Let's Talk PS Workflows

[PhilDernerJr]

It's been a while since we actually talked about photography. Much time has passed and we've got many new people here, and I'm eager to see and learn from other people's post-processing workflows.

My workflow, in Photoshop CS2, is somewhat outdated. Some of it is just wrong, and some needs a fresher way of doing it.

First I level off my shot. I use the Ruler tool to find a straight line or point of reference in the shot such as a horizon or building corner (using one on the sides of the photo might be misleading if the lens has given a rounded effect). I then go to Rotate Canvas and click Arbitrary to automatically make that ruler line level.

Then I resize, usually to 1024 pixels wide. I resize before other editing because I want to see the full result of editing in the full-size that it’s going to be viewed in afterward.

Since my photos are only for web use, I do not use a dpi setting, as I understand that anything above 72 (standard web resolution) is not noticeable on a computer screen. If I feel I exposed the photo perfectly and it’s very lively, I’ll edit to 1200 pixels wide.

I then open up levels. I find that auto-levels and auto-contrast makes all of my photos too dark. It took me a while to feel confident in applying levels manually, but I’m pleased with my results for the most part.

I then clone out the dust spots, which is the most frustrating for me. Nothing pulls your motivation away more than using the Equalize feature and seeing about 30 dots across the screen. Using a new layer for equalize makes this easier and faster, but it's still a part I despise.

I then adjust the color balance….I feel that Canon outputs images with a slight greenish tint that I get rid of by adding 1-3 notches toward magenta.

Then and only then do I do my sharpening. I put the threshold to zero, and set the radius to 0.2. The amount that I apply then is determined by my camera. On the 10D, I do a pass at 500, then a second at about 130. On the 20D, a pass of 480 or so is enough.

I then rename in a format such as AA_A320_JFK31LDEP_082709. I need to change this for sure for organizational purposes.

What do you do?

[SmAlbany]

Using CS2, my workflow is not too much different.

1. Level the horizon - I use the measure tool and arbitrary rotate just like Phil described
2. Crop (3:2 ratio in the crop tool)
3. Levels - slide the left and right sliders to fill the histogram and then eyeball the center slider to taste
3a. Sometimes levels doesn't give me a result I like. I sometimes will then use the curves tool instead. I pick the blackest and whitest parts of the image with the appropriate color samplers (can use threshold tool to identify) and then slide up the center of the curve until the image looks right.
4. Color - I rarely do much here. Sometimes a touch of +saturation - especially for Continental.
5. Dust spots. I use a duplicate layer and apply equalization. Clone out the dust by writing on the bottom layer while viewing the equalized layer. Discard equalized layer.
6. Resize
7. Sharpening. I usually sharpen in layers. Make a duplicate layer. Apply USM (0,250,.2). Erase any jaggies with eraser at 50% opacity. If still jaggy, swipe it again. Flatten. Repeat until it looks right (usually 2 times does it for me - sometimes 3)

My naming convention is Registration#_airport_date_image#

All this work and still way too many rejections. I haven't changed my workflow in quite some time. If the original image is good - this works.

Dan

[mirrodie]

Hey, I hope to hear more. I have not edited a photo in ages. And I tried again the other night and it felt so foreign to me.

My issues usually lie in sharpening. Any sharpening tips? I usually find that I CANT find the fine line between sharp and oversharp.

[Hussman75]

Thought I'd throw in my $.02 on this one... Not the most sophisticated workflow, but I swiped it from a website run by a Brit about 22 months ago. Over time, I've incorporated a few of my own changes and additions, but the basics are really his. I'd be dead in the water if I hadn't found his website... My foundation philosophy here is to bring the picture up to an "eye memory" standard- what it looked like to my eye thru the viewfinder as I shot the frame(s)...

1. After downloading the pics to my computer, I separate by type or grouping (737, 747, Airbus Narrow, RJ, Etc...).

2. Go thru each shot and determine if it's a keeper or one for the circular file.

3. Since I'm shooting JPEG, the file gets carried directly into PS Elements, opened and the work begins.

5. First up is to adjust the contrast. I do this to get a baseline correction from the original image. Depending on the lighting conditions I was shooting in I'll either use auto contrast, or do a manual adjust using the shadows/highlights function.

6. Crop the image.

7. Sharpen the image. Generally, I'll do one round of USM, as I find that any more makes the shot look too sharp, however once every so often, I'll have a shot that needs a second round- either the whole picture or just certain areas. I'm set to sharpen at between 150-180% and .25 pixels, and do occasionally bump them up or down depending on what the shot looks like it needs.

8. Make a dupe layer, and run the eraser tool over every jagged edge. Occasionally, I'll go back and do it again if an area still looks too jagged. Spaces between flaps and wings are the usual suspects that require another shot.

9. Re-size the image. As a rule, I size to 1500dpi, across the longest dimension. If necessary, I'll create a dupe file at a smaller size, but 1500 is my standard saved size.

10. Spot healing and cloning. Dupe layer, equalize and spot heal the required items. Then delete the dupe layer. I'll kill any dust spots, distant birds, tree leaves, bugs, light poles, wires, and anything else that requires obliteration. If the sky comes out looking fubar'd after fixing, I'll use the blur and smudge tools to blend the correction a bit better.

11. Now I'll manually adjust the saturation, usually by sliding the bar to the right. I like this as it makes (usually) the colors more vivid, and seems to make the picture more correct to my "eye memory." If necessary, I'll also adjust the hue and re-tweak the contrast at this point. Also, I'll go into Levels here and make any RGB type tweaks or darken/lighten the image as required.

