Select Color Range not functioning as expected

[Bill Junk]

Here is an image to which I want to make some localized adjustments.  I thought I would use the Select Color Range tool to isolate the dark green trees along a ridge at the base of Mt. Ranier.  I'm eventually making this into a B&W unage and want to be able to control the density of some specific areas.

Here's an enlargement of the area I want to select from, only selecting the dark green trees.

In the end, when I applied the Select Color Range tool to the dark green trees I didn't get what I expected.  I started with Fuzziness set to 0 and Range set to 0 thinking this would provide the greates limitation on pixels selection.  I first used the left eye dropper to select an initial color, then I used the middle eyedropper to add more of the dark green trees to the selection.  What seemed odd was that the first selection with the middle eyedropper appeared as a square.  Each time I added a little more from an adjoining area the square had another rectangle added to it keeping it a square.  As I moved closer to the edge of the dark greek trees, instead of selecting only the dark green trees, it added another rectangular area to the side that included the trees and other parts of the image which were not dark green.  I had expected to get an irregular shape that was mostly the dark green trees and from there I could adjust the Fuzziness and Range values upward to get the most useful selection.  I didn't seem to work that way. 

Here's a closeup of the selected area.

Here is the mask that was added to a Brightness/Contrast layer as a result of this selection with the Fuzziness set to 0 and Range set to 0.

I've used this approach on other images and it seemed to work as expected on some and not on others.

What am I missing in this siuation? 

[cmgap]

Have you tried using the Camera Raw filter with a color range mask and point color? Hope the screen shots will show you enough to see if this will give you the control over areas that you want to modify.

[Bill Junk]

Thank you for the suggestion.  I have not tried using Camera Raw to do more than basic, overall image correction.  I prefer to make more targeted adjustments within Photoshop and have them on layers where I can adjust the layer mask if needed.  I will go back to that image and see what I can make happen.  

[AlanGilbertson]

The ACR filter is nondestructive, since you've made the image into a smart object already. Alternatively, you can use ACR on the original, direclty from Bridge (Ctrl/Cmd+R or File > Open in Camera Raw. The masks that you create in ACR or the filter are more flexible than the one you create with masks in the main Photoshop UI, and remain fully editable.

[Bill Junk]

I'll have to take your word for it that the masking in ACR is more flexible than masking in Photoshop.  I find it a bit odd that a capability provided by a secondary editing tool is better than that provided by the primary tool.  My principle use for ACR is to provide quick overall edits to a image and I like the adjustments found in teh "Light" section.  I do find ACR better for doing initial adjustments and as you mentioned, will make it non-destrive by attaching the edits to a smart object version of the image. Occassionally I will use the Linear and Radial Gradient tools to accomplish a regional edit, mainly because they are quick to do.  However, overall I prefer to make my edits using Photoshop layers because the layer stack makes it easy to see what has been done to transform the image.   I guess it comes down to a mattter of personal preference.  I've been getting better at using the Select Color Range feature in Photoshop, but I'm still wondering why I saw the results I had in my initial post.

[Conrad_C]

Camera Raw has become very powerful. In fact in some ways, such as how masks are generated, Camera Raw and Lightroom have gotten ahead of Photoshop.

I would not say Camera Raw is always better. Camera Raw and Lightroom now have enough tools that I don’t have to move images to Photoshop very often. But there are certain highly precise correction and retouching techniques that can still only be done in Photoshop.

Maybe the best way to put it is that they’re both excellent tools now, with Photoshop providing more precision at the cost of a lot more complexity. I think of them this way: There are times when Camera Raw is more then enough and other times when only Photoshop can finish the job, and in the middle there is now a large overlap where you can personally decide where you think the break point is in moving the image from Camera Raw to Photoshop. I like to stay in Camera Raw as long as possible (going to Photoshop only if needed), others like to do the basics in Camera Raw and more of the heavy lifting in Photoshop, and both approaches can be correct.

[Bill Junk]

I appreciate your prespective.  Over the years I've witnessed the introduct of Lightroom and the advancement of its capabilities, as well as the progression of increasing capabilities in ACR.  Until recently there hasn't seemed to be as much growth in Photoshop capabilities, but Photoshop is now fully embracing the AI revolution.  It just strikes me as a bit unexpected that Adobe has put resources into advancing ACR instead of putting many of those features into Photoshop.  As you say, people do have choices and can decided to use the tools that work the best for them.

[Conrad_C]

One possible reason is that Photoshop is a fundamentally different editor than Camera Raw/Lightroom. Photoshop is a traditional editor designed around editing pixels directly, and saving the changes directly to the original file. Camera Raw/Lightroom are in the class of parametric, nondestructive editors (which most new photo apps are on desktop and mobile today), where the original is not directly edited because the edits are held separately and can only be permanently applied to an exported copy.

Each fundamental architecture probably affects both sides of feature development, in that direct destructive editing and parametric nondestructive editing offer both different opportunities and different limitations for what features can be implemented next.

So, people like me end up asking both “Why can’t Photoshop easily do mask intersections and a group of multiple sub-masks, like Camera Raw can?” and at the same time “Why won’t Camera Raw let me clone into an empty area outside the image, like Photoshop can?”

