Skip to main content
Adobe Photoshop Family

300 Messages

 • 

5.8K Points

Sat, Jan 18, 2020 1:28 AM

1

More flexible layer-like approach to profiles is needed in Lightroom

When Lightroom was first released in 2007, the philosophy behind the Develop module's adjustment panels was that you made your adjustments starting at the top panel (Basic) and worked your way down. From Lightroom 3, in the last panel, Camera Calibration, you find the camera profile selection and in those early days, you had a choice of either Adobe Standard (the default) or a camera profile to suit the camera you were using, giving your raw images an initial similar look to that produced by the camera's selected JPEG style.

The camera profile was considered the starting point for your raw processing. Indeed, raw processing must have a profile, hence the Adobe Standard as the default. In Martin Evening's The Adobe Photoshop Lightroom 4 Book, he writes

"The Camera Calibration panel (Figure 4.72) allows you to select the most appropriate camera profile to use as a starting point for subsequent Develop module adjustments."

In my understanding, the camera profile is far from a staring point since it is applied last in the Develop adjustment's pipeline. We the users have absolutely no control over this. It seems to me that the camera profile is really the ending point, not the starting point.

This didn't matter very much until Lr Classic 7.3 came along in 2018, where the camera profiles changed radically to become far more powerful, not just emulating the camera JPEG styles, but providing artistic profiles too. What's more, users can now easily produce their own profiles in ACR and use them in Lr. The location of Profiles was also changed and are now found at the top of the Basic panel, an appropriate position if a profile is still considered a starting point.

However, a selected profile is still applied last and for some this is still suitable, but not for all. Further more, only one profile has any effect on the image, the last one selected in the history.

Now in 2020, it is time for profiles to be made more flexible in their application. Being able to make profiles yourself provides endless possibilities, but these can't be fully realised due to the current restrictive application of profiles.

I would like to see profiles changed so that the user has the option of allowing a profile to be applied anywhere the processing pipeline, not just the end. Profiles would be far more useful if users could apply them as layer-like adjustments. Local adjustments are layer-like and multiple local adjustments can be 'layered'. A similar layer-like adjustment that applied a profile would be very useful. Alternatively, allow a profile to be applied in a local adjustment.

A major reason for me suggesting this approach to profiles is my desire to see Lr have a Negative function. As this feature has been requested so many times for the last 8 years and totally ignored by the development team, profiles provide an alternative approach. I have created profiles to invert raw colour negatives and use them to process 'DSLR scanned' flim negatives in Lr with some degree of success. However, the fact that the profile is applied last means all the Lr Develop controls make their adjustments based on the negative image data, not the positive I see on the screen. It would solve everything if I could choose to have my negative profile applied at the very start, providing a true starting point so that all the Develop tools worked as if the image were a positive.

Hopeful.


Responses

1.6K Messages

 • 

20.1K Points

9 months ago

What makes you believe: "In my understanding, the camera profile is far from a staring point since it is applied last in the Develop adjustment's pipeline. " and it matters?

Author “Color Management for Photographers"

Champion

 • 

5.8K Messages

 • 

102.5K Points

9 months ago

Other than your negative inversion case (which I completely agree with), how do you imagine those changes would help anything? What use cases do you imagine?

165 Messages

 • 

2.9K Points

9 months ago

"In my understanding, the camera profile is far from a [starting] point since it is applied last in the Develop adjustment's pipeline. We the users have absolutely no control over this. It seems to me that the camera profile is really the ending point, not the starting point."

This bit confuses me, and it appears to be the foundation for your request. What do you mean by the camera profile is applied last? As I understand it, camera profiles are applied at the start (i.e. Adobe Standard by default) and users can choose to adjust them when they wish during their Develop processing but ideally at the start because they're meant to be the baseline render of colours before you adjust them.

For e.g., when I shoot a session with a ColorChecker Passport, my first step after culling is to create the customised color profile for that batch of photos, restart LR so it can 'import' or add it to the list of camera profiles, and then apply it to the batch. Only then do I make any other Develop changes. I also ensure I don't use any presets that apply a profile (or modify any that I like to use to remove the profile setting).

