DPChallenge: A Digital Photography Contest You are not logged in. (log in or register
 

DPChallenge Forums >> Out and About >> Photoshop Basics Mentorship Group: Workflow Thread
Pages:  
Showing posts 101 - 125 of 155, (reverse)
AuthorThread
10/04/2005 11:01:29 PM · #101
Originally posted by bear_music:

I appreciate all the kind words of support, but I wish to make one thing perfectly clear:

What's being taught here does NOT "qualify" me as some sort of Photoshop expert. This is very fundamental stuff, and it's all stuff I taught myself over years and years of using Photoshop.


Regardless of your "classification" of proficiency with PS, this tutorial is immensely valuable to those of us without the years and years of exposure. Keep up the good work.
10/06/2005 02:24:54 PM · #102
OK, here's my second shot at it. I managed to keep the tower from turning too blue, but the reflected reed tips in the water have gotten too yellow. My impression is that some things get 'fixed' in a way after flattening that have an effect on what you can do with hue/sat and levels. For example the tower and the reed tips, when you don't do it right before flattening the image, it's almost impossible to adjust the picture like Robert did later. Is this true?

10/06/2005 05:55:05 PM · #103
Originally posted by ajschel:

OK, here's my second shot at it. I managed to keep the tower from turning too blue, but the reflected reed tips in the water have gotten too yellow. My impression is that some things get 'fixed' in a way after flattening that have an effect on what you can do with hue/sat and levels. For example the tower and the reed tips, when you don't do it right before flattening the image, it's almost impossible to adjust the picture like Robert did later. Is this true?



Correct. Flattening "fixes" color shifts and assigns them to their "true" spectrums.

R.
10/12/2005 04:33:17 PM · #104
Bear will there be anymore?
10/12/2005 06:02:58 PM · #105
Yeah... When I start breathing again... been a damned busy 2 weeks. Next installment will be using USM, and that's a tough episode to write.

Robt.
10/12/2005 06:05:00 PM · #106
Cool, just wanted to make sure, no pressure. Just when you are ready.Thanks for all your hard work :)
10/13/2005 09:41:57 AM · #107
I too want to thank you.
10/13/2005 07:12:28 PM · #108
Yep. Thanks. I just edited a bunch of photos I took at a Hustler Lingerie party last weekend. They still aren't professional, but much better than I could have done in the past. I had never been exposed to layers before.
10/13/2005 09:44:22 PM · #109
Bear, I have been one of those "following along" without saying too much. I am learning a helluva lot and wanted to pass along my thanks.
10/17/2005 12:43:14 PM · #110
Basic Photoshop Workflow — Continued, Part V

In Part I we uploaded and image from the camera into an organized file structure on our hard drive. We then saved the image as a .psd file so that we could work on it in a format that allowed us to keep layers and so forth intact in the saved version. We learned that we will NEVER do any work on our original file, but instead will clone off new copies of this file as needed to experiment with. We learned that our basic workflow is Open/save as .psd/adjust image/save final version at full size with layers intact/resize and adjust/publish to web. We looked at the concept of saving more than one version of an image, and why we'd do that.

In Part II we discussed what the Photoshop tool IS from a basic, photographic point of view, and how we would be using it as a part of our basic workflow for presenting images. We touched on what might best be done in-camera and what might best be left "neutral" so we could fine-tune it in Photoshop. We listed 7 basic things that Photoshop is used for by photographers. Finally, we promised the next section of this tutorial would discuss actual Photoshop workflow, the order in which we do things.

In Part III we took a diversion and discussed the basic use of "layers" in Photoshop. We showed how layers are not as complicated a concept as they seem at first glance, and why they are vital to an organized workflow in image editing.

In Part IV we discussed a simple Photoshop "workflow"; some of the many things we can/should do in Photoshop, and in what order we should do them?

Now it's time to examine sharpening in Photoshop. There are several tools for sharpening images, but in digital photography on of them, Unsharp Mask (USM) is the sharpener-of-choice, and we'll concentrate on that in this tutorial.

Photoshop's Unsharp Mask: the Basics of this Basic Tool.

