Lightroom Masking

[David Eichler]

I realize there is already a thread on the new Lightroom masking feature. However, I want to ask about this feature in relation to Photoshop specifically. I have only just started using the new LR feature and might not yet realize its full potential. That said, since I already tend to process many of my photos with Photoshop, I wonder whether there are any specific reasons to prefer using the masking capabilities in LR that seem to overlap with what PS does, for images that need PS for other reasons. I routinely use LR to do the basic processing, and usually some local adjustments, before sending to PS for more advanced processing and retouching.

[MDL_SD]

I think that you can argue either way on this question.  However, to me it is desirable to do as much as possible in Lightroom since all changes are non-destructive in Lightroom and it is so easy to have multiple virtual copies to explore different directions with a single image.  Camera Raw inside Photoshop gives you all of the Lightroom masking so you could also work there.
Others certainly have their own views on this.  I like using Lightroom to catalog and keyword my photos even if I edit in both Lightroom and Photoshop.

[rdonson]

For me...

-1- ease of use
-2- ability to use the LrC tools
-3- speed, less time spent getting to a good state

I also use Luminar AI and the Nik Collection for the same reasons...

As also MDL_SD says.... the LrC catalog

I will use PS when all else fails...

[digitaldog]

Biggest advantage: all parametric edits, so actually truly non destructive edits that only take place when rendering the raw as new pixels. And as this parametric editor improves in image quality, so do all your previous parametric edits along with the new(er) processing.

[David Eichler]

Biggest advantage: all parametric edits, so actually truly non destructive edits that only take place when rendering the raw as new pixels. And as this parametric editor improves in image quality, so do all your previous parametric edits along with the new(er) processing.

Once we have ensured that all needed highlight and shadow detail is available with the RAW processing, we can retain "non-destructive" edits in Photoshop by editing in layers and leaving the file unflattened, and we still have the base edited RAW file in LR. How does editing in RAW versus PS improve image quality once all needed detail is brought out in the RAW processing? Yes, there is some editing I just find a bit more convenient to do in LR, versus PS, but aside from that?

[digitaldog]

In reality, there is no such thing as non-destructive editing in Photoshop. Even with adjustment layers, eventually the data is flattened and there is rounding errors thus data loss. In high debt the loss is virtually invisible but it isn’t non-destructive.

[Chris Kern]

Once we have ensured that all needed highlight and shadow detail is available with the RAW processing . . .

It occurs to me that color probably should be added to your list.  You still retain considerable tone control after your raw file has made the round-trip between Lightroom and Photoshop, but the demosaiced image has the colors "baked in."

I still rely on Photoshop for capabilities that currently work best with a rendered image—content-aware fill being the most obvious one—but it seems to me that the two products are diverging: with Lightroom increasingly becoming an all-in-one application for photography and Photoshop a powerful tool for less literal forms of graphic processing.

(As an aside, in a prior life I sometimes interacted with the in-house graphics shop at the government agency where I worked.  Only one member of the staff was a photographer, but they all had a mastery of Photoshop that I found absolutely stunning and which which I realized I would never be able to achieve.  I don't think most of them even thought of Photoshop as a photo editor: as far as they were concerned, it was a general-purpose tool for creating and manipulating images.)

[fdisilvestro]

I would say that it depend on what you want/need to do with the picture. This is mainly because there is no present way to export the masks to Photoshop (since they are stored as bitmaps, it should be easy to do for Adobe).

In LR, masks are stored in the file structure <Catalogue name>.lrcat-data, in ACR there is a new "sidecar" in addition to the regular XMP file, so take those in consideration for your backups.

If you can do all the edits in LR, then use the mask there, but if you need further options such as replace sky or select object to replace background, then do that in PS.

Something else to take in consideration, is that "Select Sky" is not based on the raw data but on the current rendering. If you don't believe it, then just move all Tone sliders in the Basic panel to the left and select sky.

[David Eichler]

It occurs to me that color probably should be added to your list.  You still retain considerable tone control after your raw file has made the round-trip between Lightroom and Photoshop, but the demosaiced image has the colors "baked in."

Please explain why I would have any less room to manipulate color with a 16-bit, Prophoto rgb, Tiff file than with a RAW file.

