Apple and Samsung typically show off glitzy hardware innovations — dual-lens cameras; bigger screens — to mesmerize millions into buying their new smartphones. Google has taken a different approach: It wants to sell people on better software.
That was abundantly clear with the Pixel 3, the search giant’s latest smartphone, with software features powered by artificial intelligence and machine learning. When you take a blurry photo, for example, the Pixel’s camera software can capture a series of alternate images and look for a clearer shot. When you get a call from an unknown number, you can use a screening tool to determine if it is a robocall. (I recommend using this feature with caution — more on that later.)
“The phone is powered by A.I. so that we let Pixel do really useful, delightful things,” said Mario Queiroz, Google’s vice president of Pixel hardware. “The phone doesn’t get in the way. Things just happen for you.”
The Pixel 3, which will be available on Thursday, is a modest piece of hardware otherwise. It comes in two screen sizes — 5.5 inches and 6.3 inches — that are slightly smaller than phones from Apple and Samsung. The smaller Pixel costs $799; the larger one is $899.
The screen technology is OLED, which offers better color accuracy and contrast than its predecessor, LCD. The rear camera has a single lens, because Google felt its A.I.-powered camera software was so good that it didn’t need to add a second one. Like other high-end phones, the Pixel 3 is water-resistant, and its body is composed of glass to support wireless power charging, a feature that I have found marginally useful.
After testing the Pixel 3 for four days, I’m convinced that Google did the right thing by emphasizing its software. The Pixel 3 is a superior Android phone to Samsung’s top-rated Galaxy devices thanks largely to Google’s clever camera and intuitive interface.
But I wouldn’t say that the new Pixel has the all-around best camera — I prefer the iPhone camera’s more lifelike colors. Still, the Pixel 3’s software smarts offer some advantages, like the ability to take superior photos with the bokeh effect, which puts the picture’s main subject in sharp focus while gently blurring the background.
Here’s what I found about the Pixel 3’s most important features.
Let’s say you take a photo of a group of people and one person blinks. Using Google’s new photo-optimization tool, Top Shot, there’s a chance you can salvage that photo by choosing a better shot with that person’s eyes open.
To use Top Shot, you turn on a feature called Motion Photos, which are essentially three-second videos capturing the moments just before and after you take a shot. When you snap a picture, Google’s camera software can analyze those before-and-after moments for a better image to recommend.
Google’s theory is that Top Shot will change the way people take photos: Instead of snapping several shots of the same subject and manually picking the best one, users could instead shoot a single motion photo and then let Google suggest the best image.
For me, Top Shot didn’t profoundly improve my photos or change my habits. For example, in a park, I still took a half dozen photos of my dogs until they both smiled at the camera. But when I reviewed my favorite photo, Google recommended two different shots as a potential Top Shot — and I did pick the one where the dogs’ faces looked clear.
It’s impressive that Google is capable of detecting the qualities of a good photo, but I found the process of reviewing and selecting the recommended shots to be time consuming. It was quicker to use the old technique of snapping a bunch of photos and manually swiping through them to decide which you like most.
In other words: For now, I don’t consider this a must-have feature.
A better reason to consider the Pixel 3 is its impressive ability to take so-called portrait mode photos, those DSLR-like images with a gently blurred background and sharpened foreground.
This effect is useful for producing artistic-looking photos of people, food and objects. Portrait mode is the Pixel 3’s strength: Compared with both Samsung’s Galaxy phones and Apple’s high-end iPhones, which use dual-lens cameras to take portrait mode photos, the single-lens Pixel 3 camera was more reliable at sharpening the intended subject.
Here’s an example: When I had dim sum for lunch and used the Pixel 3’s portrait mode to snap a photo of five soup dumplings inside a basket, the camera appropriately sharpened the five dumplings while blurring out my partner in the background. When I took a portrait mode photo with the iPhone XS, the camera blurred out one of the five dumplings. I eventually got the iPhone camera to sharpen all five dumplings with a few more tries, but the Pixel 3 did it in one attempt.
It’s remarkable that the Pixel 3 can produce these images with a single lens. Google accomplished this largely with what is known as machine learning, which involved computers analyzing millions of images to recognize what’s important in a photo. The Pixel 3 makes predictions about the parts of the photo that should stay sharp and creates a mask around it. A special sensor inside the camera, called dual-pixel autofocus, helps analyze the distance between the objects and the camera to make the blurring look realistic.
Having this smart camera — which has the potential to become even smarter through software updates — is a compelling reason to consider the Pixel 3. Google said it planned to add more camera features in those updates, like Night Sight, which helps people take photos in the dark without using a flash.
“What we’ve been able to do with machine learning to produce image quality in photography and video, it really surpasses the need for us to do more,” Mr. Queiroz said.
Nowadays, it can feel daunting to take a phone call from an unfamiliar number because it is most likely a robocall — a scammer or spammer using an automated dialing system to steal your personal information or to sell you something. To fight robocalls, Google built a screening tool into the Pixel 3.
Beware: Using this feature comes with some risk.
The screening tool works by tapping a button when you want to screen a call. An automated message will ask the caller to state his or her name and purpose for calling. From there, the user can receive a transcription of the caller’s response and decide whether to take the call.
The tool works well: I screened a call from my partner, and Google did a perfect job at immediately transcribing her response (“My purpose is to eat dumplings, so leave me alone”).
But be wary about using this tool. The No. 1 rule when dealing with robocalls is to not answer them, according to the Federal Communications Commission. The No. 2 rule is to hang up immediately if you do answer a robocall. That’s because picking up the phone or engaging with the caller indicates that your phone number is active, which could land you on a list to be flooded with even more robocalls.
It’s unclear if using the tool would count as engaging with the robocaller. Google said that in its testing, it did not see an increase in spam calls after the screening feature was used. The company said the alternative — to ignore all unknown numbers — was not feasible because calls could be coming in from doctors, delivery drivers, restaurants and service providers.
It remains to be seen how effective the screening tool will be when Pixel customers use it en masse. I still think the safer — although imperfect — solution is to not answer calls from suspicious numbers, like those that are nearly identical to yours.
After all the tests, here’s the short version of who should buy a Pixel: You should get one if what you care about most is having a phone with a great camera that does the best job of running Google’s Android software.
But I don’t recommend messing with the robocallers. The bots will always win.