Name the number one problem with smartphone cameras. If you picked “nighttime photos,” we agree: While smartphone cameras have improved by orders of magnitude in the last decade, due to the constraints of their tiny lenses, low light conditions are still their Achilles’ heel.
But a Google researcher might have a software solution to that problem — one that actually works, that is.
Florian Kainz, who works as a software engineer at Google’s Daydream team, has developed a technique that let him take photos in very dark conditions with the Google Pixel and Google Nexus 6P phones and get results that are comparable to photos taken with a DSLR on a tripod.
To do this, Kainz basically uses burst photography to take a lot of shots of the same scene, and then analyzes the resulting photos to remove sensor noise. His solution builds on the work of another Google researcher, Marc Levoy, and his experimental SeeInTheDark app, but Kainz has taken it a few steps further — check out the details here.
In practice, Kainz built a simple camera app for Android that gives him control over several key parameters like exposure time, ISO and focus distance.
Instead of taking a single photo, the app takes a large number of photos — up to 64 — in a short period of time. The individual photos, taken in conditions ranging from moon lit night scenes to near-total darkness, were grainy.
But when Kainz computed the mean of all the frames, and accounted for some of the movement in the photos (such as the movements of the stars in the sky), he managed to get very clear and beautiful outdoors photos, comparable to those taken with a DSLR camera.
There’s still a long way to go before all this can be turned into a one-click solution.
“Arriving at the final images required a lot of careful post-processing on a desktop computer, and the procedure is too cumbersome for all but the most dedicated cellphone photographers.
“However, with the right software a phone should be able to process the images internally, and if steps such as painting layer masks by hand can be eliminated, it might be possible to do point-and-shoot photography in very low light conditions,” Kainz wrote in a blog post.
And there are other obstacles; this approach only works if the cellphone is resting on the ground or is mounted on a tripod, so it won’t help much with your blurry, indoors birthday photos.
Still, the results are promising; while Kainz did most of the work by hand, it’s not very hard to imagine most of this process being automated and making its way to a widely-available camera app.
Check out an album of Kainz’s photos processed using this technique.