After Google released a trove of photos clicked from the newly-released Pixel 2 and Pixel 2 XL to show off their camera prowess, a new discovery further ascertains why the duo deserves the ‘best smartphone camera’ title.
Google did not mention this at the launch event, but the Pixel 2 features a dedicated SoC, designed in-house, that aids in camera processing and capturing better HDR+ photos. Interestingly, the SoC is not active yet, but it will make the phone process photos faster and more efficiently than at present.
As a report on ArsTechnica notes, in addition to the Qualcomm Snapdragon 835 SoC, there’s a secondary ‘Pixel Visual Core’ SoC which has been designed by Google. At the heart of the chipset is an eight-core Image Processing Unit (IPU), which is capable of more than three trillion operations per second. Essentially, this allows the Pixel 2 to run HDR+ image processing 5 times faster and at less than 1/10th the energy than the current rate. ALSO READ: Google Pixel 2, Pixel 2 XL review roundup: A look at what reviewers say
For context, the Pixel 2 and Pixel 2 XL do not feature dual-camera setups as is the current industry fad. Both are equipped with a conventional single-lens module, however, the 12.2-megapixel rear camera is powered with machine learning-based Portrait Mode to add the DSLR-touch to your photos. So even without dual lenses, you get enhanced results.
Google says that the custom chipset will be enabled with the launch of the Android 8.1 developer preview. It will then let third-party apps use the Pixel 2’s HDR+ photo processing. This basically means, Facebook or Instagram’s camera app will be capable of producing images at par with the Pixel 2’s native app. ALSO READ: Google poked fun at Apple for ditching the 3.5mm audio jack last year, only to do it with Pixel 2
The Pixel Visual Core SoC has not been designed keeping only the Pixel 2 in mind. Google says the chipset can handle ‘the most challenging imaging and machine learning applications’. The company is already preparing the next set of applications designed for the hardware. Given the strong emphasis on machine learning, we might be able to see some applications in augmented reality.