The Photoshop Thread - Any Experts Here to Help?

Cafe Society
1

I haven’t seen a general Photoshop thread here. :slight_smile: Perhaps the Dope has some graphics artists or Photoshop Gurus that can help?

I’ve been using Photoshop ten years. I’m still learning this beast. :slight_smile: I use 7.0. I see no reason to waste money upgrading.

The thread is open to whatever you want to discuss. Let’s learn together. :slight_smile:

I’ll start with unsharp mask.

I’m still struggling to fully understand unsharp mask. You have Radius, Threshold, and Amount. ScanTips defines them…

My problem is knowing when these are set correctly. Too much Radius causes large halos. Beyond that, I find it tough to see much difference in small changes.

I tend to leave radius at 1.6 and Threshold at 10. They seem to work with most images with faces. Change threshold to 8? I don’t see a difference unless I zoom way, way in. Even then I don’t know what the hell I’m supposed to be checking.

Can, anyone explain how they set these? How do you check and see if it’s right?

Here’s an example. Photo of Rep Giffords and husband.
original

reds were over saturated. Used a hue layer -15 to desaturate. Levels layer 12, 1.15, 255 I didn’t adjust the white (255) because there’s already a glare on the guys head. I didn’t want it to flare more. Shifting the mid-point lightened it up.

unsharp mask Amount 112% Radius 1.6 Threshold 10

unsharp mask Amount 108% Radius 1.6 Threshold 10

unsharp mask Amount 104% Radius 1.6 Threshold 10

The change in sharpness is very subtle. I’d probably go with the 104%. I tend to be pretty conservative because I don’t fully understand the process. It bothers me that I make a change and don’t know if it makes the image worse or better. :smack: I waste hours in a photo slideshow flipping back and forth trying to decide which image is better.

How do you know it needs a different radius? or Threshold? What are you checking? What’s the procedure?

2

2nd Challenge…

How would you clean this image up with curves? I have spent days and days trying to figure out curves. I’ll get pissed, throw the book down. Try again a few months later. Curves kick my ass every time. I’ve spent 5 years off and on trying to figure out a procedure to set curves.

I use levels because I understand it better. I always make an image worse trying to use curves.

3

I guess there’s no photoshop gurus here then?

I stumped the Dope. :smiley:

I’m an analytical guy. I like procedures and lists. That only takes you so far in Photoshop. There is an artistic element too. Graphics artists can do amazing work in photoshop. They are working with their eyes and creativity. It goes beyond rules that say click this tool, click that button, and so on.

4

There are entire books written about sharpening in Photoshop. If you really want to learn about it this book covers it all.

Your sharpening parameters are going to change depending on what your final use of the image is. In general, if you’re printing an image, you’re going to sharpen it far more than if you’re using it for screen. Pictures for print generally look oversharpened on the screen, but print very well.

My typical parameters for sharpening for detail for a full-resolution 12 MP image to be printed at 8"x12" is something like 0.8-1.2 radius, 220%, 0-10 threshhold. I will often also do a second pass for mid-tone contrast, using parameters like 10-30 radius, 20%, 0 threshold. This is called “hiraloam” sharpening (high radius, low amount.) Photographers will use this in conjunction with traditional detail sharpening to get even more “pop” from their images. Dan Margulis has a good treatment of the subject on that link.

I have to run, but there’s also good information of the basic parameters here.

5

Yeah, when I talk differences in sharpening, I mean something like 100% vs 150% vs 200%, not four percentage points at a time. In print, such a small difference would not be noticable.

6

That there’s the problem. The wall I keep running up against with photoshop is selections. Mainly because I’ll try to use tools to do the selections for me. I’ll try to set up my photoshoots (product shots) with monochromatic backgrounds that I can easily select out, sharp edges that I can easily crop etc. I want to be able to enter numbers into windows and have it select what I want selected dammit. But in the end, it takes someone with a steady hand and a good eye to go around the object to do it right and make it look good.

As it turns out, photoshop tricks left brainers into thinking it’ll make them artists, but you still have to be a right brainer to master it. To a certain extent, that’s why digital cameras are so damn fancy (to get techie people to think that if they buy this camera they can take amazing pictures) any pro photog will tell you it’s not the camera, it’s the photographer.

7

I get confused by Masks. I’m not even sure why they’re so commonly used by the experts because they barely figure into anything I do, there are always alternatives that seem to be just as quick and manipulable, if less “efficient” (i.e. I create extra layer copies and opacities to get what I want).

I avoid unsharp mask because it almost never does what I’m really after. Instead I copy the layer, use the default “sharpen” on it, and then change the layer’s opacity to blend with the lower one. Saves shitloads of experimentation with mystifying settings.

8

The only time I ever use masks is for complex selections (remember, I’m not that good at selections to begin with). Flipping over to the quickmask highlights everything I’ve selected in pink and makes it easier to see. You can then use other ‘regular’ tools to aid in making selections.
I think most PS books have at least one chapter on masking and I think I’ve even seen entire books on it.

9

Heh. Stumped the Dope? Nah, there’s just no right answer. Just like there’s no right answer as to which of your sample pictures is best.

