Processing steps

[Fox Paw]

I'm still trying to get to where I know what I'm doing when processing photos. At this point, I have a fairly set pattern of steps, but I don't know if it's a good pattern. I'd be curious as to what steps others usually take in processing.

For me, at this point:
1. Adjust as much in Raw as I can, except for sharpening, and convert to a tif.
2. Straighten or correct perspective if needed.
3. Crop if needed.
4. Levels layer to modify the tones and correct color casts.
5. Many local curves layers to modify overall and local contrast.
6. Local color balance or saturation layers, if needed, to take a color up or down.
7. Sometimes a small overall saturation boost to some or all colors.
8. Sharpening, usually with high pass sharpen but sometimes with USM.
9. Print it and look at it for a few hours or a few days. If something's wrong, I'll figure out what (with luck) and go back and fix it (with luck). Or start over (often), if I realize that I took the photo in the wrong direction from the start.

Anyone willing to share their habits?

[PhotoJet]

I do as much of the processing in the RAW file as I possibly can. When I open the file in PSP, I want it as close to what it should be as I can get it. I will adjust brightness/contrast levels and such in PSP and apply any "after-market" effects, too prior to printing. But I use the Canon Photo Professional software to do as much as I can prior to converting... including sharpening.

[Fox Paw]

PhotoJet, is your Canon software Digital Photo Pro? I think we have the same thing.

Everything I've read indicates that sharpening should be left for last, or nearly so, because it's destructive. That's the only reason I don't sharpen before converting. When I have sharpened in Raw, though, I can't say that I see much difference in the final product. Sharpening remains a mystery to me. I know how to do it but I never really know how much.

[Walter]

...

Sharpening remains a mystery to me. I know how to do it but I never really know how much.

I've heard that sharpening should be done last, and according to the print size. I rarely sharpen and if so, only locally to a minimal degree (USM or Smart sharpen).

I can't stand nikon's capture nx, so I usually just use it to convert my d300 shots to tif. My cs2 bridge can handle the raw files from my d80, and I'll occasionally adjust white balance and white-black points, sometimes exposure.

convert to tif
clean sensor spots
more clean sensor spots
even more clean sensor spots
& again.

Adjust levels
-and locally

Adjust curves
-and locally

Optional:
-Magic
-More magic
-Back off magic

Dodge and burn

Save

Not much help am I? :irked:

[ann]

process raw files in either capture nx2 or photoshop, save as tiff files,

i pretty much following George DeWolf's recommadations
crop, contrast, brightness, color, defects, sharpnen

I use the following steps to do the above:

check levels and curves and then may add some creative pp, sharpening for printing. I rarely chop, but would do that first when i do
if i don't make a print, i don't sharpen.
For a long time i didn't sharpen until i started reading Bruce Fraser books and his logic and samples changed my mind. I always felt my images looked sharp, but when i started printing i started sharpening for the paper type.

[Overread]

My typical work through:

1) Open RAW file with Photoshop Elements 6 - this starts up Camera RAW (v 4.2 for me) which I sort of understand

2) Adjust the White Balance as needed - generally I do this by eye and often end up slipping into custom, sometimes the auto settings work and sometimes they totally fail

3) Adjust the exposure and recovery as needed - sometimes a little adjustment is needed here, normally I use recovery a lot more than I do the exposure slider.

4) Contrast, brightness and blacks - no real set method just by eye again and tweaking things - I might even adjust recovery and exposure again based on changes - I do keep an eye on the histogram for all this and try to keep burn outs to a minimum (but also watch on the shot where they occur.

5) Clarity tool - boost this up as much as needed, making sure to view the image both at fullsize and at a size where I can see it all to make sure that I don't get halo effects or a false shadow appear around the subject (can happen so keep an eye out for it).

6) I might adjust vibrance and saturation but often not.

7) On the Detail tab Noise reduction - Colour is set to 25 as standard

8) Crop and straighten if needed (though I'll admit I often forget this stage and tend to do this in photoshop)

9) Open Image (opens it in PE6) note that I set the depth to 8-bit because elements does almost nothing for 16bit. If I had the option I would output as 16bit.

