In the new Pixel 4 patch, Google tends to correct auto white balance

Google released an update on the Pixel 4 smartphones that some users brought to light last month to correct the auto white balance problem. The fix was part of the security patch in November that is now going out to all Pixel users, although the camera repair is only for the latest models.

In the new Pixel 4 patch, Google tends to correct auto white balance
In the new Pixel 4 patch, Google tends to correct auto white balance
How it looked (Pixel 4) • How it should look (iPhone X)

A few days ago, the Pixel 4 auto white balance, which appeared to change solid, subtle colors, was brought to the attention of a reddit user. The phone turned the room into yellow instead of red when taking a photo in a room that only has a red light source (a Philips Hue LED bulb set to bright red). On other phones, this problem was not found which correctly captured the light.

The problem was corroborated by Android Police writer Ryne Hager, who took a video of a crimson theater curtain that looked like a very light pink hue after the image was processed by the Pixel 4 lens.

In the new Pixel 4 patch, Google tends to correct auto white balance
The curtain is supposed to be deep red (Image source: Android Police)

The issue seems to be down to the new machine learning based auto white balance algorithm that Google introduced with the Pixel 4, which instead of processing the image the way it is tries to correct it to what it should be. Google demoed this on stage during the Pixel 4 launch event, where the phone corrects snow that looks blue due to light from the sky to white.

Machine learning, however, is just as good as the data it receives and is bound to travel sooner or later. When a completely red room was shown, the ML machine probably did not know how to handle the situation and did its best to delete the red picture, which left the image with the other two primary colors, blue and green, which together created the yellow image.

The question is solved after the latest update and that the camera now renders a red room as red correctly.

Though, it is worth remembering that the program may not yet be bullet-proof and can still be fooled into extreme scenarios for which the machine learning algorithm had no preparation.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *