Google’s Pixel and Pixel XL take some of the best photos in the mobile world thanks to their internal Gcam software, and Alphabet’s X department, the birthplace of the software, has taken a moment to pen a blog entry telling the curious story of how the technology was born and eventually found its way into the Nexus lineup’s replacement. The story starts back in 2011, with Sebastian Thrun, the head of Google X at the time, trying to figure out how to finagle a decent camera onto something as tiny as Google Glass without compromising the other needs of the device, such as long battery life and being lightweight.
The issue was quite simple; a decent camera simply could not fit onto the Glass hardware without making it less convenient and easy to wear. While some attempts to approach the problem from the software side were made, the team didn’t manage any serious progress until Marc Levoy got involved. Levoy, a computational photography specialist from Stanford University’s computer science department, Levoy started up the Gcam department that year, and by 2013, their work on image fusion technology was in Google Glass, then got revamped as HDR+ and made its way to the Nexus 5 and Nexus 6 in the default camera app. The tech works by taking a huge amount of different shots in a very short time, then automatically merging those shots, combining the most well-lit, accurate, and detailed parts of each picture to create a composite photograph.
The technology was improved over time, until it finally ended up as the powerful iteration of HDR+ that serves as the default photo mode on the Pixel and Pixel XL. Gcam is still hard at work to this day, delivering powerful imagery and photography improvements to the likes of YouTube, Google Photos, and even virtual reality applications. Going forward, the Gcam team is looking to bring new technological advances to the world of digital photography. Machine learning to control things like white balance and exposure could be just the beginning; for now, there are no particular forward-looking projects planned, but machine learning is a definite next direction for the department.