Is Luminar 4 the future of photography?

by Jan 13, 2020

I’ve been thinking for a while now about what my future photography workflow will be. Am I going to continue to tie myself to Lightroom? Or is there something else out there I should be thinking about moving to?

I kicked the tires on Capture One since so many photographers I know swear by it. I think it’s quite a good program, but I hated the UI and I could simply not see myself using it for any length of time. I experimented with Luminar 3 and it simply didn’t seem better than Lightroom, and it had no DAM (Data Access Manager) to speak of. So when Luminar 4 was announced, I was intrigued, because it looked like the company was going all in on AI and computational photography. Was that all hype, or was there substance behind it?

When they released Luminar 4, I grabbed a copy and installed it as a plug-in to Lightroom, and since then I’ve been experimenting on images with it, sometimes on older images, sometimes processing an image in parallel both with Lightroom to compare the outcomes side by side.

I’ve been meaning to write about this, but with holidays and all, I haven’t gotten to it, but Jeff Carlson and Kirk McElhearn’s latest episode of PhotoActive Podcast is all about Luminar — and Jeff is writing a book for Rocky Nook on it, so he’s been using it a lot through the beta process. It’s a good discussion on the program and gives some nice background on it if you haven’t read anything on it yet, and I recommend it to anyone possibly interested in the program or who are curious about this whole “AI processing of your image” concept. That reminded me that I had meant to write about my experiences with Luminar 4.

The future of photography is computational

There’s a lot of chatter going around with the term computational photography, but it may not be clear what we mean by that. While an argument could be made that any kind of post processing work done on a digital image is computational photography (and on a purely technical basis, be correct), what we tend to call computational photography are software tools that allow us to create images that aren’t possible in camera. In the early days of digital photography limited sensor dynamic range often meant we needed to do HDR processing to combine multiple images to get a usable image. Stitching images into panoramas, or using focus stacking techniques to get extended depth of field are other fairly common techniques we use today.

What has us talking about computational photography, though, are some of the newer and more advanced techniques that have been emerging in recent years. Apple’s Portrait mode, which creates an artificial boken around a subject, or the new low-light night modes in both IOS and Android are two examples of the kind of thing that can only be accomplished with intense software algorithms. That type of processing is now being brought into “mainstream” photography via Luminar 4 and its AI tools, in a way intended to make high quality processing of images easy for a non-expert photographer.

In my testing of Luminar, the improvements you can get on an image can at times be stunning, when “all” you are doing is moving a slider or three, is amazing. As someone who remembers the early days of digital photography when processing meant loading up Photoshop and using layer masks and a lot of brush work and filters and … and …

And today, what used to take hours in Photoshop takes five or ten minutes in a tool like Luminar.

One has to believe these kinds of tools are the future, whether or not you like the idea of someone who hasn’t sweated over layer masks and sharpening protocols and noise reduction and all of those other things we’ve had to learn to get the most out of our images.

I have an image I love to use for these examples, because I love the image, and because processing the image is in general a hot mess because technically, it’s pretty messed up. I originally took it in 2011 and fell in love with it, and it’s been one of my most popular images over time. I’m including three versions of the image here for your amusement.

Morro Bay Rock sits on the harbor during sunset, Morro Bay, California. Processed by Photoshop as a 3 image HDR

morro bay harbor at sunset, California. Processed by Lightroom.

morro bay harbor at sunset, California. Processed by Luminar 4.

The first is the processing I did in 2011, and because it’s 2011, that included a three image HDR merge, and about 3 hours of cleanup and tweaking in photoshop until I was happy with it. Looking at it with my eyes today, Morro Rock is horribly over-processed and over-sharpened, to name just the most obvious flaw, but it is, honestly, very 2011 era digital photography.

The second was when I decided I didn’t ever want to see that over-processed rock again and took it down to the original raw image and processed it from scratch, which I did about six months ago. That was done entirely with one image — no HDR — and in Lightroom, and uses four radial filters and a gradient filter, and it took me about 40 minutes to process. The final one I did for this article using Luminar 4, using only the standard tools. There is no brush work, no masks, no local adjustments. It’s all global adjustments. No, seriously. And it took me about ten minutes.

That Luminar-processed images needs some more work, but I’d argue it’s better than my 2011 version, and I am still a beginner at understanding how to get the best out of Luminar. I also want you to compare the skies in the first two images at the blown out sun going behind the rock. Now, look at this new image and, um, the blown out sun is gone. Luminar fixed that, in what I feel is a very pleasing way. When I noticed that I was rather blown away, because I didn’t do that, the program did. It’s kept the detail in the sky well. I think the detail in the rock is better than the second image, and the lighting across it is a lot more even. There’s a lot to like in this image. It still needs work — the rock has a rather unpleasant green cast, which I see in a much reduced way in the 2nd image as well, and the fog banks behind and to the sides of the rock have degenerated into noisy blotches — but that’s all fixable. (and neither are new problems to Luminar, and the source of much cussing and head scratching in my past).

So in about ten minutes, with very shallow knowledge of Luminar, I was able to reproduce an image that took me 3 hours to wrangle in Photoshop in 2011, and Luminar was able to take a rather flawed raw image and bend it to my will in effectively no time.

