One of my favourite things to do as a photographer is to look back upon old photos I shot many years ago and see how I can improve them with new skills I’ve acquired since then. The most useful tools for a photographer are not cameras, lenses, and lights, but instead are time, perspective, and reflection.
Photography technology moves at a blistering pace, and while a photographer can constantly chase the latest and greatest tech, taking a step back into the past is a valuable exercise. You may have noticed my Instagram feed is a mix of newly shot photos as well as reworked classic images from years past. This is because I’m constantly trawling through my old photos, trying to spot images I overlooked the first time or images that can be re-processed to look a million times better. Some photographers advocate keeping images only for a year or two before discarding them as part of a good digital asset management (DAM) strategy. Granted, it takes a ton of storage space to keep every photo you’ve ever shot — but storage is dirt cheap, so there’s no reason not to.
My primary post-production tool is Adobe Lightroom, which gains new features several times a year. A lot has changed since the initial release in 2007, meaning any photos I processed way back then might have a lot more life in them than I was able to extract at the time. New camera profiles, localized tools, and other additions means I can make photos I didn’t know were possible nearly 10 years ago.
Not only can a regular review of images allow me to see where I went wrong with lighting and posing, but it also allows me to see where I may have made mistakes in the images I selected for retouching — I often find a similar image I like more than the select I made many years ago. As I’ve progressed as a photographer, my eye has gotten better and my tastes have changed, so it’s a small thrill to find a new gem in an otherwise sad pile of pictures from when my skills were sub-par.
Below are some rather embarrassing examples of how badly I screwed up my post production in the past. Ridiculous though they may be, I appreciate looking back at these images and fixing them — it’s been a good learning experience so I know what not to do in the future.
Let’s take a look at a diptych I’ve processed on three separate occasions over the years. The initial images of Melissa in the bathtub were created in 2007. I’m not sure what I was thinking then, but apparently a cool blue tone seemed appropriate at the time. Generally, avoiding blue tones with skin is a good course of action, but there are times it can be made to work. This wasn’t one of those times. It didn’t work. At all. The skin tone and overall feel of the image is grey and “dead”. Melissa looks like a wet zombie. It’s also worth noting how primitive camera sensors and Adobe’s camera profiles were in 2007 — the incredibly unappealing texture of the noise in the shadow areas being the prime indicator here.
Click to enlarge
I reprocessed this image with Lightroom in 2012, pushing those blue tones even harder, but this time giving Melissa a more appropriate, human-like skin colour. Unfortunately, I overdid it with the noise reduction and skin smoothing. Detail in the hair has been obliterated, and while the skin colour is indeed more human, the overall look is more robotic plastic than anything. That being said, I made the eyes come alive in the photos by brightening them up, revealing the brown iris, and eliminating the dark circles under the eyes. Strangely, I didn’t edit out the seam on the wall like I did the first time around. Overall, this was a half-assed second effort.
That brings us to 2016, where the third time is a charm. The noise reduction has been toned down. The wall seam is gone. The eyes have been processed the same way. Blue tones are completely gone! Considering this was shot with only the available tungsten lighting in the bathroom, this diptych represents the original scene as best as I can remember it. In a nod to current processing trends which go for a film look, I’ve added a subtle layer of grain across the entire image, pulled the blacks slightly, and given the colours and contrast a slightly muted treatment.
Obviously my own improved skills and improved software have helped me bring the best out of this image, even though it was captured with a 10MP camera from 2005. But who knows how I’ll feel about these images another 4-5 years from now. Will my own skills be leaps and bounds better again? Will software continue improving? How much more information could possibly be in that file for me to take advantage of?
Moving on to 2009, here’s an image from one of my first ever shoots with Samantha. The processing is not as cringe-worthy as the images of Melissa above. In fact it’s really not that bad at all. Her skin is on the warm side, but still has detail and texture. No obvious sensor limitations are visible. The overall treatment is natural and pretty “normal”. So why go back to this one? Well there are a few little things that bother me. With Sam’s torso being lighter than her face, the eye naturally wants to go to the lower half of the image first, or to the hotspot on the hardwood floor. Her dark hair blends in to the dark background a bit. Sam’s arm and hand look weird with the bed sheet obscuring them. Lastly, the reflections on the fridge in the background really bother me. It’s tough to say whether I noticed or cared about any of this seven years ago — I really can’t remember.
