Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: AltGirl on February 28, 2011, 08:30:01 pm
-
I am struggling with a Photoshop challenge that I can't seem to resolve and would love any help that I can get...
I am looking to pull out both the highlights and the shadows from a given image into two separate layers or files. What is the best method of pulling out highlights and shadows in the curve dialogue box? Understanding that the bottom of the curve is the shadows and the top is highlights, how do I work the curve so that I am left with just the highlights or shadows from a given image?
Please ask any questions you need to for clarification.
Thanks!!
-
You may want to try using the "blend if" blending option. There was an article on this web site a while back. You may be able to find it with a search. Here is a link to another article. http://members.ozemail.com.au/~binaryfx/blendif.html If I understand your question correctly, "blend if" should easily do what you want.
-
I'm not sure what you wish to accomplish but you might try this.
1. On the background layer do a Select>Color Range and pick "Shadows" from the drop down box.
2. Ctrl+J to make a new layer and label it "Shadows"
3. Repeat steps 1. and 2. for Highlights.
This will give you two layers, one for each.
You may find that you prefer a stronger selection than this in which case i suggest using a Luminosity Masking technique discussed at Tony Kuyper's site -
http://goodlight.us/writing/luminositymasks/luminositymasks-1.html
-
Hi,
I'd suggest you can use the light and shadows dialogue.
I have written a small article on using luminosity masking, here: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/46-fixing-sky-with-luminosity-mask
The article recommended by Jack Varney is much more extensive: http://goodlight.us/writing/luminositymasks/luminositymasks-1.html
Best regards
Erik
I am struggling with a Photoshop challenge that I can't seem to resolve and would love any help that I can get...
I am looking to pull out both the highlights and the shadows from a given image into two separate layers or files. What is the best method of pulling out highlights and shadows in the curve dialogue box? Understanding that the bottom of the curve is the shadows and the top is highlights, how do I work the curve so that I am left with just the highlights or shadows from a given image?
Please ask any questions you need to for clarification.
Thanks!!
-
I’m familiar with Tony Kuyper’s luminosity masks. I used to use them and think they’re great. However, since I discovered “blend if”, I use it instead. I think that luminosity masks and “blend if” accomplish basically the same objective, but “blend if” allows precise control over what is included and excluded as well as how abrupt or gradual the transition between the two.
-
... how do I work the curve so that I am left with just the highlights or shadows from a given image?...
You can add a few fixed points to the curve in the areas you do not want to change (say, non-highlights), and then manipulate the rest of the curve (i.e., highlights portion of the curve).
-
I have used Erik Kaffehr's technique for adjusting the sky values and I highly recommend it!
-
In two words: use a grayscaled version of the image as the layer mask; then tune the mask using curves.
-
I am struggling with a Photoshop challenge that I can't seem to resolve and would love any help that I can get...
I am looking to pull out both the highlights and the shadows from a given image into two separate layers or files. What is the best method of pulling out highlights and shadows in the curve dialogue box? Understanding that the bottom of the curve is the shadows and the top is highlights, how do I work the curve so that I am left with just the highlights or shadows from a given image?
Please ask any questions you need to for clarification.
Thanks!!
You can select the lightest half of the image (the highlight end of the curve) by opening the channels tab and ctrl clicking (say) the red channel thumbnail.
Use this selection to create a mask in a curves adjustment layer. Call that your Lights layer.
Duplicate that layer and invert the mask and call that your Darks layer. You can then independently work the highlights and shadows using these layers.
I have assumed here you are familiar enough with PS to work with basic layers and masks but if you need a step by step explanation then feel free to come back and ask.
It would be good common forum courtesy to provide feedback as to whether any of the responses to date answer your request.
-
Luminosity masks. Much better than the Select>Color Range>Highlights/Shadows because the masks are self feathering so you get a very smooth transition between tones. With the Color Range method you can get very defined lines between tones depending on how severe your Curve adjustment is and feathering the selection doesn't work overly well.
The Layer Blending Properties (Blend If) can work as well but I, personally, don't find it as subtle or elegant a solution as luminosity masks. It is less work though.