[David Eichler]

In reality, there is no such thing as non-destructive editing in Photoshop. Even with adjustment layers, eventually the data is flattened and there is rounding errors thus data loss. In high debt the loss is virtually invisible but it isn’t non-destructive.

Why is it inevitable that the data is flattened? You can keep an unflattened file indefinitely until you actually have to process it out for display or delivery to a client, and you can still retain the unflattened file after doing that. Plus you still have the Raw file to go along with it, in case you ever want to process from scratch again, say, with a different RAW processor.

[David Eichler]

Something else to take in consideration, is that "Select Sky" is not based on the raw data but on the current rendering. If you don't believe it, then just move all Tone sliders in the Basic panel to the left and select sky.

In any case, I do not find the Adobe sky selection process to be very useful. Maybe in the simplest situations, with very clear demarcation between the sky and the rest of the scene, it can work okay. But in my experience it selects too much or too little, and it is much easier to modify the mask in PS (and get a more precise mask to begin with). Plus PS has the Refine Edge feature, which LR doesn't have.

[digitaldog]

Please explain why I would have any less room to manipulate color with a 16-bit, Prophoto rgb, Tiff file than with a RAW file.

For the same reason you can't unbake a cake.
And depending on the sensor native “color space”, ProPhoto RGB a theoretical space based on a display, can make that unbaking even less possible.

[digitaldog]

Why is it inevitable that the data is flattened? You can keep an unflattened file indefinitely until you actually have to process it out for display or delivery to a client, and you can still retain the unflattened file after doing that. Plus you still have the Raw file to go along with it, in case you ever want to process from scratch again, say, with a different RAW processor.

You want to view this image outside of an Adobe product? You want to print it, get it on the web? The data has to be flattened.

[David Eichler]

You want to view this image outside of an Adobe product? You want to print it, get it on the web? The data has to be flattened.

Since the majority of the work I am doing requires using PS, one way or another, that flattening for output is a part of the process. So, it is really a moot point for someone like me.

[David Eichler]

For the same reason you can't unbake a cake.
And depending on the sensor native “color space”, ProPhoto RGB a theoretical space based on a display, can make that unbaking even less possible.

Sorry, not a proper use of an analogy, imo. You claim the result is the same (cannot undo), but the process of baking is not at all comparable. I would like to know the technical reason that is your basis for the claim (which is very different from heat applied to flour, egg, and water yields a cake). If you can provide a link to something that explains it, I would appreciate it.

[digitaldog]

Since the majority of the work I am doing requires using PS, one way or another, that flattening for output is a part of the process. So, it is really a moot point for someone like me.

Me too. But it doesn’t change the facts of what occurs with the data in either product.

[digitaldog]

Sorry, not a proper use of an analogy, imo. You claim the result is the same (cannot undo), but the process of baking is not at all comparable. I would like to know the technical reason that is your basis for the claim (which is very different from heat applied to flour, egg, and water yields a cake). If you can provide a link to something that explains it, I would appreciate it.

Changing pixel values isn’t the same as creating those RGB pixels from sensor data.
Start here:
http://www.lumita.com/site_media/work/whitepapers/files/pscs3_rendering_image.pdf

[David Eichler]

Changing pixel values isn’t the same as creating those RGB pixels from sensor data.
Start here:
http://www.lumita.com/site_media/work/whitepapers/files/pscs3_rendering_image.pdf

Thanks.

[digitaldog]

Thanks.

Karl’s piece is IMHO, the best Adobe white paper ever published and one of the best on photography, perception, photo history, digital imaging. The guy is brilliant.

[Rand47]

The ability to add to and subtract, and “intersect” with the various masks is quite nice.  The ability to label them is a Godsend… finally, Hallelujah!   I typically use a smart object workflow that I learned from Charlier Cramer and do most of the “heavy lifting” (as much as possible) in LrC.  Then specific surgical things in PS w/ layers.  Being able to jump back to ACR and re-tune what I’ve done as the effects of PS work affect the overall look of the image is nice.  So, I don’t see it as an either / or as much as a nice integrated suite of tools.

Also, after soft-proofing and printing from LrC, I hope never to have to print from PS again, thank-you-very-much.

The DAM aspects of LrC either are valuable for you or not, but they are “there” and powerful.  E.g. show me every image I shot with a 70-200 Nikon at ISO 800 on a Z7 II with keyword red, rose.   Hard to do that in PS.

Rand