In Google Pixel 3, the developers of the search giant rethought the method of creating portrait images, replacing software algorithms with a neural network that can not adjust the camera to one of the specified scenarios, but create them independently depending on the shooting conditions. Thanks to this, it was possible to significantly improve the quality of the background blur even with the help of a wide-angle camera without the need to equip smartphones with an additional telephoto lens.
Despite the fact that the algorithms responsible for creating portrait photos in smartphones of previous generations worked quite well, Google decided that they are far from ideal when it comes to the quality of determining the depth of the frame. So it was decided to make the new Pixel camera smart. Especially for this purpose, the developers have designed a unique case that can accommodate five Google Pixel devices, located with a small offset relative to each other.
Shooting with parallax effect
With the help of such a cover, Google was able to simultaneously make five frames, which were later combined and formed a parallax effect due to the displacement. Subsequently, this data was processed by the neural network, understanding exactly how the depth of the frame is formed. Thanks to this knowledge, Google Pixel 3 and 3 XL can now take portrait shots that are close to professional in terms of detail and background blur quality.
The research that Google is doing in this area is of great importance for the entire industry. The search giant was the first to prove that the creation of portrait images is possible even without the use of a telephoto lens, pushing for the implementation of a similar technology Apple, and added the findings to the application “Google Photos”. Thanks to them, users can manually increase or decrease the intensity of the background blur already in the finished pictures, which many smartphones simply do not.