P: A Revolutionary New Masking Technology

Lee Jay's automasking demo (above) is indeed impressive - I'd say:
- His automasking brush skills are far superior to 99% of Lightroom users, or
- That took him over an hour to do, or
- He cheated ;-}
(or all of the above)
My mileage varies quite a bit too...
PS - What Lee Jay means (he'll correct me if I'm wrong no-doubt), is that the auto-mask uses an adaptive algorithm. In other words, if you start brushing on blue but continue brushing into the browns, it'll assume you must want to paint browns too, or you wouldn't keep brushing on them. I would have thought however, that if you release the mouse then all bets are off again. I'll be doing more experimentation...

Lee Jay, what do you mean that it continually adds color? I haven't been able to select blue sky behind trees, which should be very easy you would think.

Masking is probably one of the editing tools that needs the most improvement, so I second this suggestion.

To clarify the feature request:

I am proposing that Adobe expand the (auto)masking tools in Lightroom.

Although improving the existing brush's automask would also be worthwhile, that has nothing to do with the subject of this feature request, unless you think the brush's automask obviates the need for other forms of automasking - I don't.

Put another way: I don't see a more U-point-like masking technology as a substitute for the existing brush masking, but a complement.

Yet again: I think it would be a nice addition to also have a masking technology that does not require painstaking brushing for those instances when its possible to specify what you want to target more simply and quickly in other ways. Hopefully, the mask generated in this fashion could be modified using the existing (hopefully improved) masking brush.

I agree with your assessment of the current brush auto-masking, and in case you haven't already, its worth re-reading Lee Jay's "key element" clause above.

My proposal is not for an alternative to the brush's auto-mask, but an addition to the masking repertoire.

Sorry for going into the solution instead of trying to elaborate the problem - I think you're right that Adobe could engineer a better solution than I, once they set their mind to it. I just hope they do set their minds to it...

I'm not recommending the present brush's automasking be gutted. I use it too sometimes, and agree that with some TLC it could be improved. I'm hoping it can be maintained along with some other masking tools.

PS - I'll be taking your key element into consideration as I practice - thanks.

I understand what you are saying Rob but I think we are better off by describing the problem the auto-mask tool does not solve and let adobe come up with a solution than attempt to provide the solution. I agree with Lee Jay - I like the automask idea better than u-point and that the automask/adjustment brushes need some TLC including:

• 
  
• halo control on high contrast edges
  
• needs to mask all develop changes
  
• some control on how/what is being masked
  
• in-depth documentation/tutorial
  

"Targeting light blue sky behind trees, without selecting the dark blue car below the tree, nor the tree branches / leaves, nor the clouds..."

Rob...I did all of the below with automask.

http://photos.imageevent.com/sipphoto...

I've said before that automask needs some TLC and that it could use some additional complexity (something like tolerance and edge controls), but I already find it more powerful and flexible than Nik's solution. However, it's not at all intuitive to use and even after messing with it for over a year I hadn't figured out the key element which is that it continually adds colors to the selection if you drag over them or click on them. My intuition told me that the set of selected colors would remain fixed at initial mouse-down.

Thanks Marc - I looked at the Photoshop masking demo (http://t.co/l4z4Ivl)

But I really think Lightroom needs something more like U-points.

I've used Photoshop's masking tools some, as well as VertusTech's Fluid Mask, and Topaz ReMask, and onOne's Mask Pro..., but none of these are really what's needed in Lightroom, IMO.

They are all aimed more at selecting objects, e.g. for extraction and compositing.

U-Points are aimed at selecting specific colors, luminosities, and textures. As implemented by Nik, they either work great, or you switch to plan B.

I'd like it to be: They either work great, or you refine them until they do.

Don't get me wrong - I can easily see Adobe tapping into their existing masking experience, and Photoshop code-base - this is why I think they could do such a good job. Its just that the masking technology as is, misses the mark for Lightroom purposes, IMO.

I mean, one idea:

Maybe an A-path, instead of a U-point. And the path density represents the force of the effect, whilst path width indicates the geographic scope of the effect (in u-point parlance: the "radius"). I might have to get one of them tablets ;-}

Then along the path, one could adjust color tolerance, luminosity, and structural / textural attributes. This would solve one of the problems with U-points: Sometimes a big point is too big, but bunches of small points make "waves". And, there is no way to hint at what should be included/excluded.

And another problem with U-points (sorry, I told myself I wasn't going to go here) is they transcend the other selection mechanisms, meaning you can't use the eraser to erase parts, or marquees to scribe a boundary around... (you can add additional manual masking in, but that's not the same - very much like whats being discussed in the gradient mask thread here - erasing a gradient mask is different than trying to reverse the affects with a brush afterward.

Anyway, I assume if Adobe invented something similar from scratch, it wouldn't violate Nik's patent.

One facet of my present vision in a nutshell - an A-path: like a U-point, except without the limitations of u-points - integrated in with the rest of the mask drawing features, so its not all-or-nuthin - e.g. slap a few A-path's down and maybe fine tune the targeting criteria - that gets you most if not all the way there, then tweak the mask using whatever mask manipulation methods are available, then start tieing local adjustments to the mask.

Summary:
--------
The key word here is auto-masking. User specifies geographic extent, and tolerances for the things that determine inclusion vs. exclusion, then the software generates a perfectly seamless mask, which may or may not perfectly include/exclude what is wanted, which can then be used as is or modified using other mask manipulation tools...

Example Use case:
-----------------
Targeting light blue sky behind trees, without selecting the dark blue car below the tree, nor the tree branches / leaves, nor the clouds...
Drop an A-path in the sky, then specify the range of colors and luminosity to be affected. Drop an "anti" A-path on the car, and specify a different luminosity range. Voila, you can now adjust the sky and leave the car alone (estimated time: 2 seconds). This is impossible to do right now in Lightroom, because the auto-masking associated with the brush would not be adequate to get just the sky, and the HSL sliders would include the car since it has the same blue hue - only the luminosity differs.

Final thoughts:
---------------
Photoshop's masking may be better for selecting a person, but I'm just after adjusting one of the colors in her blouse, without adjusting like colors in her scarf...

Rob

This type of thing would greatly enhance the use and marketability of Lr.