Am I impressed with Luminar? Absolutely. Does it successfully take complex processing and make it as simple as pushing and pulling a few levers? Yes. Is this the future of digital darkroom processing? Yes. Am I going to adopt Luminar 4 as my digital darkroom tool in the future?

Well…

Sean Tucker recently did a Youtube video titled To Edit or Not Edit your Photography? All about what is and isn’t appropriate in post processing. And hint: his opinion matches mine, which is whatever you feel is comfortable and appropriate to do to an image is fine, all all of those people who are trying to tell you an image has to be processed the way they want it processed are wrong.

I bring this up to bring up a point about Luminar 4 that I’m pondering and I don’t really have a good answer for: In the image example above, I specifically worked to reproduce the image in Luminar in a way that looked like my previous versions of the image. But when I’m working on new images for the first time, I find myself conflicted with Luminar a bit.

Am I processing the image the way I want it to look, or is Luminar guiding me to the image it wants me to create? Is this my image or is this Luminars? The AI capabilities are powerful and impressive, but I keep tripping up on a philosophical question: is that really my image, or is it the image the AI wants me to have? And I don’t have a good answer for that, and it’s keeping me from using Luminar more for the moment.

This is not the “image must be right in camera” argument that Sean destroys so wonderfully. It’s more fundamental: when I’m processing an image in Lightroom or Photoshop, the job isn’t done until I have the image the way I want it. With Luminar 4, this is more ambiguous to me: am I processing the image until I have the image I want? Or until the AI has given me its interpretation of the image and I accept it.

This may seem a silly argument to have, and if so, feel free to ignore me and adopt Luminar happily. But one worry I have with the AI tools is that they are going to encourage us to all process images towards an “accepted” definition of quality, not a personal one. And that’s a big problem for me to get past.

I have some other issues with Luminar, more logistics based. The main one: If I want to migrate away from Lightroom, I have to stop using Lightroom as a DAM as well as a processing tool. Luminar’s DAM is, well, not up to what I need for a DAM to be comfortable leaving Lightroom. This means my use of Luminar will be as a plug-in to Lightroom. doing so requires converting an image to TIFF for processing, and having multiple copies of an image in the library, and all of the added complexity and chaos in the workflow this creates, and I’m not yet convinced that added workflow complexity is worth the added value Luminar gives me.

It’s also totally unclear to me what a possible Mac and IOS workflow for image processing could be with Luminar, and any alternative to Lightroom I might adopt I’m going to have to understand what that can or will be. right now, the IOS side of Luminar is nothing at all. Adobe Lightroom (vs my normal tool Adobe Lightroom Classic) may not be great at this yet, but the roadmap and the integrations between it and Lightroom Classic seem clear to me as a roadmap, and until I see something better, it’s hard to shift away from these tools until that’s part of my solution as well. Like Adobe or not, they’re further along at this cross platform integration, and better at it, than anyone else. (Please feel free to correct me if you think you can).

Which is kind of my second challenge with Luminar: I’m pretty good with Lightroom these days, I think, although I keep learning new tricks to keep getting better. And while I can get really good images out of Luminar without the thousands of hours of experience I have with Adobe’s tools, I’m not entirely convinced they’re better images. Most seem to be as good. I can do them faster in Luminar (most of the time), but Luminar doesn’t seem to be giving me a huge improvement in the quality of my images, just a different take on them of equivalent quality. Your mileage with your images will definitely be different, I’m sure.

My Bottom Line

I think Luminar’s work is impressive, and it does really interesting and good things to images. I think if you’re at all curious about its possibilities, you should download a copy of the demo and push some of your images through it. You may find you love it.

That said, Luminar will continue to be in evaluation mode for me for the foreseeable future, because it’s not what I’m specifically looking for, which is a full replacement for Lightroom. That saddens me a bit, and if the continue to innovate the DAM parts of the tool the way they have the processing engine, there’s hope — and to be honest, nobody else is doing anything I feel lets me replace Lightroom either.

And I’m really curious about that philosophical question I raised, whether or not the end product is the image I envisions, or whether it’s Luminar’s AI vision based on the instructions I gave it. Because there’s a bigger issue in that thought, because at the extreme end of that are the growing problem of deep fakes and synthesized images, and I’m just not sure right now where I fall in the spectrum in terms of what’s acceptable to me.

You may well feel entirely different — and that’s fine. I also think the less skilled you are in post processing, or the less serious you’ve been in creating “pro quality” type imagery, the more Luminar will be a wonderful tool for you. Part of me things the best thing that could happen here is for Apple to buy Luminar and add the capabilities of Luminar 4 into the Photos app. I think the result could be incredible and I’d be all over that with my iPhone images, and perhaps all my images.

So parts of my love what I’m seeing in Luminar and what it says about what the digital imagery universe is going to look like over the next five years. Other parts of me aren’t so sure — but for reasons that may not actually exist, and will certainly not be relevant for most users of Luminar. And my take for you: if you’re curious about Luminar as a replacement or supplement to your current digital darkroom tools, grab a demo and give it a try. The learning curve is a lot lower than Lightroom, and using it seems to be a lot simpler in general, and so I think it could be a big improvement in their final image quality for a lot of people.

And I do hope I see improvements to the DAM to make my personal decisions on this easier down the road…