-
The Layer Blending Properties (Blend If) can work as well but I, personally, don't find it as subtle or elegant a solution as luminosity masks. It is less work though.
Bob, just curious why you find Blend If less subtle. Maybe I’m missing something and should reconsider using luminosity masks.
It seems to me that Blend If is capable of at least as much if not more subtlety than luminosity masks. For example, if you want to isolate highlights, you can select to include 100% from level 255 to 197, and then gradually decrease what is included with a gradual decrease to 0% at level 162. Of course, you can use whatever level numbers you want. Thus, Blend If allows complete control over the level where you start to exclude pixels as well as how abrupt or gradual the transition. The luminosity masks I’m familiar with didn’t permit such control. Perhaps I need to take another look at luminosity masks.
-
Yep, absolutely agree that's how it works. And I'm not saying it's a bad approach. I just, personally, prefer the luminance masks. I find that with the layer blending properties I get a less smooth result than with the luminance masks. It's not dissimilar to using the Select>Color Range>Highlights approach. The selection isn't feathered the way it is with the luminance masks so the effect of any Curve/Level/HS adjustment is more abrupt and more quickly veers toward posterization with any kind of serious adjustment to the selection.
-
I think I understand, Bob. Feathering is different than just a gradual blending from, as in my example above, level 197 to level 162, or even level 197 to level 0. For example, if I have a highlight area, say around level 200, next to a dark area, around level 95, the Blend If gradual blending doesn’t blend such an abrupt transition well, whereas feathering does because it feathers based on pixel radius not luminosity, correct?
If so, it makes me wonder if it’s possible to create a mask using Blend If and then apply feathering to it.
-
Another approach altogether is to open the image in Camera Raw or Lightroom and use the excellent tools for adjusting highlights and shadows in the Basic panel (Exposure, Recovery, Fill, Blacks) for achieving the end result you want, then export the image back to Photoshop. You can do this with TIFFs, PSDs, JPEGs and of course raw files. Making maximum necessary use of these tools at the raw processing stage of course vastly reduces the need for a lot of gymnastics in Photoshop afterward. Not to say the other techniques don't work, but try the easy way first - which in this area happens to also be highly effective a good part of the time.
-
Bob, just curious why you find Blend If less subtle. Maybe I’m missing something and should reconsider using luminosity masks
The most important distinction is that luminosity masks are built from image data whilst blend if is based on numbers that have no intrinsic relationship to the image content. In general, more accurate and subtle masks can be built from image data. Making selections and masks from channels is a good example.
-
The way luminosity masks intersect is different from applying feathering to a selection, Dean. The intersection of luminosity masks is more of a gradual overlapping of the selections. So when you adjust one mask it will affect others to a lesser degree which helps keep the tonal gradiations smooth. Even when you feather a selection, it's a discrete selection and doesn't impact any others even ones that may overlap.
Nick, is that really accurate? What I mean is that the layer blending options Blend If condition applies the blending of layers based on brightness values in the image. If I want to hide high brightness values, I move the white slider to the left and values above where the slider is will be hidden based on those values in the image layer. If I split the slider the hiding of those values will be more gradual. So it is using brightness values in the image.
-
The most important distinction is that luminosity masks are built from image data whilst blend if is based on numbers that have no intrinsic relationship to the image content. In general, more accurate and subtle masks can be built from image data. Making selections and masks from channels is a good example.
I too really don't understand this statement. Everything in digital imaging is based on numbers. Complex mathematics underlies every manipulation we do. The numbers make the image - it's ALL numbers under the hood. A blending operation, whatever tool or technique you use, is blending numbers between the active later and the layer below it in order to effect a change of image appearance.
-
Hi,
Luminosity mask are quite flexible and can be very subtle. I often use channels for luminosity masking or image->calculations. The idea is to try to utilize the mask in the image.
This article shows one way of doing it: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/46-fixing-sky-with-luminosity-mask
It is very well possible to achieve the same result with different tools. It's well possible that some tool is better for some task than some other, but it is always the result that counts.