For the question at hand, both ways have a good solution. To isolate a color in Photoshop, Select > Color Range has been a color pro’s go-to feature for probably over 25 years. To isolate a color in Camera Raw/Lightroom, a parametric Color Range mask works great.

[Bill Junk]

Thanks for the technical background.  Of course a software architectural choice will facilitate certain types of feature expansion and limit others.  Maybe ACR was added as a compromise to allow useful features that were implemented in Lightroom to have a somewhat equivalent capability in Phoroahop, but had to be implemented as an add on because of Photoshop's architectural style.  I'm still puzzled a little since everything I do in Photoshop is non-destructive.  All modifications are either made via layers or as filters applied to a smart object (the original image).  Now back to Select>Color Range.  There are a few idiosyncronies in the way it works that I have yet to master.

[Conrad_C]

Maybe ACR was added as a compromise to allow useful features that were implemented in Lightroom to have a somewhat equivalent capability in Phoroahop, but had to be implemented as an add on because of Photoshop's architectural style.

By @Bill Junk

Just FYI, the actual sequence was different.

Early digital cameras only shot in rendered formats like JPEG or TIFF. Camera makers started offering the option to save raw sensor data, and photographers jumped all over that because they would get more control, like how you can get a higher quality image from film if you not only control your printing, but you are also in control of how you process the original film.

As I said, pixel editors like Photoshop only work on rendered images, so although raw shooting was getting popular, Photoshop did not know (and still does not know) how to open camera raw images. You had to process the raw file first in a separate geeky third-party raw converter utility. Adobe decided to simplify that, so they wrote Camera Raw as a front-end plug-in for Photoshop that let you develop camera raw data in a user-friendly way, and it sends the rendered result to Photoshop. This was around 2003.

Around 2005, Apple released Aperture, which treated raw files from a whole-workflow perspective, one of the earliest parametric editors, a real break from tradition. In 2007, Adobe released Lightroom, which took the Camera Raw engine and similar to Aperture, built around it to give photographers a fully raw-based, end-to-end (import to print/upload) workflow, including virtual organization and bulk metadata handling. Countless other apps followed that model. The architecture of these applications is, as we’ve seen, quite unlike Photoshop or the other tradtional direct pixel editors.

In the beginning, Aperture and Lightroom had the features to make the most important basic corrections before rendering to an RGB format. For most people, that was not nearly enough, so images generally had to be edited further in Photoshop to get them to be good enough for final delivery. What has been happening, especially in the last ten years, is that software companies have been learning how to build on those parametric editors so that they’re now capable of things like local (masked) adjustments, noise reduction, advanced tone mapping… Today it’s much more likely that an image could be done from start to finish in Camera Raw/Lightroom. That is why today there’s more overlap between them and pixel editors like Photoshop.

[AlanGilbertson]

Adding to what Conrad said, it's important to realize that a camera raw file is not an image. This isn't an intuitive idea, since our cameras and our tools always show us an image, but a matrix of cell values with a Bayesian sensor distribution is far more like a somewhat complex spreadsheet than it is like a picture.

Photoshop is an image editing tool. A raw file is just data that must be interpreted into a raster image format before it's visible or printable. Conversion can be done in-camera, which is why cameras produce jpegs instead or, or in addition to, a raw file. A raw file must be converted to an image before Photoshop can do anything with it.

ACR was in Photoshop long before Lightroom showed up. I believe it dates to Photoshop 7, so that's almost 20 years earlier.

[Bill Junk]

Thanks to both of you for the background on ACR, Lightroom, and Photoshop. I was using Photoshop before Lightroom came out but had forgotten much of the history details.

[davescm]

@Bill Junk 

Try resetting the Select Color Range tool and also  check your eyedropper settings for sample size and layers.

I downloaded the corner of the image that you supplied in your first post and all is working as expected here, I could not replicate your full square selection.

Dave

[Bill Junk]

your use of the Select Color Range looks fairly good.  I went back to that image and tried again on Layer 0 with the same settings you used:  Localize color clusters unchecked, Fuzziness set to 23.  The first click with the selection eyedropper added stuff from all over the image.  Then after using the add eye dropper almost the entire image was selected. I tried it in both the standard release of Photoshop and also the beta release.  I also tried several different sampling sizes.  The larger the sampling size was the better the selection was up to a point. but I could never get it as isolated as you were able to do.  What I did notice is that if I have Localize Color Clusers checked, then the closer the Range value is to zero the more the selection takes on a rectangular appearance.  I did go into ACR and tried the Luminace selection tool on another image and I was able to get a much better selection of an irregularly shaped region than I could have obtain in Photoshop.  It's too bad that Photoshop doesn't have a similar selection capability with similar controls, particularly being able to feather the brush easily.

[AlanGilbertson]

The two easy ways to do brush selections in Photoshop are with the new Selection Brush (which is circular only, but has a full range of hardness), and the pro's faithful standy, Quick Mask. The latter involves first applying a black fill (which shows as red in Quick Mask), then painting white with whatever brush size and hardness are suitable. When you exist Quick Mask mode, whatever parts of the image you painted with white are selected. You can use either method to fine tune a selection.

[Bill Junk]

Thanks for the suggestion.