When I set out to improve this image, I found a better one from the same set to work with instead. The image is very similar, but taken from a slightly different angle to eliminate the fridge in the background and prevent the sheet from doing weird things to Sam’s arm. Again processing for 2016 trends, I’ve muted the overall palette, pulled the blacks slightly, and given the image a “shot on film” vibe. There’s also more detail brought out in the hair, Sam’s face has been made the same brightness as her torso, the hotspot on the floor is toned down, and the light falling on her is more interesting overall by virtue of her scooting a little bit closer to the window blinds.
Click to enlarge
The 2016 version still isn’t perfect, of course. The trendy processing might look weird, dated, or even awful a few years from now. The little curl of hair on the right side of the image distracts me. Sam’s pinky finger sticking out away from the rest of her fingers drives me mad. The background is busier than it was before, but at least it’s darker than the highly reflective fridge, so Sam will be the focus. You might even prefer the original image and that’s okay. I like the new one better. I can see a future where I come back to this image and make it a hybrid of the two versions. Something with slightly stronger colour and contrast, but not too much.
While we’re reviewing photos of Sam, here’s one from 2010. This is just all kinds of awful. Sepia toned but not quite. Rounded corners? What the hell?
Click to enlarge
Just like the previous example, a slightly different shot was selected for my 2016 interpretation. The processing is better in every way — especially because there’s actual colour now! I was also able to tame the backlighting a little more, so it doesn’t have as much of an obnoxious glow. Big improvement.
Here’s one more image of Samantha from that same day. Again with the sepia-but-not-quite styling. What was I thinking? There’s also no detail in the shadows and it looks like I did nothing at all to smooth out the skin tone on her legs.
Click to enlarge
The new edit is again a different image, where Sam has turned ever so slightly away from the camera, which makes the torso appear slimmer due to the angle being cut down. This was shot on grey seamless paper, but I’ve added a slight cool tone to the shadows which is a popular, very 2016 thing to do. Sam’s shirt actually has detail now — you can see the folds and patterns. And just like before, the image has relatively normal colour instead of the almost-sepia mess. Lastly, I used this Content-Aware technique to expand the width of the image.
The yellow processing on this original image of Marietta was my attempt to mimic a pale yellow style that photographer Clayton Cubitt had been using at the time. I failed. He did it so much better, of course. My attempt was much too strong. That being said, I don’t hate this image as much as the other originals I’ve showed you here. While it’s overly warm, it doesn’t look too dated to my eye. But there’s always room for improvement!
Click to enlarge
For 2016, the first thing I did was adjust the crop. The original had too much shoulder, too much arm, and too much bikini, consuming nearly two thirds of the photo with the head crammed at the top of the frame. The new crop eliminates some of this bottom portion, giving Marietta’s face more prominence in the frame. The yellow tone has been neutralized and the colours much more true to life. This was shot on an overcast day with some of the flattest light I’ve ever seen. While Marietta is a great model, the light wasn’t very interesting, so I originally added the yellow to try and jazz things up a bit. As you can see now, that wasn’t really necessary. An interesting portrait was still possible without resorting to colour trickery. I gave more attention to the eyes and lips in the new edit, and the overall sharpness and contrast were both increased to give this portrait some more snap. Definite improvement, and I don’t see myself going back to touch this one again.
So how has all this constant review and reflection helped me be a better photographer? For one, I’m making better selects now after a photo shoot. I’ve long advocated that photographers should not review, select, and start processing images immediately after a photo shoot. My usual workflow is that I’ll download and backup my images immediately after the shoot, then wait a few days before looking at them. When I do pore over them, I make some initial selections and then wait another day or two before refining my selections down to a smaller number. Only after that do I start processing. Even though this process is slower, I feel that it helps me get better images. By reviewing old photo shoots and selecting better photos from them, I’m more likely to pick the best ones the first time around for current shoots.
Secondly, I’ve learned not to go to such extremes when processing images. Moderation, subtlety, and conservative choices are better (for me) than wild, over-the-top expressions. This also fits in better with current processing trends, though it likely won’t be long before the pendulum swings the other way again.
Lastly, and related to the previous point, I shoot a little differently now. For the last little while, I’ve been routinely intentionally underexposing my images slightly to get the best possible raw file I can get (maximum detail in both shadows and highlights). When future software gets new features and abilities, my files will be in the prime position to take advantage and produce even better looking images.
How often do you review your past photographs and what have you learned from them? What steps do you take to make yourself a better photographer?