These Lightroom Classic Updates Prove AI Can Really Help Artists

May 12, 2023

Generative AI and the shady ethics of training software on copyrighted images have turned artificial intelligence into a curse word among many photography communities. But, if there’s one way the technology can help rather than hurt artists, it’s this: making tedious, non-creative tasks drain less time and energy. That’s what Adobe is doing with its latest round of Lightroom Classic Updates. The April 2023 Lightroom Classic updates focus on two areas: reducing noise and smarter masking. Both are powered by artificial intelligence.

The smart masking tools that Adobe has added to Lightroom over the last year have been a game-changer for my editing process. It’s shaved hours off my editing time as a wedding and portrait photographer. That’s because Lightroom Classic uses AI in the right way — it masks out specific areas but still allows me to retain control over the effect. For example, the whitening teeth adaptive preset goes overboard to a blinding toothpaste white. But, I can use the Select People, check the teeth box and ensure the warmth I added to the overall photograph doesn’t create distracting yellow teeth. The same applies for skin softening — I can soften the skin so it’s not distracting, yet it still has some texture.

Not only does the smart masking save time on individual photos, but it works remarkably well to apply the same effect across multiple images. Batch editing with local masks is an incredible time saver. And now, those tools have new masking options to select facial hair and clothing.

The April 2023 update also brings a Denoise tool that I think many photographers will like. It creates some really nice results. But, sadly, it’s a destructive edit that can’t easily be undone and isn’t compatible with batch editing.

We’ve added the following to our Lightroom CC review:

Smart Masks

Nestled among the tools like crop and healing are the brush tools. While this used to only let you hand select an item or use a radial or gradient to apply local adjustments, the brush tool has seen major changes over the last few versions. All of them immensely increase how simple it is to make adjustments that apply only to one part of the photo. You can now select the subject, the background, or the sky easily. For portraits, you can select an entire person or parts such as eyes, skin, and hair. The program will differentiate between multiple people too.

The AI-powered options can also be used when batch editing very different photos. For example, I can use the brush tool to whiten teeth and then apply those settings to a group of images with the AI automatically finding the teeth for me. This is a huge feature for me when editing hundreds of photos from a wedding day. It saves so much time compared to re-doing each mask by hand for every single image.

The latest tools to come to these local edits are the ability to select clothes and facial hair. This makes it simpler to say, change the colors of a shirt or darken a mustache. The clothing masking seemed pretty accurate, though it selects the entire outfit and not just the shirt or the pants — the option to differentiate between types of clothing would be a nice thing to see in later updates. The facial hair masking was less specific, selecting a bit of the skin too. That could be a problem if you adjust things like saturation, but it didn’t look odd with a simple darkening.

This suit used to be brown. But, I would have had to manually subtract the white shirt to only change the jacket and pants.

The list of adjustments that can be made to a mask isn’t as complete as the adjustments that can be applied to the whole photograph. But, Adobe just added the option to adjust Curves within a mask. This is a great tool to use in conjunction with the select sky.

Lightroom Clas

While Smart Masks have been a game changer, there is one big addition I think Adobe needs to add: HSL color tools inside of the masking edits. When I remove skin redness, I often don’t want to remove the redness from the lips or clothing. If I had the HSL tools inside one of those smart masks, that would be very easy to do. The ability to adjust specific colors within that mask would be great, as the current local color options are limited.

Also, I would cry tears of joy if I could lighten the lenses of transition glasses in batch editing. This used to be a far-off dream, but if the software can recognize clothing, maybe masking the glasses is next.

DENOISE

One of the highlights of the April 2023 update is an AI-based Denoise algorithm. The same noise sliders still exist tucked into a new Manual Noise Reduction sub-section. Those longstanding sliders are still the fastest way to edit noise, making the tools still helpful for mid-level ISOs. The highest ISOs, however, can benefit from the new DeNoise. However, the lack of a few key features — like bulk editing — makes the tool feel a little overhyped.

The Denoise tool opens a pop-up window where editors can use the slider to preview the effect before clicking OK. Like AI-based masking, the Denoise tool can take a few minutes. It also creates a new DNG tool rather than replacing the original file. This is because the Denoise tool isn’t non-destructive like most of Lightroom’s tools. Translation: You can’t undo the Denoise. You can go back to the original version of the photo, but you’d have to sync or re-do any edits made after the Denoise algorithm. Editors can compensate for that by using the Denoise last. But, that still means going back and adjusting the amount of the Denoise will be annoying. Sadly, the new tool also isn’t compatible with batch editing.

Original
Edited with Denoise
Original (Cropped)
Edited (Cropped)

Overall, I think the Denoise tool did a good job of smoothing out noise with less softening than the original sliders. It will still smooth out some detail, but the slider lets you choose how much to preserve. It’s a great tool, but it’s not earth-shattering. I




Source link

Comments 0

Leave a Reply

Your email address will not be published. Required fields are marked *