So, through a series of related events, I came to be watching a live training session today on processing wildlife photos. It was a great session and I learned some useful techniques.
The techniques were taught in Lightroom Classic. Now, I can cope with some different sliders between PL and LR. I know there are some differences and in some cases I prefer the PL approach, in others the LR one.
But… significant use was made of “Subject select” and “Background select” and “Object select” and… more importantly, intersecting masks.
I have argued in the past that most of my masking needs can be met with Control Points and the occasional Control Line. There are times, however, where even the magic of Control Points can be a chore. Watching the simple act of “select background, intersect with linear gradient” was, however, hurting my head on how I would reliably and repeatedly do this in PL.
To be fair, when practicing the techniques on two photographs, I discovered “select subject” (and therefore “select background”) got it wrong in a significant way on one of them, but the other LR mask-adjusting tools made it somewhat trivial to fix.
I’ve also criticised LR’s auto-masking options but, again, in the two use cases I have tried today, it did a great or passable job. Ironically, considering my prior criticism was based on giving it easier photos to work with.
It got me to wondering whether I might end up just using PL as an inefficient version of PureRaw. Because, as clever as auto-masking is, nothing holds a candle to DxO modules when it comes to noise reduction and lens correction.
The content based masking is about the only thing I miss since I ditched Lightroom for DXO Photolab. And the other thing that annoys me is when I accidently select multiple photos and change them all with no way back…
Abandoning PhotoLab for Lightroom feels a little icky. I started with PL3 and have been extolling its virtues ever since. I’ve owned every version, and those of ViewPoint and FilmPack.
Depending on what shows up in PL9 (of which I have no idea since I seem to be out of the loop on beta testing), I could be looking at PureRAW 6 instead.
Your thoughts on PL are reasonable. I and I am sure others have requested content-specific selections. I recently returned from two weeks in the Veneto and I tried editing in PL8, but selecting sky, city skyline, or specific objects was time consuming to get the look I wanted, but doable. I saved the PL files as DNGs after NR, switched to CO, and was able to quickly select what I needed and fine tuned the masks with three erase brush options. I use Apple’s Cleanup for the rare object removal task. LR is a likely a better one-stop option and I can understand why you liked using it. Let’s see what PL9 brings as they seem to have NR sorted.
I agree it works well when selecting a mostly uniform sky then inverting. The process becomes more complicated (at least for my skill level) when there are multiple colors and luminosities. This is where CO’s “AI” or magic brush selection/erase saves time. Picking shapes was better with LR if I correctly recall my brief time with the program.
I agree, but I don’t always get a natural-looking blend in the way I can get with CO. I like DXO PL because it is like Capture NX, which I used when I shot Nikon so I stick with it. But this was before “AI selection”. Tech changes. IMO, DXO needs to go beyond lens correction and NR or PR and NIK are their future. That’s OK and that might be their business model. In that case, I’ll use PR with CO, but I’d like to only use PL. Looking forward to PL 9’s offerings.
Andy Hutchinson recently did a YouTube video discussing issues in this thread. Andy, like me, uses both Lightroom Classic and DxO PhotoLab. And yes, Adobe LrC and Photoshop, and others including Luminar Neo and ON1 Photo Raw, can be very good at masking for specific elements. It often requires some manual refinement, but it can save a lot of time in selecting elements.
But, Andy argues convincingly, it doesn’t always produce the best edits. If you have a sunrise-lit sky, select the sky, and edit the color, then invert the selection and adjust the non-sky color differently, it may make for a dramatic but very unrealistic scene, because light and color don’t separate out that way at the horizon; they blend across the horizon. Linear gradients can work much better at blending color between diverse elements.
People are drawn to new powerful features. Sometimes the older way is better.
I like Andy because he always gives context. I suspect I am missing that context in your report (I’ll watch the video later) because, while a linear gradient will work well for a sunset, it’s pretty terrible for isolating a bird from the background. Which is what I’m primarily using it for.
The author offers valuable advice (or ideas) on the tension between AI and traditional gradient masks in this case. I’m not a landscape photographer, but I think the tips can be generalized to some extent.
Each product adds features every now and then. They need to do it in order to stay ahead of or with the pack. Kind of self-fulfilling bet in the race of commerce.
As new tool emerge, we can test them and find out in which case they provide easiest access to realising our intentions.
Multiply the number of cases by the number of intentions and get something that throws absolute terms into the marketing waste basket.
The horrid part in all of this is in regular and frequent switching rather than in the fact that new tools emerge, taking whatever app to the top while other apps silently work to replace whatever is on top.
…while a linear gradient will work well for a sunset, it’s pretty terrible for isolating a bird from the background. Which is what I’m primarily using it for.
Sure, and I use Lightroom masking to do that too, isolate a bird. But this video’s about landscape rather than wildlife images.
Part of what’s driving it is that Lightroom Classic just came out with a Landscape masking tool which allows the editor to quickly select specific elements like sky, water, natural ground, etc. and easily apply separate edits to any or all of them. Adobe is using AI to power this masking tool, and for some images, its selections are pretty accurate; for others especially where elements overlap, like leaves over sky as seen in Andy’s thumbnail above, or where different elements are close in tone or color, the accuracy may miss and non-AI selection tools may work better.
Sure. But using PL’s tools to select a bird is no picnic either. Especially if both bird and background have multiple colours, even some shared. Use a brush? Some plumage is a nightmare to get to the edges of.
In the end, except for the very simplest cases, the AI approach is going to be much quicker and easier…If it works, of course.