Inside Photoshop elements 6

note if its a macro shot add in lots of work with the spot heal tool - tones of it if the shot comes from my MPE lens (seriously it gives a whole new meaning to sensor dust!). I have also yet to find the add magic button :(

1) First stage is Curves/levels. I have a curve plugin that lets me use curves without the silly sliders that are the elements default. Since going RAW I find that auto (on either this or levels) tends to do almost nothing since the whitebalance has already been fiddled with. So I will often play with the dipper tools (often I find I just use the black and white when I can)
After this its curve playtime - I kind of know what I am doing - I tend to fix the midpoint of the curve with a dot and then adjust the upper 3rd depending on the secene - lifting it up (which causing the lower 3rd to dip) is effectivly boosting the contrast of the shot, often works but if I have strong whites present I might just dip the lower 3rd down a bit (darkening the darks) whilst keeping the whites either the same or dipped a little just to help control them.

2) Duplicate the base layer (my curve plugin does this when it does the curves, if I don't use that I always make sure I have an untouched base layer of the shot to go back to)

3) Enhance - adjust lighting - shadows and highlights. I'll often use this to further help with highlight control on the shot (though it could probably be similuated with better curve control) as well as to show up a little more details in the shadows if needed. Again its very much playing with sliders to get a pleasing effect - I don't tend to adjust the midtone contrast very much either.

4) All shot dependant - playing with saturation, contrast and brightness settings - layermasks often come into play to mask off or limit the effect. How much is done depends on the shot - with a good shot very little with ones with more problems or a look I'm going for more is done.

5) Should really be where I save the image as a PDS or a Tiff and then flatten all the layers followed by duplicating the new base layer in preparation for noise and sharpening. I will admit I often don't bother since websize (my main current output) is often not affected. But if I am bothering I do it here - I also title the save file with a note that no sharpening or noise is done on the shot

6) Apply noise reduction and sharpening (unsharpen mask - amount as needed, radius 0.8, threshold 4) as needed.

7) For the internet resize image to 2000pixels on the longest side and sharpen again - then resize to 1000pixels on the longest side and sharpen a final time (often this last time is very little in amount). Then save and upload to flickr.


Note I've never done local curves or levels - is that levels and curves with layermaskes? IF so how does one decide on the areas to choose to apply them to?

[Walter]

...

Note I've never done local curves or levels - is that levels and curves with layermaskes? IF so how does one decide on the areas to choose to apply them to?

Pretty much that with masks. I just look for where-ever may need it, lighten a shadow or bring out something.

Translate "Magic" as listed above to shadow/highlights occasionally.

My current phase is doing as little as possible.

[Overread]

This is well worth watching - an interesting approach
http://www.youtube.com/watch?v=hn8FKlslwl4

watch the whole series

[René]

My way is usually:
1: Import in Aperture

2: Quick preview and delete unwanted rubbish!

3: Crop Image and Straighten

4: Adjust Highlights/Shadows

5: Adjust colours - hue - saturation etc

6: Use Definition on image

7: Rate 1-5* and add keywords and name and copyright info

8: Decide then to keep in colour or BW and export the image

9: Take another sip of coffee before it gets cold!

[ann]

thanks Overread,

i plan on giving those a view. I tend to use cs2 as i am used to the layout however, it won't support my new camera so i am now using either 3 or 4 when i need to convert from RAW. These may be very helpful in creating a better comfort level for the layout of the program as it has changed a bit.

[Fox Paw]

I missed this thread while it was developing. Interesting.

Overread, I don't know what others do, but I usually do local curves adjustments as follows. Set the freehand selection tool to feathering of maybe 8, to keep the boundaries of the local adjustment from being noticeable. Select the area you want to adjust. Add a curves layer, doing whatever you wish to do to the selected area...lighter, darker, more contrast, less contrast, color curve. After you make the curves adjustment, add a Gaussian blur of maybe 5, which blurs only the edges of the adjustment layer and further helps to avoid a sharp boundary between the adjusted and unadjusted areas. That's it. The same approach works for local color balance or saturation adjustments. If I'm adjusting a sky, I'll use the magic wand to make the selection rather than the freehand tool, but otherwise it's the same.