Here's where we left off on the last lesson:

At this point we have a "corrected image" that's ready to be fine-tuned. It's now time to sharpen, resize, border, and save for web. Issues of sharpening are a subject unto themselves, and they interrelate with sizing, so we will come to that with the next section of the tutorial. For now, be aware that the sharpener-of-choice is "Unsharp Mask" and take care to examine your image very closely as you sharpen it so you avoid introducing "artifacts" (usually halos between bright & dark areas) during the sharpening process.

We should first discuss what sharpness actually is in photography. That may seem silly (we all "know" what sharp is, right?) but it bears examination. For one thing, an image can be in focus and still not appear to be "sharp". "Sharpness", in fact, is perhaps too general a term for what we are discussing here. We are not discussing whether an image is in focus, and this is what some people refer to as "sharpness". What we're working with here might better be termed "acutance", but we are going to stick with "sharpness" for this basic tutorial.

Remember our work with Levels? Here's what I want you to do; open up any ordinary (no extreme contrast) image in Photoshop and add a levels adjustment layer to it. Now, in the levels dialogue box, see the bottommost bar? The one with 3 sliders, white on the right and black on the left? Notice how if you slide the right pointer left (towards the middle) your whites get gray. Likewise, when you bring the left slider to the right, your blacks get gray. This bar sets the maximum dark and maximum light for your image.

Slide both pointers towards the middle and watch your image lose contrast. Notice how the less contrasty it is the less sharp it looks.

Now return the pointers to their origins at far left and far right and go up to the histogram, the tool we worked with before. On this one, the pointers set the white point and the black point; everything to the right of the rightmost pointer will be white, and everything to the left of the leftmost pointer will be black. So when you slide both of these towards the middle of the histogram, the image becomes more contrasty. Notice how, up to a point, the more contrasty the image is the "sharper" it looks.

Sharpness and contrast are very closely interrelated, especially in the digital world. Sharpness, basically, is about the perception of edges; when we bleed off contrast, we make edges less noticeable, and vice versa.

Now, perceived contrast in an image is directly related to the size at which the image is presented and the distance at which it is viewed. Prove it to yourself by looking at any full-tone image you have hanging on your wall up close and from a great distance. The closer you get to it (the "bigger" it is visually) the more "resolution" you can see and the more details you can see. Step back far enough and you see no detail at all, so that the image becomes less and less differentiated.

It's a fact of photography that the larger an image is printed, the less contrast/sharpness needs to be added to make it appear "sharp" to our eyes. Open up a straight-out-of-the-camera, properly exposed image and use the magnifying glass to blow a section of it way up. See all the nice little details, sharp and clear? Now shrink the same image so it's about the size of an image viewed on DPC and watch the details (and the apparent "sharpness") disappear. As we lose details, sharpness goes south on us.

What this means to us on DPC, where we display 640-pixel images, is that as we shrink the image size (and lose a LOT of detail doing it) we need to "increase sharpness" (which basically means increasing contrast at a very fine level) in order for the image to display as a "sharp" image.

This is why the most important sharpening, for us, is done AFTER resizing the image. If you're working with low sharpening or no sharpening in-camera, one of your first steps in the workflow would be to apply the appropriate sharpening to the entire image at full size, but for the sake of this basic tutorial I'm assuming most people have their in-camera parameters set to produce a clean-looking image straight from the camera. So, for us, sharpening is done after resizing.

Unsharp Mask: How to Use It

Paradoxically enough, USM works by using blur to define edges. Sharpening happens when we increase the contrast of an "edge", a juncture between light and dark, and Photoshop USM works behind the scenes applying a certain amount of blur to the image and then digitally comparing the blurred image to the original image to "define" the edges within the image to which sharpening will be applied. The decision Photoshop is making, on a pixel-by-pixel basis, is to compare each pixel with its neighbors and decide whether it should be lightened, darkened, or left the same. If Photoshop decides the pixel is located on an "edge", it will adjust its value; otherwise it will leave it alone.

