My Digital Photography Workflow: Post-Processing
We're on part three of my three part mini-series on digital photography workflow. In part one I covered my in-camera workflow, and in part two I covered my organization workflow. So now that the photos have been captured, stored, and organized, we're ready to start the editing process.
I usually don't jump right into editing immediately after organizing my photos unless I have a lot of extra time to burn. When I do have an hour or two for editing, I look to my “Process” label (as I explained in the organization workflow) and pick one that suits my mood. Some photos I know I'll be turning black & white, others I know will be in color, but for most of them I have no pre-thought plans. When I see the untouched photo, though, I start running through all kinds of ideas on what I might do with it in Photoshop.
Once I've picked a photo for editing, I locate it on the hard drive (easy task with Picasa) to find the RAW file. Like I mentioned before, I don't allow my organization software to index my RAW files — only their JPEG counterparts. This helps to keep the viewing process fast while removing multiple views of the same photo (JPEG and RAW) in the software.
A RAW file must be processed through a RAW editor prior to doing any additional editing within Photoshop. I use Adobe Camera RAW 4 (ACR), which comes with Photoshop CS3. In ACR, the first thing I nail down is the white balance. This can be done with presets, custom adjustments to the temperature and tint, or by allowing the software to automatically choose for you. I usually see how the auto white balance looks, in addition to the presets. I pick the best looking setting, then I'll fine tune the temperature and tint.
The next step to RAW editing is dealing with all the exposure and color adjustments. This includes things like exposure, recovery, fill light, blacks, brightness, contrast, vibrance, and saturation. I usually just hit the “Auto” button and see how it turns out. It usually gets pretty close to being acceptable, not needing more than a few tweaks here and there. The important thing to watch for during this phase of editing is that your histograms don't get clipped on either end — that means you're losing information.
There are several other sets of adjustments that can be done in the RAW editor (curves, lens adjustments, etc), but I typically don't bother. I like to do more of the editing in Photoshop so I have the flexibility of using layers and layer masks. When I'm ready to open the file into Photoshop, I check that it's set to Adobe RGB and 16 bits/channel at a resolution of 240 pixels/inch. Then I open a copy of the file by holding the “Alt” key and clicking “Open” — I do this so I don't change the camera settings in the original. RAW files are actually two files: one for the data and one for the camera settings. The data file can't be changed, but the settings file can. I like to preserve the original settings.
The file opens up into Photoshop with the settings that have been applied in ACR. The first thing I do is save it as a Photoshop Document (PSD). I save the file to the same directory as the JPEG and RAW file, and with the same file name. So now, the photo has four files with the same name: JPEG, RAW, RAW Settings, and PSD.
After saving the file, the very first thing I do to the photo is touch-ups. This includes getting rid of any dust spots, abnormalities, things you just don't want in the photo, and anything else that needs to be taken care of. I consider cropping to also be a part of touch-ups, but I don't typically crop unless the photo absolutely needs it. Cropping should be done up front so you don't waste your time or energy on parts of the image that won't exist in the final output. After cropping, I add a new empty layer to the image and name it “Touch-ups”. I use this layer to do all my cloning or healing brush activities so that I'm not destroying the original data. There's a neat little option for these tools that allows you to sample from the current layer and all layers below. This allows you to clone like normal without picking up the stuff you're cloning out from the layers below.
I'll go through a lot of experimenting with various adjustment layers, masking, conversions, blending, etc. I don't really have a formula for what I do to each photo — each one is different. But whatever I do, I do it non-destructively. By that, I mean that I can simply turn a layer off to remove that step in the editing process. I never touch the original pixels. Adjustment layers are key for ensuring data retention, but adjustment layers aren't always available for every type of editing. This includes things like layer blends, sharpening, blurring, and other effects. For things like this, I create a blank layer on the top of the layer stack and merge all the visible layers into it. To do this, highlight the empty layer and press Ctrl+Alt+Shift+E. The resulting image shouldn't look any different, but now you have a copy that you can do some destructive editing on. The only downside to this is if you decide to readjust one of the layers below, it won't show up in the copy layer. There might be a way around this by using CS3's smart objects (or whatever they're called), but I haven't played around with them much. Regardless, if I decide to sharpen, it's the last step in the process.
When I'm all done editing, I double check that I've saved it to the PSD file. Then I convert the image to sRGB from Adobe RGB and 8 bits/channel from 16 bits/channel. Assuming the image looks pretty much the same as before, I save a high quality copy to a JPEG with the same filename plus the letter “a” to designate that it's been altered (PICT3246.JPG ==> PICT3246a.JPG). Now the photo has five files associated with it. This new JPEG gets indexed in Picasa and I can place it into the correct finished product label, give it a description and title, and do extra keywording. I save the altered JPEG in sRGB so I can quickly make scaled down copies of it for the web without having to open the PSD file and do all the conversions over again.
If I decide to print the photo, I'll softproof the PSD file in Photoshop using the correct printer profiles. Depending on where it's being printed (at home or in a lab), I'll do all the necessary adjustments and color management for that printer and either print straight from Photoshop or save it to a TIFF or JPEG to be printed, again depending on where it's being printed.
That's pretty much it from start to finish. Everybody does their post-processing differently, and we all have our little things that we're adamant about doing. No single process is right, so use what you can from my techniques to improve your own technique.