Google recently unveiled the Pixel 8 Pixel 8 Pro , and among its updated roster of features , the caller keep circle back to its Real Tone capableness – but it ’s not a characteristic undivided to the most late theoretical account .
alternatively , Google Real Tone has been a feature of many late Pixel phones , down to the budget - rivet Pixel 7a and all the way up to the Google Pixel Fold – though carrying into action is n’t uniform among all Pixel phone , with the most recent Pixel 8 series deliver the most accurate skin tone yet .
With that in mind , here ’s everything you need to know about Google Real Tone , including what it is , how it work and where you could happen it …
What is Google Real Tone?
Real Tone is a Pixel photographic camera feature article design to improve truth when multiply dark peel tones where previous Google earphone have fallen brusk .
Google has solve with more than 60 lensman of dissimilar ethnicities to take aim the Pixel tv camera to capture darker skin tone more accurately , forestall them from appear darker or brighter than how they expect in real life .
Real Tone is a part of Google ’s ongoing effort to address skin tone diagonal in AI and tv camera applied science , and it has continued to improve its operation since its original iteration on the Pixel 6 .
“ This racial diagonal in photographic camera engineering science has its source in the way it was developed and tested”,explains Google on its website .
“ photographic camera sensors , processing algorithms , and redaction software package were all trained for the most part on datum sets that had light pelt as the service line , with a limited plate of hide tonus taken into account . This bias has lingered for decades , because these technologies were never adjust for those with darker cutis ” .
Because Real Tone is baked into the Pixel ’s camera , it is also support on many third - party apps , like Snapchat .
How does it work?
Rather than being based on one single app or technology , Real Tone follows a theoretical account that addresses six core areas in Google ’s imagery technology .
First of all , the Pixel camera is trained to find a diverse reach of faces to ensure the television camera can get an in - focus trope in a smorgasbord of lighting experimental condition .
Next make out clean counterbalance . The auto - lily-white balance is design to better chew over a variety of peel tones .
The automatic exposure is responsible for how bright photos snapped on the Pixel look , so Google has amend this as well .
moreover , the company developed a unexampled algorithm to reduce the impact of isolated lightness on an image , insure dark skin tone do n’t come along washed out when framed by a sunlit window , for example .
obviate blur is also central to getting a estimable photo , so Google has used the AI in its custom - madeTensorchip to keep images sharp in lowly - igniter conditions .
Finally , the auto - sweetening feature in Google Photos work with photos aim with smartphones outside of the Pixel , mean other Android user can optimise the colour and lighting in their images after taking them .
All of the above is applied to the Pixel ’s photographic camera , which is then tested by professionals to expand Google ’s simulacrum datasets .
“ We go along to work with image experts – like photographers , cinematographer , and colorist – who are celebrated for their beautiful and accurate imaging of biotic community of color ” , said Google ’s Image Equity Lead , Florian Koenigsberger .
“ We take them to test our cameras in a wide-cut range of problematic light conditions . In the process , they take thousands of portraits that make our image datasets 25 clip more divers – to look more like the human beings around us ” .
More late , Google has open - sourced the Monk Skin Tone ( MST ) scale , a scale developed by Harvard professor Dr Ellis Monk that is more inclusive and ply a tolerant spectrum of skin tones than the current tech industry standard .
Which Pixels support Real Tone?
Real Tone launched on thePixel 6 in 2021 and can be found on the observe phones :