What does stump me is why you aren’t just using the preview function in Unsharp Mask, instead of “jumping back and forth.”

As has been said, different photos with different target uses do best with different degrees of sharpening. If you’re printing at 150 ppi then the same sharpening Radius is going to have a more powerful effect than if you’re printing at 200 ppi.

One thing about Photoshop is that there are a dozen different ways to do the same thing. People come up with techniques that work for them, and they are rarely the same ones other people use.

Regarding Curves: My usual technique is to first use Adjust Levels to get a full range of black-to-white before going to Curves. In Curves my rules of thumb are:
–Don’t let the curve run (nearly) level for any significant distance
–add two handles and try an S curve to increase the contrast.
–Or add a single handle and make the curve concave or convex to lighten or darken a significant section of the image.
–Use the eye dropper to check out where the important part of an image falls on the curve. I’ll check where a facial tone hits the curve and then add a handle there to make the face darker or lighter.

I rarely use the separate RGB curves. Just too easy to mess thing up.

About the shiny head:
One retouching problems is adding believable pores to skin. (I’m not an actual PS guru, btw.) But if a face has blown out highlights that I need to fix, this is my current technique:

  1. Duplicate the image on a new layer.
  2. Create another layer above that one to do your retouching on.
  3. Set a soft paint brush to an appropriate working size and set the brush technique to Dissolve.
  4. Set the brush to maybe 10% or 15% opacity, and then use the eyedropper to select a color near the blown highlights.
  5. Lightly paint the highlight with the “Dissolve” specks, and then apply a Gaussian blur filter at a setting that leaves the specks visible but soft. Maybe 0.6.
  6. Select a different nearby color and repeat the paint/blur routine.
  7. Spend 12 minutes building up layer on layer of blurred specks of slightly different colors.

It looks kinda like skin pores and you can slowly shade the facial contours with darker and lighter sampled colors. It’s also pretty forgiving in that you can paint right over random hair strands without it looking bad.

10

If it works for you, it works for you, but I think sharpen is a poor solution for most photographic sharpening needs. It sharpens too many pixels of the image (too low a threshold), at too small a radius, and too low a level. If I had to guess, it’s like Sharpen’s settings are something like 125%, 0.5 pixels, 0 threshold. I’m usually at settings much higher than this, like 220%, 1.2 pixels, 5 threshold. But it depends on the image. For something with very blocky details (not fine detail), I might be at a radius approaching 2 or 3 pixels, with a slightly higher threshold, too. For something with really fine detail, a radius of 0.5, with a lower threshold. And with the uncustomizable Sharpen filter, there’s no way to do localized contrast adjustments like you can using the hiraloam technique mentioned above.

And even unsharp mask (though I do use it often) is not the ideal solution when you have smarter algorithms like Smart Sharpen out there that do an even better job of finding the areas you want to sharpen in a photo and avoiding those you don’t.

11

Sharpening doesn’t actually figure into my work too often. One technique I sometimes do, though, is to sharpen before any anticipated resizing.

12

The exact opposite of what I’ve been taught. Sharpening should always be the last thing you do. Sharpening and then resizing leads to big halos around details.

13

That hasn’t been my experience. If you resize (to smaller) after sharpening, it softens the extremes, and reduces or eliminates haloing.

Depends on what effect you’re going for, I suppose.

14

Yep, in general, with photographs, sharpening should be the last thing you do.

15

This bears repeating again and again and again, *ad infinitum, *and it applies to more than just Photoshop. People have gotten the idea that all they need is the latest gizmo, and it substitutes for years of education and experience. And the real professionals have learned that they need to use ***both ***hemispheres of their brains.

16

Reminds me of a story (yes, I AM a teacher, why do you ask?).
And I tell it to my Photoshop students:

Lindsay Buckingham was producing John Stewart’s solo album (that’s post-Kingston Trio, not pre-Daily Show). He got a call to do a Fleetwood Mac gig in the middle of mixing. He got off the phone, turned to John and said:
***“OK, you finish.”
**“Me?I don’t know anything about mixing!”
**“But you’ve seen me do it. Want to know the secret?”
**“Hell, yes! You’re a genius.”
**“See all those sliders and dials? Move 'em around until it sounds right.”

And… John Stewart quickly became a sought-after producer, and I keep seeing his name on excellent albums.

So I tell my students: Move those sliders (in your adjustments and filters) until it looks right.

17

That’s what I’ve seen too. I always resize a little after sharpening to hide the jaggies and other artifacts. Even resizing to 97% is enough to clean up after sharpening.

18

Then you’re sharpening with the wrong parameters. Sharpening should be the last step in post-processing, in most cases. If you’re sharpening for print, then don’t worry about the jaggies so much. Photos sharpened properly for print always look oversharpened on screen.

19

I’ll have to try over sharpening and then printing.

Are you printing in CMYK or RGB? I’ve read that the commercial outfits with offset printers all request files in CMYK.

20

I don’t really deal with offset printing that much, so all my printers ask for RGB files. Even the few books I make on offset print presses, the printers ask for RGB as well. I used to deal with CMYK separations when I was in the newspaper industry, but nobody I work with now asks me for CMYK files.