12. If required, at this point I'll do a noise reduction. Typically, I don't need to, but once in a great while I'll get stuck in a corner and have to go this route.

13. Save the image, and move it to the correct file folder.

HTH!

[Metal484]

This is some great stuff. I haven't used photo shop in ages and I only recently started getting into photography (especially plane spotting and sporting events) so this thread will be a lot of help when I will have trouble editing my images which I'm sure I will.

[pgengler]

My workflow starts in Adobe Bridge (CS4):
1) Add metadata templates with contact and copyright information to all photos.
2) Add location metadata to photos.
3) Add keywords (tags) to photos; I have a couple of categories (Airline, Airport, and Plane are the most commonly-used top-level ones for my aviation shooting) with gradually more specific subkeywords. For example, a shot of a British Airways Boeing 747 (registration G-BYGC) gets these keywords (bracketed ones aren't actually included in the metadata but are purely organizational):
[Airline]
- British Airways
[Airport]
- KJFK
-- JFK
-- Kennedy Airport
Plane
- Boeing
-- Boeing 747
--- Boeing 747-400
---- G-BYGC
4) Rate each photo, 1-5. My rating scheme is like this:
5: the most amazing shot(s) ever; the sort I might actively look to sell. So far I don't have any of these.
4: really good photos that are among my favorites.
3: photos I think are worth showing; anything at this level or above gets posted to my own site.
2: photograph is not flawed but isn't great; may be part of a group of similar pictures where one was chosen as a 3-star photo but I don't need 12 photos of the same thing.
1: photograph has some unfixable flaw: subject cut off, too far away, blurry, etc.
5) Open the photo in Adobe Camera Raw and make some adjustments, such as:
* Level photo using ACR's adjustment tool
* Crop to desired size/ratio
* Remove visible dust spots
* Tweak exposure, brightness, contrast, and white balance settings, usually using "Auto" as a starting point

I prefer to get as much done in Camera Raw as possible, since these changes are saved in a "sidecar" metadata file, so I can always go back and change any of them later without having to start from the unmodified original image.
At this point, I open the image in Photoshop. Since most of the work has been done in Camera Raw, there's not much to do here. I might do a more intensive search and destroy for dust spots (especially if I'm going to upload the photo to JP or Anet), and usually try doing auto-levels and sometimes auto-color; if I don't like the results I'll do Levels manually and sometimes tweak in Curves.
For sharpening, I use this technique:
* Change the image to use LAB color
* Select the "Lightness" channel only (the image will appear black and white)
* Apply an Unsharp Mask. I forget what my usual settings are but I think it's something like 0,100%,0.1. (The LAB color technique usually needs smaller numbers than sharpening the RGB image).
* Convert the image back to RGB color.
* Save as JPEG.

[jerslice]

1. Open files in RAW - make adjustments to noise, contrast, sharpness, and most likely the white balance, if necessary
2. Convert to exif-jpeg or to .tiff if I intend to print
3. Open file in CS4
4. Crop
5. Adjust levels/curves
6. Adjust vibrancy, color balance, saturation
7. Apply USM as necessary
8. Spot sharpen/soften
9. Blur out grain when/where visible and if I feel like it
10. Select and turn sky into a layer if visible - run despeckle on it
11. Make adjustments to background layer apart from the sky is necessary and/or feasible
12. Flatten image
13. Create copy layer, equalize - remove those dang dust spots while praying there aren't too many and they aren't in any exceedingly inconvenient locations
14. Repeat 13 until they're all gone or blended in - skip from here to 15A if printing
15. Resize to what the shot warrants for internet viewing - usually 1200; never lower than 1024, rarely larger than 1600
16. Save as file name straight from the camera in the jpa.net folder as jpeg
17. Cntl+Alt+Z my way back to full size
18. Add border
19. Watermark for flickr, other sites
20. Save to Finished Product Folder under name from camera as jpeg

15A: Make additional local adjustments - remove ALL visible grain on subject
16A: Stairstep image size if necessary to print size
17A: Make additional local adjustments
18A: Watermark if needed
19A: Save in required file format (.tiff for professional, very large print; jpeg for 12x18 and under) in the READY TO PRINT folder and rename file according to airline and type

[SmAlbany]

My workflow starts in Adobe Bridge (CS4):

Very interesting workflow. You do some things there that I have not tried. I spent some time in the past shooting RAW but then I got away from it because I really wasn't getting much bang for the extra work - not to mention dealing with larger file sizes. I may have to try that flow and see how it goes. By the way, is RAW required for the metadata? I assume that allows you to search for specific tags? I have CS2 so I assume that it has that capability.

Thanks for the post
Dan

[pgengler]

By the way, is RAW required for the metadata? I assume that allows you to search for specific tags? I have CS2 so I assume that it has that capability.

Not for the keyword/tag or contact/copyright information metadata; this gets saved in the EXIF headers for JPEG images. With more recent versions of Adobe Camera Raw (definitely with CS4, maybe with updates for CS3, not sure) you can actually open JPEG images in ACR and do the same non-destructive edits and have the changed saved to the 'sidecar' XMP file. I don't shoot JPEG so I don't know if all the options are available or how well they work on JPEG files, but the option is there.

As for searching, Bridge will only show you the keywords/tags that are present in images in the current directory; there's no way to get it to search multiple directories to find images with specific keywords. There are other programs that do this, however.

Most of the metadata-related workflow is copied from/inspired by suggestions from The DAM Book (where DAM==Digital Asset Management), by Peter Krogh. It covers just about everything involved in managing photos from start to finish so that you preserve the most detail and have it backed up, etc. I'd definitely recommend the book to anyone doing anything approaching serious photography.