If I make other Develop adjustments and then change the profile, chances are I'll need to redo some of the color adjustments.

So, when you say profiles are applied last, do you really mean only during processing that leads to the development of your negative versions (i.e. you Develop your photos to their final tone & colours and then use your custom-built profiles to produce the inverted negative versions of them)?..

Champion

 • 

2.2K Messages

 • 

37K Points

9 months ago

I have created profiles to invert raw colour negatives and use them to process 'DSLR scanned' flim negatives in Lr with some degree of success. However, the fact that the profile is applied last means all the Lr Develop controls make their adjustments based on the negative image data, not the positive I see on the screen.
What's needed is a separate raw data invert function that is applied before any other controls during demosaicing. This has already been requested at the below post with no status to date. For all other raw editing purposes the camera profile is best applied last to insure the full dynamic range of the image data is available to the Tone controls for highlight and shadow recovery.

https://feedback.photoshop.com/photoshop_family/topics/inverting_in_camera_b_w_negative_scans_to_pos...



474 Messages

 • 

12.2K Points

9 months ago

The whole premise that LR applies changes based on panel position (or anything else) is invalid.  As far as I understand the only adjustments that have any sort of "order" is the use of the spot removal tool where you use the destination area of one removal operation as the source for another.  As I recall there is one other but I can't remember what it is. 

In all other cases the "adjustments" become mathematical formula's for the value of any particular pixel and the order applied or the order that panels are shown is totally irrelevant.   For example, lets say you added 1 stop of brightness to the entire image.  Then in the gradient tool a particular pixel had its highlights reduced by 0.5 stop and since that pixel was originally real bright it got the whole 0.5 stop reduction.   So the adjustment of the luminosity of that pixel is original value +1.0 stop -0.5 stop for a net of +0.5 stop.   it doesn't mater if it was done the other way around with original value -0.5 stops +1.0 stop the answer is still a net of +0.5.  LR and ACR does not apply one adjustment, then another then another.  Each pixel gets a mathematical formula that encompasses all the adjustments that apply to that pixel and the net of that formula is applied.

Dan

1.6K Messages

 • 

20.1K Points

9 months ago

Correct: recommend edit order is not the same as processing order.

Author “Color Management for Photographers"

Champion

 • 

2.2K Messages

 • 

37K Points

9 months ago

The whole premise that LR applies changes based on panel position (or anything else) is invalid. For example, lets say you added 1 stop of brightness to the entire image.  Then in the gradient tool a particular pixel had its highlights reduced by 0.5 stop and since that pixel was originally real bright it got the whole 0.5 stop reduction.   So the adjustment of the luminosity of that pixel is original value +1.0 stop -0.5 stop for a net of +0.5 stop. 
That's not entirely true. Try applying a +4.0 Exposure Gradient filter to an entire image that has bright highlights and then apply -4.0 Exposure with the Basic panel Tone control. If what you're saying is true it should look identical, but it doesn't due to non-linear summation of the Local and Global controls. This used to be much worse in LR 4 with the Local controls applied first with no summation, which often produced highlight and shadow clipping that was unrecoverable with the Basic panel Global Tone controls. This was partially "fixed" in LR 5 or 6, but it's still not as you describe above.


In the case of the Tone control it is clearly applied last and overrides all of the other controls


474 Messages

 • 

12.2K Points

9 months ago

I'm not seeing what you are seeing.  I took an image and made two VC's from it.  I then "reset" both VC's to their original out of camera state.  In one, I first adjusted the overall exposure to +4 followed by making a gradient which covered the entire image (drew gradient outside the image) and on the gradient set the exposure to -2.  On the 2nd VC I made the same changes but in the opposite order.  I then compared the two images and they are identical.  No change in the histogram when flipping back and forth between the images and no change in the look between the two.  



Can you please explain your tone Curve example?  i don't think I'm understanding your point.  You show the basic panel Whites slider at 100 and the blacks slider at -100 and the tone curve shows the darks starting at around 25% and the brights ending at around 75% but these are two completely different adjustment from the whites and blacks sliders in the basic panel and even though they both affect how the brights and darks look, they are not interchangeable.  Are you saying the result is different if you do the basic panel first and tone panel 2nd vs the other way around?  My experiment in doing the Basic Panel White/Blacks sliders first then the tone curve 2nd gives the same results as doing the tone curve first and the Basic panel 2nd.  Look of the image is the same,  histogram is the same



Tone curve 1st then Basic panel 2nd


Basic Panel 1st then Tone Curve 2nd

Champion

 • 

2.2K Messages

 • 

37K Points

I'm not seeing what you are seeing.
Dan I was responding to this statement, which as demonstrated is incorrect. However, I do agree that the order of applying the Global and Local control has no impact on the final  rendering.
The whole premise that LR applies changes based on panel position (or anything else) is invalid. For example, lets say you added 1 stop of brightness to the entire image.  Then in the gradient tool a particular pixel had its highlights reduced by 0.5 stop and since that pixel was originally real bright it got the whole 0.5 stop reduction.   So the adjustment of the luminosity of that pixel is original value +1.0 stop -0.5 stop for a net of +0.5 stop. 

300 Messages

 • 

5.8K Points

9 months ago

Dan, you are not doing what Todd explained with the global exposure and full extent gradient exposure to negate the global adjustment.

What Todd wrote was to apply the local gradient with +4 exposure followed by the basic adjustment of -4, thereby resulting in a net 0 exposure adjustment. The problem is, the result isn't the same as the image you started with. Example, the Reference on the left is the original raw image, the Active on the right is a Virtual Copy with local full image gradient exposure +4 followed by global exposure -4:



The RGB values displayed in the histogram panel for the current sample point show quite a difference from the original values.

If you do the local adjustment after the global one, then the result is pretty much the same as what you started with. In the screen shot below, on the right is a Virtual Copy with +4 global exposure followed by a local full gradient exposure -4. Notice the RGB value are almost identical to the original image now.



Regarding the tone curve, Todd is simply saying that the tone curve adjustment is applied after the basic adjustments. In addition to that, the local adjustments are applied after all the global adjustments, including the tone curve. With Todd's 25/75 tone curve, add a full gradient and push the whites to 100 and blacks to -100. The histogram shows that tones push past the limits set by the tone curve.

Tony

474 Messages

 • 

12.2K Points

Tony, the point I was answering was the topic of "does the order matter", not the question of does the Exposure slider in the Basic panel work using the same mathematical algorithm as the basic slider in the gradient tool.  I'll give your experiments a try tomorrow or Monday.

Champion

 • 

2.2K Messages

 • 

37K Points

If you do the local adjustment after the global one, then the result is pretty much the same as what you started with. In the screen shot below, on the right is a Virtual Copy with +4 global exposure followed by a local full gradient exposure -4. Notice the RGB value are almost identical to the original image now.
Anthony these are not the same settings applied in different order. In my example +4 Local Exposure and -4 Global Exposure settings were applied and the order of application makes no difference. In your second example you are using -4 Local Exposure and +4 Global Exposure, which produces different results. As Robert Somrak points out this is because the global exposure is “image adaptive” and the local exposure is not. 

474 Messages

 • 

12.2K Points

Anthony - I'm on LR 8.3 so perhaps 9.1 is different but I can't replicate your results.  in images below, The left one had Gradient Exp set to +4 then Basic Exp set to -4.   In the right photo I did the Basic Exp to -4 then the gradient Exp to +4.  The two images are identical.  Flipping from one to the other the histogram doesn't change at all.  So, in my experiment the order of application did not matter.



I did the same experiment with with tone curve (25%-75%) followed by a full image gradient.  When I adjusted the Exposure with the gradient, order didn't matter.  However when I instead adjusted the whites (+50) and  Blacks (-50)  the other order did make a difference.  I really don't know what to make of that result.

Champion

 • 

2.2K Messages

 • 

37K Points

I did the same experiment with with tone curve (25%-75%) followed by a full image gradient.  When I adjusted the Exposure with the gradient, order didn't matter.  However when I instead adjusted the whites (+50) and  Blacks (-50)  the other order did make a difference.
Using LR 9.1 I see no difference in rendering due to order using a  Tone Curve with 0, 25.1 and 100, 74.9 settings and Graduated filter with +50 Whites and -50 Blacks. I tried both Adobe Standard and Adobe Color, opened the files as layers, Blend Mode Difference, and Levels applied. They are identical.

300 Messages

 • 

5.8K Points

You're right Todd, they aren't the same adjustments in a different order. My mistake and thanks for pointing it out.

It is still an interesting result that means care needs to be taken when trying to negate global exposure increases using local adjustments.

Tony

300 Messages

 • 

5.8K Points

Dan, regarding the Tone curve and gradient test, Just as Todd found with Lr 9.1, I too found no difference between the rendered images and changing the tone curve/graduated filter order.

Tony

300 Messages

 • 

5.8K Points

It is still an interesting result that means care needs to be taken when trying to negate global exposure increases using local adjustments.
I've messed up again Todd. I meant to say

... care needs to be taken when trying to negate local exposure increases using global exposure decreases.

1.6K Messages

 • 

20.1K Points

9 months ago

Tony, if you render out both to a TIFF in the same color space, maybe subtract the two in PS, what do differences result?

Author “Color Management for Photographers"

300 Messages

 • 

5.8K Points

Andrew, opening both as layers in Photoshop should be the similar to exporting as tiffs, so this is the result:



Tony

300 Messages

 • 

5.8K Points

The difference posted previously was between the virtual copy and the original raw image, with virtual copy having the local adjustment before the global. Here is the difference between the virtual copy and original with the local adjustment after the global (much less).



Tony

300 Messages

 • 

5.8K Points

9 months ago

Here is an example that might explain what I mean by a Profile being applied last. I have created two virtual copies from an original raw file. To one copy I applied the Adobe Standard profile and then used the White Balance Selector tool to set a Temp and Tint from a sample point in the image. The result is shown below:



I then applied the Artistic01 profile to the second virtual copy and set the white balance using the White Balance Selector tool to sample the same point in the image as used previously. The result is shown below, with the first virtual copy set as the Reference on the left and the second copy as Active on the right:



It is clear that the white balance in the copy with the Artistic01 profile is based on the colours before the profile has been applied, otherwise the Temp and Tint values would be quite different from what they are. The Temp and Tint values are exactly the same as those computed with the Adobe Standard profile applied, however, the RGB values in the histogram panel are nowhere near neutral in the Active image which has the Artistic01 profile applied.

This indicates to me that the profile, at least the Artistic01 profile, is applied last.

Why does it matter? That may depend on the profile, as it does with an inverted tone curve. Adobe has no idea what profiles users will create nor how they want to use them, so in my opinion, this restriction limits their use. There may be very good reasons to apply camera profiles and B&W profiles last, but what about other artistic profiles?

Tony

60 Messages

 • 

1.1K Points

This doesn't indicate the profile is applied "last", just that applying the white balance comes before applying the transformations in the profile.

1.6K Messages

 • 

20.1K Points

Yes, as DCP profiles are by design, WB agnostic and the user is supposed to be able to apply WB independent of said profiles. 
The only people who actually know the processing order of edits are deep inside Adobe. 

Author “Color Management for Photographers"

Champion

 • 

750 Messages

 • 

13.3K Points

9 months ago

Part of the reason the original image isn’t the same as image with the +4 global exposure and the ~4 local gradient exposure is because the global exposure is “image adaptive” and the local exposure is not.  

Champion

 • 

2.2K Messages

 • 

37K Points

Thanks Bob that explains the non-linear interaction between the Local and Global controls. Since the Globals are used to adjust the overall image making them image adaptive is helpful. However, the Locals are used to adjust specific areas so It makes sense to apply them linearly. ERGO–Never the twain shall meet!