curves are not included in the transform (aside from a basic gamma 1.8 to 2.2 for example).  This is not the case with original RAW to working space conversion in the RAW processing software where the color space transform (which is not "accurate" and the output curves are mixed together into a "look".)  When you profile to a color chart, you'll defeat the "look" part.  Maybe you'll like it better, that's entirely possible, but it will not be the same look.

Maybe it's my English (I'm German speaking): It looks like you didn't understand what I was explaining in my last post or I do not understand what your goal is.
Profiling can have a look-component too. Maybe it's better understandable with an additional page about that.

To all: If there is something wrong here. Please criticize me. I'm only profiling one camera. Therefore I never did comparable profiles over different cameras or brands. But if I understand it correctly it should be possible.
As far as it comes to the default profiles within RAW-converters I think they take colour behaviours of OOC-JPGs into account and that's why they are not identical but have a branded look.

[JRSmit]

Is the "look" element not just the choice of the maker?
Therefore a subjective thing. Whereas profiling from a given chart like xrite Color checker is more objective.

[sebbe]

Is the "look" element not just the choice of the maker?
Therefore a subjective thing. Whereas profiling from a given chart like xrite Color checker is more objective.

Common profiling is always with a target like the Color Checker Passport. There are other possibilities like measuring the CFA of a camera but that's not in the budget of most of us.
The profiling software reads the shot of the target and calculates the optimal (lowest avg. dE) profile out of it. As this "objective" profile is not very pleasing or need more adjustments within the RAW-processor some profiling software gives you the chance to customize the result on top before the profile is made. Therefore you can add a personal look. But it's up to you to use the profile without any additional customizing or add some changes (tone curve, colour shifts) on top to give the profile a personal look.