Since I started this thread, my approach has changed a lot. I'm now doing 90 percent of the processing in Lightroom and usually using Paint Shop Pro only for a few tweaks. Lately I've been processing landscapes once for the sky and once for the ground--and sometimes once more for shadow areas--and combining the results in PSP with masks or the blend range tool (= "blend if" in Photoshop, I think).

[Matt Needham]

Open in Adobe Camera Raw. Apply my default ACR settings for the camera and/or situation. I have several presets made up as starting points for processing typical shots from various cameras, and then a low light (high ISO) version for each camera. Mainly they switch camera calibration to "camera neutral" which is my typical picture mode, modified curve, and adjust the sharpening settings slightly (usually more amount and less radius compared to ACR default). I follow the Bruce Frasier/Jeff Schewe school of sharpening which calls for 3 stage sharpening: capture, local, & output. I am also a commie and heretic. ;) Not really, sort of, but sometimes when I tell people I use 3 stage sharpening I get the "Burn the witch!" treatment. :twisted:

In ACR I adjust the image lightness first if necessary. The exposure slider pulls the histogram from the brighter end while brightness pulls it from the middle. I use one or a combination to brighten or darken the photo. If I need to adjust the black point I may do it now or later, or I may just handle it with the curve.

I do white balance adjustment next.

Crop and straighten.

Then I adjust contrast, clarity, recovery, fill, vibrance, and saturation as needed. ACR is non-destructive so it doesn't really matter what order. Usually it's just contrast and clarity. The other sliders often aren't adjusted.

Adjust vig, and fine tune the curve.

Search and destroy dust and acne. Use the ACR local adjustment brush, gradient, etc... to locally adjust as I can. Local burning, dodging, contrast, clarity, sharpness, saturation, color tint, etc...

If needed open in Photoshop as 16 bit ProPhoto RGB. Do any local or global adjustments I couldn't do in ACR. I think the ACR local adjustment brush is wonderful, and for most of my photos it works great, but occasionally it's just not the right tool for the job.

If I don't need PS I just keep the adjusted raw file. If I made adjustments in PS I save as a 16 bit ProPhoto RGB tiff if flattened and psd if I need to save layers.

At this point the photo is finished except for resizing and output sharpening. I tend to sharpen 12"x18" and smaller prints all about the same, but for larger prints I get better results assessing each one individually. My final global sharpening is done in PS. I use a high percentage/low radius Smart Sharpen, followed by a low percentage/high radius unsharp mask on a copy layer set to luminosity blend.

For resizing I use Bicubic Sharper and Smoother in PS. If I'm down sizing for web I run my watermark action (found it with google search), and then use bicubic sharper as my very last step (after global sharpening). If that looks over sharpened I redo using plain Bicubic. If I'm increasing size I use Bicubic Smoother. Some folks say increase size in 10% steps. I've done extensive personal testing and can see no difference between that and just doing the up-res in a single step, so I just do it in one step. I have found that sometimes it's better to do the output sharpened before the up-res, and sometimes after. It just depends on the subject, texture, etc.... By turning layers off and on it's easy to see the results from either.

Convert to 8 bit, sRGB, and save as jpeg

[Chris of Arabia]

I've tried following this little lot a good few times now (cbrush (http://www.cbrush.com/)), but I'm still not convinced over the way it handles certain things, for instance, noise seems to get far worse just after it's done the 'add contrast using USM' step, which you then have to start dealing with later on. It has though got me far more comfortable in using many of the tools in PS, and also in understanding what you can and can't do with layers and masks - bloody awful example photo though...

[Matt Needham]

...noise seems to get far worse just after it's done the 'add contrast using USM' step...

Try using different settings that he recommends. For 8-12 mp images I like a percentage of 10%-20%, and a radius of 20 to 60. Also adding a bit of threshold will keep the usm from effecting noise so much. If you give it a threshold of 3 you are telling it to ignore details 3 pixels in size and smaller.