Now, open a DPC-sized image and set your parameters so you are viewing your 640-pixel image at "actual pixels"; the size it will be displayed at on DPC. You can do this in the "view" menu. Now, for this exercise, go to "layers/duplicate layer" and make a copy of the background to work on. We are doing this so we can easily discard our mistakes and start over again from the base layer. In Basic Editing challenges, do your sharpening directly on the base layer. Now go to "filters/sharpen/Unsharp Mask" and open the dialogue box.

You'll see 3 fields that can be adjusted, either by moving a slider or by typing a value into the box at the right of each slider bar. They are Amount, Radius, and Threshold. Each of these fields does something different, and they all work together interactively.

1. Amount: This adjustment tells Photoshop where to draw the line between a "light" pixel and a "dark" pixel. As you move the slider to the right, dark areas get darker and light areas get lighter. Move it too far, and you produce pronounced haloes around dark/light intersections.
2. Radius: This adjustment is basically doing exactly the same thing as "radius" in Gaussian Blur; it's telling Photoshop how much blur to use in the edge-finding stage.
3. Threshold: This adjustment tells Photoshop how much difference should be present between the light pixels and the dark pixels before sharpening is applied.

(This has taken me a long time to write because I am trying to reduce a very complex topic into a form that can be understood without too much technical mind-bending. Once again, I caution that this is a very simplistic representation of the topic and I request that others not step in to "correct" me with more precise details.)

Okay, back to actually USING the USM tool; you have an image open, so start dragging sliders around willy-nilly, way to the extremes, and watch how much things change. Amazing, huh? Your task is to find a happy medium for the specific image. You'll adjust Radius to define your edges. The larger the radius, the larger the area (the "border") that will be lightened or darkened. Contrasty images need less radius, flat images need more.

You'll work with Amount to effectively increase or decrease the amount of sharpening that's applied at the specified radius. The larger the radius, the larger the amount you'll use, as a rule.

Finally, you'll use Threshold to fine-tune. As a rule, we use Threshold at very low values in continuous-tone images (photographs). A threshold between zero 1.0 is all you're likely to need.

Typical reasonably-sharp-from-camera images might use an Amount of 50%, a Radius of 1 pixel, and a threshold of zero. Less-sharp images might use an Amount of 100-200%, a Radius of 2-3 pixels, and a Threshold of 1.

As always, once you've made your USM adjustment you can go immediately to "edit/fade USM" and play with fading it in and out. You can overcook USM a bit then zoom in on the image and fade it until there is no visible haloing at magnification.

There are MUCH more sophisticated ways to use USM, including layering varying degrees of USM in different layer modes, doing the sharpening in the luminance channel on an image converted to lab color, and doing 2- or 3-pass sharpening runs. I'm aware of these, and indeed my standard USM flow (which I've saved as an action) uses 4 passes on 3 different blending modes, but that one's not legal for basic editing. So we'll leave it at this for now, at least.

The best thing for you to do now is to start playing with USM on a selected image. Work the variations. Get a feel for what the sliders do. Find a setting that works for you and your camera; every camera's different. Just remember: The goal is an image that looks sharp but not harsh, natural, and that has no visible haloing along the light/dark interfaces. When people say an image is "oversharpened" they are referring to one (or both) of these things.

******************

This pretty much covers our basic, simple workflow. I'm open to requests for what to get into next, people. Borders is a possibility. What do YOU want to see?

Robt.

10/17/2005 01:25:22 PM · #111
Bear, I've also been lurking and following this mentorship. First, I'd like to say thanks for putting the time into this, it has helped me imeasurably.

Second, I'd like to vote for a lesson on borders. Seems like that would be the next logical step. Also, I'd love to see a discussion on "Save for Web" vs. "Save As..." and doing your own settings.

Thanks for your work on this!
10/17/2005 01:31:36 PM · #112
Originally posted by livitup:

Bear, I've also been lurking and following this mentorship. First, I'd like to say thanks for putting the time into this, it has helped me imeasurably.

Second, I'd like to vote for a lesson on borders. Seems like that would be the next logical step. Also, I'd love to see a discussion on "Save for Web" vs. "Save As..." and doing your own settings.

Thanks for your work on this!


