If Google’s Pixel line of phones is best known for anything, it’s impressive software that elevates the photos its cameras capture. This year’s Made By Google event, which kicks off at 6 p.m. ET on Wednesday, Aug. 12, will likely see the Pixel 11 phones shown off — a standard, Pro, and Pro XL version most likely — along with the next Google Pixel 11 Pro Fold. Rumors also point to a Pixel Watch 5. And everything will likely get a good dose of Gemini AI. Last year, Google added features to its Pixel phones that relied on Gemini generative AI technology like Pro Res Zoom to compensate for the limitations of small smartphone cameras. My big question is, with the Pixel 11 line being announced, will Google genuinely advance mobile photography or introduce more gimmicks that can frustrate the photo process for photographers and everyday people?
It’s a tough line to walk. Generative AI technology for images has advanced rapidly over that time, but so has the public’s resistance to AI slop and photos created entirely from prompts.
I want to believe Google can step up the photography game this year.
Pro Res Zoom technology vs. image quality #
Pro Res Zoom was the Pixel 10 Pro’s signature photo technology, a novel solution to an everyday smartphone camera problem.
It’s a common sight when people try to take photos with their phone: They zoom in on a scene by pinching and spreading with two fingers on the screen. But past a certain point — 10x zoom on the Pixel 10 Pro — all the camera is doing is digitally enlarging pixels, which diminishes detail. At 100x, images look like you’re squinting at them with poor eyesight and no glasses.
Google’s answer is to recreate, using generative AI, what’s in that zoomed-in area. The phone guesses at the content and generates an artificial rendering of what it should be. In some situations, it can work well, like with distant buildings or other architectural features that aren’t the focus of the shot. But if people are in the scene, forget about it.
Rumors suggest the Pixel 11 Pro phones’ version of Pro Res Zoom will go up to 120x magnification, so I don’t expect things to change much. However, the pace of generative AI improvements could mean it does a better job rendering the extremely zoomed-in scenes.
Ironically, Google’s earlier Super Res Zoom feature (which launched with the Pixel 3 in 2018) turns out to be the superior technology here, despite it being elbowed aside by Pro Res Zoom last year.
At up to 30x magnification, you still get the same problem from digitally enlarging pixels, but they’re not as large as at 100x. Super Res Zoom applies AI-based upscaling to sharpen details and fix many of the fuzzy artifacts created by digitally zooming. In my tests, it was difficult to tell which zoomed photos were Super Res Zoom images versus optically zoomed versions.
If Google can extend Super Res Zoom higher into the zoom range, even up to 50x zoom without resorting to generative AI reconstruction, I’ll call that a win.
Night Sight improvements #
Photographing in dark situations has been one area where smartphone cameras have excelled. If you’re ever able to view the Aurora Borealis, unless you’ve got a good camera and a tripod, your phone will give you far better handheld images of the light show in the sky.
Current phones do a good job with night photography, but we can always ask for better.
Dark photos tend to suffer from visible noise artifacts and issues with exposure and color. Too many smartphone cameras flex the ability to boost the exposure in a dark scene, but go too hard in artificially brightening a scene – I never want a night shot to look like it was exposed in the middle of the afternoon.
I’m optimistic, though, because AI technology for removing noise has improved a lot in the last couple of years. And if a camera can increase the lighting in a scene and still make the scene look realistic, then the developers can adjust the algorithms to turn down the exposure.
Photo editing with AI #
The goal of anyone taking a photo is to “get it right in-camera” and avoid spending more time editing it later. That’s really the entire driving force behind smartphone cameras: millions of people are capturing trillions of photos that will never be edited, so the camera needs to do as much work as possible to get it right the first time.
And quite often, it does. The typical person taking a snapshot with their phone is thinking about capturing a moment when they tap the button, not shutter speeds or exposure compensation.
But some photos do need editing, and smartphones have more than enough processing power to handle it. The hurdle is more often the complexity involved: not everyone is comfortable dragging sliders for brightness, contrast, white point, shadows and so on.
The Photos app does include an Ask feature, allowing you to tell the app what needs to be changed. If a photo looks washed out or has flat lighting, ask the app to make the fix. In my testing, it often does a good job.
What I’d like to see is the same feature adjusting all those sliders for me so I can then go back and tweak the settings if I want. Instead, I can only save the result as another copy of the image in my library.
A plea to lose the generative AI slop #
Whereas most of Google’s image-editing features that use AI build on the photo’s initial composition, there are plenty of generative AI features that enable users to radically alter them. I’m genuinely curious to know if people go for “remix” features that take a picture of someone and can change the background, restyle what they’re wearing, turn them into a sketch and so forth.
These features seem to exist not because people are clamoring for them, but because AI models can do them, and therefore they get worked into product releases to show the market and management that “we’re doing AI.” Maybe it’s the longtime photographer side of me that bristles at these types of generative creations (“glop?”).
It would make sense for Google, with all its investments in generative AI, to lean further into developing AI features that change images beyond recognition, simply to boast about their expanding toolbox. If I’m any judge, I don’t think anyone’s asking for that. Google has the resources and the talent to make meaningful advancements in smartphone photography. I’m hoping this year’s updates live up to that potential.