Google Pixel 6 with its own Tensor Processor

Google Pixel 6 with its own Tensor Processor
Google Pixel 6 with its own Tensor Processor

The Pixel 6 is priced from $599 and the Pixel 6 Pro is priced from $899 and will be available on October 28.


The Pixel 6 comes with 8GB of RAM, a 6.4-inch OLED screen that supports a 90Hz screen update rate, an 8MP front lens, a matte aluminum case, and polished aluminum bezels in Sea Foam, Coral Pink, and Storm Black.

The Pixel 6 Pro comes with 12GB of RAM, a 6.7-inch OLED dual-curved screen with narrower bezels, 120Hz screen refresh rate, 11.1 megapixels front lens, and a chip that supports ultra-wideband technology (UWB) for more accurate positioning, and for the first time, the Pixel 6 Pro comes with a triple rear lens and a periscopic telephoto with 4x optical zoom. In addition, the Pixel 6 takes away the 1/2.55-inch image sensor (Sony IMX363) that has been used for three generations and replaces it with a large base image sensor, the same matte aluminum housing, and polished aluminum frame, and colors such as Sunshine Yellow, Cloudy White, and Stormy Black.


The Pixel 6 series features a glass body and color-blocking design that fits well with the Material You UI pre-installed with Android 12. The front lens features the popular center single perforated lens design with a matte black camera bar on the back, while the rear camera module features a 50-megapixel wide-angle lens with 150 percent more light intake and a 12-megapixel ultra wide-angle lens. Both phones can be quickly charged to 50% in 30 minutes, but Google has said it will no longer provide a charger, so it will need to be purchased separately.


Self-developed Tensor chip enters the phone

Apple just abandoned Intel, Google will abandon Qualcomm? Can the artificial intelligence-focused TPU (tensor processing unit), and that custom hardware, help Google stand out from its competitors? This is also a key point of Google's current launch.


Google calls Tensor "a milestone in machine learning" and "the biggest mobile hardware innovation in the company's history", a chip designed by Google Research Institute over four years, based on the past 20 years of computing experience and designed to drive machine learning and artificial intelligence. The chip was designed by Google Research over four years, based on the past 20 years of computing experience, and is designed to drive the development of machine learning and artificial intelligence, and is the Pixel's biggest innovation to date.


The chip is said to enable Google to easily translate advances in artificial intelligence and machine learning into actual consumer products. Google has not announced the performance specifications of the Tensor and foundry, but according to previous revelations, the chip is a Samsung 5nm process with 8-core Arm cores, close to the performance of high-end Exynos SoCs.


Google TPU was introduced back in 2016, is specifically for accelerating machine learning processing unit, has been in service in their own data center for many years, also launched the enterprise solution, now is the first time into the phone. Quick queries and longer audio tasks, such as live captioning or recorder applications.


What can a processor that excels at machine learning do in a phone?

Tensor also enables new Pixel 6 features such as motion mode, more accurate face detection, and real-time translation of text into different languages as you type it, currently available in English, French, Simplified Chinese, Japanese, Spanish, German, Italian, Portuguese, Hindi, Polish, and Russian. Pixel 6 and Pixel 6 Pro can simultaneously translate the corresponding language through subtitles and audio of movies.


For the smart voice enhancement, Google still spent a lot of time in the conference to introduce, but also invited the "queen of storage" Mariya Kondo to demonstrate a Google real-time translation dialogue.


Specifically, in terms of intelligent voice, Google first emphasized its own voice recognition capabilities.


For example, "Bryan" is often misidentified as "Brian", but after a few corrections, when you hear "Bryan", the Pixel phone will no longer mishear it. Pixel phones will no longer be misheard.


In addition to the technology, Google showed us the text message dictation function, in addition to text input, you can even use your voice to enter emoji, just say one more "emoji" after the keyword. At the same time, you can also say "send" to send, "clear" to clear the message, or when you see a typo in the dictation, you can directly click on the typo and repeat it, without cutting in and out of dictation mode.


And if you can't speak the language of the person you're communicating with, Google can translate the chat in real time, whether it's in a message box text exchange or in a real-life voice exchange.


In addition, Tensor has improved the "Call Screen" feature for cell phones, letting you know the average wait time for the top 5,000 companies in the U.S. when you make a call, or convert automated voice content to text and display it on the screen, or you can hand the call over to Google Assistant, which will send you a message once you're connected. and once connected, it will send you a notification.


Google also said that the Tensor chip will handle specialized machine learning tasks and is far more power efficient than previous Pixel phones.


There's also a dedicated image processing core that helps the Pixel improve photography. For example, thanks to Tensor's powerful performance, the Pixel 6 and Pixel 6 Pro introduce the Magic Eraser feature, which is unique to other competitors, and when someone else enters the shot by mistake, the system will identify people or things that probably shouldn't be there and then offer you "Delete" suggestion, you can choose to delete at the system's prompting tap. Of course, you can also manually select what to delete at other times.


Other upgrades include Face Unblur, which automatically detects faces in the frame, and if it finds a blurred face (such as a head movement, etc.), the phone automatically activates a second shot and lets you take two photos at once. The AI will then paste the clear face from the wide-angle camera onto the photo taken by the main lens to make a clear hybrid photo.


In addition, the Tensor chip also gives the phone's camera a variety of different ways to play, including but not limited to Action Pan, Long Exposure, and more.

Alan Smith