"Save for Web" is a simple choice. Here's why: in Photoshop, any time you save a jpg or save-as a jpg, the dialogue box shows you the degree of compression (which you can adjust) and the resultant file size. Sounds good, right? BUT PS also appends a text file of exif data to the image, and that's not included in the total. So you can save-as jpg at, say, 147kb, and when you go to uplaod it DPC tells you your file size is larger than 150kb, and you have to go back and twiddle. Pain in the keister. Save for Web leves the exif data off, and you get a clean file at the specified size.

Plus with Save for Web you can specify the filesize you want it saved to, and the program chooses the compression to make that happen. Use Save for Web for all images you wish to post to the web.

R.
10/17/2005 07:27:12 PM · #113
As usual very helpful and I will be having a play in the next few days :) Any homework ? Always helps when I am forced to do something!
10/17/2005 07:36:16 PM · #114
Re: Part V

Thanks, bear. You're still da man!
10/18/2005 01:58:36 AM · #115
Originally posted by joynim:

As usual very helpful and I will be having a play in the next few days :) Any homework ? Always helps when I am forced to do something!


OK, HOMEWORK

The following image is straight-from-RAW, resized and jpeg'd. NO sharpening, no exposure adjustments, no color saturation, no contrast tweaks, nada. All camera adjustments zeroed for this one (and indeed for all my RAWs). I have processed a similar image from this shoot and it's very nice color, nice dramatic sky, sharp as a tack.

Your assignment: use USM to sharpen this image: show us one you think is "correctly" sharpened and one you think is clearly, obviously oversharpened.

Extra credit: do the full monty on it; hue/sat, levels, etc. If you're doing it all by basic rules (what we've been "teaching") then see how well you can contain the sky while letting everything else have some life. It won't be easy, but try using selective color in the white channel on the sky.

Feel free to select the sky (use magic wand, then "select similar", then the rectangle marquee with the alt key held down to subtract any stray selection below the skyline) and work with the sky and the rest of the image separately, if you like; that's one way to deal with this sort of extreme contrast. Another way, if you have CS2, is shadow/highlight tool, but that may not be enough by itself in thsi case, depending on what effect you are looking for.

Have fun.



Robt.
10/18/2005 06:09:24 AM · #116
OK bear, here's my try:
Not sharpened................Sharpened enough.............Too damn much

Also took a shot at the overall adjustment of the image, tweaking the levels and maximising the Icelandic potential of the sky. Used levels, hue/sat, selective colour (mainly white channel as you suggested) & shadow/highlights. USM settings were (NIL), (65/1.2/2) and (165/2.2/0) respectively.
10/18/2005 06:13:25 AM · #117
Hmmmmm..... Now that I look again at my alleged 'finished' image (middle one above) I see I've washed too much green out of the original. I guess that happened while trying to boost the highlights in the darker areas?
10/18/2005 07:12:15 AM · #118
Thanks for this new installment. Great job as usual.
Genrel adjustments made are:
Levels: It had nice looking histogram to begin with, so I only took the midtones abit to the left, too lighten the trees a bit.

Hue/sat: overall increase in sat + some tweaking to the blue, green and yellow channels. I got some artifacts in the clouds that I am not sure how to avoid.
Like ubique mentioned the sky here screams for an icelandic/heida burning treatment but of course that's nor allowed.

USM: I did 2 passes (something learned not so long ago and LOVE)
1-pass: I don't remember the exact numbers but something like 25/50/3. Too add contrast
2-pass: somehting like 110/0.8/0



in this one I selected the sky and darkened it with the brightness control. There are some areas betweeen the leafs and branches that I missed but I didn't hane anymore patience. sorry.

10/18/2005 12:25:15 PM · #119
Ubique: Really nice work on the sky, all things considered. You're lightening/contrasting the foliage more than I'd consider optimum, but that's a personal judgment, not an absolute one. Consider this: clearly the scene is backlit, right? Anyone who knows beans about photography is going to see the artificality of lighter, contrastier foliage. That's how I'd look at it. As for sharpening, I believe the lesser-sharpened of the two is still too sharp.

