One of the problems of being a keen photographer is that it’s easy to slip into a mode where you treat every photograph as if it were a professional assignment. You can end up carrying around a lot of kit, and then spending a lot of time editing photos.
I now make a conscious effort to decide my objective in advance. Is it artistic – where the kit and the time will be justified by the result – or am I simply seeking to capture a memory of an event, aka a snapshot?
Even for proper photography, I rarely carry a DSLR these days: my trusty Sony a6300 copes with most things. But when I’m just after a snapshot, I try to keep things simple …
For snaps, I just use my iPhone SE – which lacks four capabilities but is otherwise a perfectly decent camera.
There are, of course, a huge number of iOS apps you could use to do work on iPhone shots, from Instagram through to the iOS version of Lightroom. But again, I try to remember that I’m simply aiming to capture a memory, so try to limit myself to a 10-30 second edit in the on-board Camera app.
But unless I’m uploading immediately to social media, I have a strong preference for working on a Mac rather than an iPhone, even if it’s just a quick edit. And that’s where Photolemur comes in.
Photolemur is a Mac app that uses AI to automatically edit photos – either individually or in batches. Instead of applying the same edits to every photo, or asking you to choose a filter, it analyses the content of each photo and attempts to perform the appropriate edits for each one. It does this with some mix of 12 adjustments:
- Color recovery, to boost the saturation of strong colors
- Sky enhancement, boosting blues and sharpening cloud patterns
- Exposure compensation, boosting shadows and recovering highlights
- Smart dehaze, boosting contrast and saturation to cut through fog or pollution
- Natural light correction, adjusting color temperatures to match the time of day
- Foliage enhancement, boosting greens and fall colors, and sharpening leaves
- Noise reduction, to remove the speckles often seen in low-light shots
- Tint perfection, adjusting color temperatures to suit the scene
- Face retouching, helps clear up blemishes (this feature is a work in progress)
- Horizon straightening, similar to the built-in camera function
- JPG fix, aims to clear up any artifacts caused by JPEG compression
- RAW processing – if you shoot in RAW, Photolemur will process in this format
The app started life with a subscription payment, which I personally dislike, but has now switched to a one-off purchase cost. At $30, the cost really depends on your basis for comparison. If you compare it to standard filter apps, which are usually free or cost just a few dollars, it’s expensive. If you compare it to full-on photo editing apps like Photoshop and Lightroom, it’s cheap.
Which is more appropriate? Let’s take a look at some examples. To make it a realistic test, I chose photos taken on the iPhone, all of which I would categorise as snapshots rather than photos. In other words, none of them took more than a few seconds to compose, and were all taken with the memory in mind rather than trying to create anything particularly artistic. This is what I see as the target market for such an app.
The first one I tried gave good results. Here’s the before:
And the after:
The original is a little underexposed, due to the bright sunlight on the carpet behind. Photolemur has boosted the exposure and upped the saturation to compensate for the backlight, but most notably has increased the contrast quite significantly. This really helps to bring out the texture in the fur. For a fully automatic process, I think this is a great result.
Next, I tried it on a photo of my partner. It’s quite a challenging shot to do anything with as it’s shot directly into the sun and thus has a blown-out background.
Here’s the after:
It’s done a good job of recovering the color in shadow area. Check out the right side of the scarf and the wall to the right, for example. It’s also done as much as could be done with the highlights – look at where the sun filters through the hair, for example, and the blooming has been removed. In Steph’s face, the skin tone has been warmed and teeth whitened. It’s also very subtly narrowed the face, which some women might appreciate but is slightly detrimental here. Finally, it’s boosted contrast to the building being constructed bottom-left – though a downside of this is that the rain-streaks on the window are now more prominent.
Let’s try it with another shot into the light.
Here, the most prominent change made has been an exposure boost, which brings out more of the color in the pink champagne, and also a little more detail in the buildings outside. However, a downside is that the random stranger to the right, previously rather hidden in darkness, is now a more prominent element of the shot. This is a good illustration of the difference between an automatic and manual edit, where you would really want to selectively adjust compensation.
So far, so good. One really good result, one worthwhile improvement and one that gave mixed results but works well for the actual focus of the shot.
However, I found the app didn’t always perform so well. Here’s an example of where it delivered very mixed results. It was taken late in the day, so there was a lovely yellow-ish quality to the light as it headed toward sunset. (You may be detecting a slight theme to these snaps …)
Photolemus has adjusted the white balance to almost completely remove this light.
On the plus side, the contrast boost has brought out the texture in the label, and emphasised the droplets, which works well. But overall, I’d say this is worse than the original.
Finally, here’s an example where there’s no question: Photolemur has significantly degraded the image. First the original (taken at an SF props exhibition in London):
Now, it’s not a great shot – the Darth Vader helmet is rather lost in the darkness. Photolemur has obviously noticed this and boosted the exposure.
But doing so has brought into view lots of distracting elements inside the case, and especially the reflection in the glass. That’s an objectively worst version of the shot.
Conclusions
Photolemur certainly doesn’t replace conventional photo-editing software like Lightroom – but then it’s not trying to. It’s aiming to make quick, automated improvements to snapshots. More often than not it succeeds, but not always. I suspect it does best with the kind of landscape shots you mostly see on the Photolemur website.
I can’t claim to have tested it with enough photos to give any kind of statistical guide to its hit-rate, but it certainly made worthwhile improvements to the majority. Others were more hit-and-miss, while some did more harm than good.
This, then, is the problem with the current state of AI technology when it comes to apps like this: they aren’t yet as intelligent as we’d like them to be.
Could I recommend it? It’s tricky. If $30 is a trivial amount to you, then maybe it’s worth getting so you can suck it and see as you use it. Run your snaps through it and compare before-and-after to see which ones you prefer. It’s my bet that you’ll prefer the after versions most of the time.
But others may take the view that you can buy a lot of iOS apps for $30, and there are a few powerful Mac apps at this price level too. Finally, even if you view the cost as ok, you might say that having to take the time to compare before-and-after shots defeats the point of an app that’s all about automated ease.
The good news is that you can download a free trial with no batch feature and a watermark. Try that, and you can decide for yourself.
Photolemur costs $30 from the Mac App Store, or direct from the developer. You can download a free trial version here. A family license for up to five devices is also available.
