Keith's Odyssey to Planet Fitness: A bit of a conundrum

Some of you knew that I was signed up for a Lightroom course this weekend. (For the non-photographers, Lightroom is a complex program used to edit photos, similar to Photoshop in some ways.) I took a basic Lightroom course from Neil Zeller about 2017, and later watched a few tutorials that were not as helpful as one would like. I will exit out of such a thing if the instructor has promised to show the steps to get from A to N, and at step E just they're getting into the meat of the matter, they start talking about alternate ways of getting to F and G, or that since you're here it's just as easy to go to Q, which I have no intention or desire to do.

It's infuriating. I want them to show all the promised steps in order, no less and no more, and only then talk about some of the alternate ways of doing it, because most bloatware has altogether too many ways of doing the same thing. There is almost certainly a nonintuitive "don't do that!" during the sequence of steps, and it's all too rare they talk about that. Often there's an assumption they've made that they haven't told you about, that they assume you've made because isn't it obvious.

That could be a whole separate rant but we won't go there today.

As it turned out for complicated reasons, it worked better for Neil to spend a day with me and my mothership for one on one coaching. It was enormously productive. We reviewed my Lightroom settings and preferences, and how my workflow treated files and folders. All good because I'd heeded his advice from 2017. Then I got signed up for BackBlaze for off site backup, and started that. The internal drive was fast since there really isn't all that much on it. A day later and it's not even half done the external drive. (You are backing up your computer, right? RIGHT??)

From there we sort of covered two paths. One is making individual images better by learning to use the existing tools that I either didn't know how to use or was unaware of, and the new tools in a recent upgrade. Let's just say there are a lot possibilities and it's easy to go wrong.

The other path was tools and techniques for dealing with hundreds, and sometimes thousands of event photos in a batch, which most photographers never do. One of the new tools is totally the cat's meow for that sort of work, and I'm not talking about the AI culling thingie.

We picked out a variety of images and explored the tools. Landscapes, portraits, events, nights, reflections, some with complicated problems like odd lighting or dust spots on the lens or sensor. Some of it was the steps needed to do a thing, sometimes the philosophy of better images, and some of it was the hands-on practice for me to build some muscle memory. I've been playing with some other images to try different things on images that will never be seen. There were lots of virtual copies and undo keystrokes along the way. Very little swearing, in case you were wondering.

The conundrum is we tweaked some images that have already been blogged, and yet they're so much better now. I'm likely to do more of them, just because they're good images and can be better. Some of the comments have suggested improvements, some of which I was incapable of, and might be now.

Maybe they'll show up in the 'why didn't I blog this and maybe I did' section. Or maybe I'll blog them and see if anyone notices they're a repeat. For sure images that I want to put in another book are going to get another pass. Maybe I'll add a new section to of the day, called reimagined. Hmmm.

I still struggle with the transition from photograph to digital image. How much tweaking to an image is fair game? Sean and I have discussed this extensively during our road trips. I have no firm set of rules to follow. One of the strong suggestions is that if someone can tell you've made the adjustment, it's probably too much.

I for sure won't put in something that wasn't there. Unless I want to make it look like an alien landscape for some reason, I want it to be pretty much what I saw, and what another person would see if they go to that spot and made allowances for weather. Images tweaked, tuned, and gently massaged, but not over baked. I look at some of my former edits and wince. I don't want to smooth skin to the point the person looks plastic.

Removing distractions is a slippery slope. There needs to be something around the thing to give context, but there's a spectrum of backgrounds from pure white or black, to a bland wallpaper, so an interesting pattern, to the perfect background complementing the thing, to distracting, to so wild and crazy it becomes the thing.

Here's some before and after photos, some of which you've seen. Keep in mind the before image is actually after I had edited it the first time. You don't want to see the raw image.

In the pairs of photos, A is after, B is before.

1A.

1B. The major difference is removing the distraction of the airplane in the upper right, and opening up the shadows in the lower quarter, and tweaking some of the other settings.

2A.

2B. I was was pretty pleased with this image, but knew it could be better. With some coaching on using the tools, I did a much better job masking and tweaking the two bits of land. Denoise was awesome. Plus removing some distractions, changing the aspect ratio, and a few other tweaks.

3A. I'm not sure if this image was ever blogged. The differences should be obvious.

3B.

4A.

4B. Cleaned up some, but not all of the gunge in the water. I still want the scene to look real, and we all know that water is never perfect. Tweaked colour balance a bit and worked on the sky and it's reflection in the water. Plus making it level. Can't imagine how I overlooked that. This is a big image, about 5' x 1.5' at 300 dpi. There are still tweaks that could be made, and I would if someone wanted to buy this image, but the danger is that it's a rabbit hole with no bottom.

5A.

5B. Mountain landscapes can be surprisingly tricky. They can be hazy and turn blue, with the sky blending into the mountains. Lots of masking and tweaks to treat the sky, mountains, foothills, and fields separately. Plus removing a distraction.

6A. I have a certain amount of skepticism about AI. A lot actually. Lots of people are trying to use AI for things because they don't want to pay a person to do them. Except people are better, albeit not as fast. People writing computer code with AI are building time bombs, unless the right person reviews it very carefully. AI generated stuff often looks plausible at first glance.

I worry that an AI will wake up and be smarter than Skynet. And we thought that viruses and rats were tough competitors. I'd never do this to a race image, but this is purely the AI engine finding and removing distractions. It removed the person on the path, the person holding the bike, then another pass removed the bike. I suppose if I were trying to make this a perfect image, I'd try to remove the trash bag and see how that looked. Then again, these are working people during an actual race, not models posing for a magazine cover. Plus running the removal process is typically 30 seconds per image for my M4 Macbook Pro. Neil's souped up mothership does it in 5 to 10 seconds. That adds up for 500 images to the point it's completely uneconomic.

6B.

I had been thinking that Laverty Falls would be next blog, so they're the first images I looked at after Neil went home. Let's just say I looked at the images with new eyes, and made some changes. Stay tuned.

Of the Day

Driftwood (NZ)