Ref: Your first example is a very natural rendition of the scene, all in all. Sharpening seems about right. Like with Ubique's version, I feel, personally, that the foreground foliage is too light, but... The sky may be a little too blue. Don't worry about the artifacts; they are happening 'cuz you are working from a lo-res version of the image. Working from a full-size original, it would be better. Your "darker sky" version darkens the sky with the wrong tool. Try tweaking the selection in levels instead, or better yet with curves also, though we haven't discussed those. Levels to set the light & dark points, curves to force the contrast on the midtones.

Everyone; this is actually a bad example for use in "selecting" the sky and working with it; I tried myself, and the interstices on the overlapping, central tree are driving me batty. So much so that I wouldn't even DO it (the picture's not worth it) but I'll post my own example shortly ignoring the really bad effect where the tree and sky meet.

Robt.

Message edited by author 2005-10-21 04:58:01.
10/18/2005 12:43:38 PM · #120
Okay, here's mine as far as I'm gonna bother taking it. The sky NEEDS to be bright in this image, it's the natural state of things. I've pushed it as far as I can, I'm getting a trace of artifacts myself. "Colorizing" the bright areas is resulting in a most unnatural looking sky, so I've abandoend that approach. Yucky picture, actually.



Robt.
10/21/2005 05:49:11 PM · #121
Hi to all. I've found this thread very informative. Robert, your notes are easy even for a novice like me to understand. I have a question, however. I was just getting used to unsharp mask in Photoshop Elements and now I just got CS2 with the smart sharpen tool. I have no idea what settings to use in smart sharpen and, would you use it instead of or as well as unsharp mask? Any thoughts. Thanks - Julie
10/22/2005 03:09:47 AM · #122
Originally posted by JulieG:

Hi to all. I've found this thread very informative. Robert, your notes are easy even for a novice like me to understand. I have a question, however. I was just getting used to unsharp mask in Photoshop Elements and now I just got CS2 with the smart sharpen tool. I have no idea what settings to use in smart sharpen and, would you use it instead of or as well as unsharp mask? Any thoughts. Thanks - Julie


I don't have CS2, so I can't advise on this. Try forum search on "smart sharpen", I know it's come up before.

Robt.
10/22/2005 10:05:14 AM · #123
Here are mine Bear, first the oversharpened one:


200/.7/2

Not awful, but 640pix images just can't take anything over about 100/.5/2-3

Next, the one I tried to whip into shape using curves, hue/sat and USM:


Didn't spend a whole lot of time on this, but to try to give some life to the greenery (yellowrey) and boost contrast a little. Came out looking a little to dark in the foreground.
10/22/2005 11:19:52 AM · #124
Originally posted by strangeghost:

Here are mine Bear, first the oversharpened one:


200/.7/2

Not awful, but 640pix images just can't take anything over about 100/.5/2-3

Next, the one I tried to whip into shape using curves, hue/sat and USM:


Didn't spend a whole lot of time on this, but to try to give some life to the greenery (yellowrey) and boost contrast a little. Came out looking a little to dark in the foreground.


I don't consider the first to be oversharpened for web viewing, not at all. In fact, the second isn't sharpened enough, for my monitor. I'm curious; what sort of monitor you using?

The colorized version is about as good as we've seen so far, nice job.

R.
10/22/2005 02:38:14 PM · #125
Originally posted by bear_music:

I don't consider the first to be oversharpened for web viewing, not at all. In fact, the second isn't sharpened enough, for my monitor. I'm curious; what sort of monitor you using?

The colorized version is about as good as we've seen so far, nice job.

Apple eMac/CRT, not LCD, which might account for how harsh the first looked on my monitor as I tweaked it. I pushed that one about as far as I generally push a full res JPG when I'm editing for a challenge or for a print, about 200/.7, which is generally way too much for a web sized image. Of course, when I use my usual workflow, I've already USM'd an image at full size, and then apply just a little tune-up once I have resized it for web, prior to saving for web.

Message edited by author 2005-10-22 14:39:28.
Pages:  
Current Server Time: 04/18/2024 08:22:21 AM

Please log in or register to post to the forums.


Home - Challenges - Community - League - Photos - Cameras - Lenses - Learn - Prints! - Help - Terms of Use - Privacy - Top ^
DPChallenge, and website content and design, Copyright © 2001-2024 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 04/18/2024 08:22:21 AM EDT.