Best regards
Erik
The most important distinction is that luminosity masks are built from image data whilst blend if is based on numbers that have no intrinsic relationship to the image content. In general, more accurate and subtle masks can be built from image data. Making selections and masks from channels is a good example.
-
I too really don't understand this statement. Everything in digital imaging is based on numbers. Complex mathematics underlies every manipulation we do. The numbers make the image - it's ALL numbers under the hood. A blending operation, whatever tool or technique you use, is blending numbers between the active later and the layer below it in order to effect a change of image appearance.
Y'know what, ignore what I said. You are both right, what I said does not make sense on re-reading.
I think I must have meant is that whilst blend-if can indeed be used much like luminosity masks, lum masks are more versatile given that you can also quickly make selections and hence layer masks directly from them. This is because the blend-if operation is essentially based on tones only, whilst masks made from channels and luminosity can also be made from tones but applied to specific areas of the image. I'm sure combos of all the above can be used to the same ends though - you use the tools that you think are appropriate I guess.
-
I'm sure combos of all the above can be used to the same ends though - you use the tools that you think are appropriate I guess.
Actually, I would suggest doing just that...combine the power of Blend If with the power of luminosity based masks...since what you have then is the ability to target with the Blend If and modify with the luminosity mask as a layer mask. Best of both techniques...
-
Another approach altogether is to open the image in Camera Raw or Lightroom and use the excellent tools for adjusting highlights and shadows in the Basic panel (Exposure, Recovery, Fill, Blacks) for achieving the end result you want, then export the image back to Photoshop. You can do this with TIFFs, PSDs, JPEGs and of course raw files. Making maximum necessary use of these tools at the raw processing stage of course vastly reduces the need for a lot of gymnastics in Photoshop afterward. Not to say the other techniques don't work, but try the easy way first - which in this area happens to also be highly effective a good part of the time.
This is absolutely correct, but I find I need to do more adjustments with masks in PS after getting the image as close as I can in ACR. I find that with ACR, the different sliders affect the entire (or almost the entire) luminosity range, although to different degrees (so if I am adding fill, it also is having some effect, though smaller, on the mid-brights), whereas with masking I can adjust tones in a certain range without affecting the tones elsewhere at all.
-
That's why adjustments like Fill and Recovery are cautioned to be used with a light touch and if something more extensive is required, another method is better used.
-
That's why adjustments like Fill and Recovery are cautioned to be used with a light touch and if something more extensive is required, another method is better used.
When you say "are cautioned" - who is doing the cautioning? I've found these adjustment can be effectively used over a wide range of settings - depending on what else you do in the same application to complement the adjustments.
-
Here - http://www.earthboundlight.com/phototips/camera-raw-recovery-fill-light.html
Martin Evening in his Photoshop Lightroom book.
I've seen it written and recommended elsewhere as well but am not going to scour the internet looking for the sources right now.
In my own experience, I agree. I know that means nothing. ;D
-
Martin Evening in his Photoshop Lightroom book.
See the note on page 274 of the Lightroom 3 book. He notes that the 2010 improvements to LR have changed things with respect to the problem that you note.
-
See the note on page 274 of the Lightroom 3 book. He notes that the 2010 improvements to LR have changed things with respect to the problem that you note.
Indeed, and Bob Johnson, in that dated article, didn't really discuss this so much as a problem, but to advise quite cogently that one shouldn't try adjusting things beyond the limits of what the image itself can deliver. This seems sensible to me.
-
I don't have his LR3 book but I still see problems when trying to push either Recovery or Fill.
Mark, I think I'm saying the same thing. Use judiciously and don't try to push too far.
Anyway, it's not worth getting into a urinating contest over.
-
Anyway, it's not worth getting into a urinating contest over.
Absolutely - I agree. All I was suggesting is that it's good to maximize the appropriate use of these tools before doing more esoteric stuff in Photoshop.
-
All I was suggesting is that it's good to maximize the appropriate use of these tools before doing more esoteric stuff in Photoshop.
Absolutely no disagreement there, Mark. ;D