Center crop of a 12MP iPhone image and a 12MP Mirrorless Camera Image
One of the fantastic things about computational photography, though, is that it scales with the hardware. Improve the hardware and, in many cases, the computational photography algorithms can be applied more subtly, thus reducing the impact of these tradeoffs.
Predictable Results
It’s hard to depend on your smartphone camera in the same way that you can depend on a standalone camera. When you carry around a standalone camera, you learn what it can do, how it will react to different scenes, how the images look, and how much can be accomplished later when editing. In short, standalone cameras are predictable.
This is a critical piece of the puzzle for smartphone cameras, as the phone is making so many decisions which are completely hidden from the user that it is impossible to confidently know what the resulting images are going to look like.
This goes far beyond typical “auto” modes on cameras - smartphones are… smart. For example, a phone will expose a scene differently based on what is in it. Is there a person in the shot? Is there significant movement in the scene? Can you see the sky?
And these are not bad features. In fact, they automate many of the adjustments a photographer makes when composing a picture. What’s missing is consistency it is incredibly difficult to predict how your phone camera will handle a scene until you’ve actually opened the camera app!
The other half of predictability is how the image data is handled after the picture is taken. When I look at a picture that’s been taken on a smartphone, I never know how much I will be able to edit it before it falls apart.
How much noise is there in the shadows this time? Can I tweak the highlights so that the sky isn’t as blown-out?
Who knows.
Post-processing in this way is niche - and a bit of a pro feature - but even still, we’ve recently seen some steps in the right direction. We just need more.
Smartphone cameras should over-capture in the shadows and highlights, and retain the data, if the user wishes. Once you know that you can rely on post-processing to refine your photos, you can begin to capture more freely.
Predictability might be seen as a nuanced subject for such mass market devices, but in reality, everyone can benefit from it.
Allowing people to build confidence in their cameras opens up new photographic opportunities which may otherwise have been skipped-over. Creating images that have more room to be edited helps everyone achieve the look they want, whether it’s in Lightroom or Instagram.
The Undefeated Convenience King, with a Catch
Taking a picture has never been more convenient - nowadays, you always have a “good enough” camera on you. This phenomenon can be thought of as “Image Quality vs. Convenience” but there is another metric that tells a different story: “Image Quality vs. Portability”. At the risk of raising some eyebrows, the Image Quality vs. Portability today rarely exceeds that of the film era. Cameras as far back as the 1950s offered pocketable sizes and big image quality.
Film is wildly inconvenient compared to digital, but you could throw your camera into your back pocket and get stunning pictures. In many ways, the camera my grandfather took on family outings back in the day produced better pictures than my modern-day smartphone can!
Even modern standalone cameras rarely deliver better on Image Quality vs. Portability. There are a number of reasons why this is the case - all of which are outside the scope of this post - but much of it has to do with modern lens design and the pursuit of the perfectly-corrected lens.
Many film cameras had objectively worse optics, but they still made images good enough to exceed our current smartphone standards!
Hardware goes a long way with cameras, so it’s time that we leverage it more in the realm of smartphone cameras. Combine that with using computational photography to address any remaining (or engineered) aberrations, and image quality takes a big leap forward.
At Glass Imaging, we are engineering our camera hardware and software hand-in-hand so that they complement one another, getting the best out of both of them.
Why Does it Matter?
Today’s smartphone images are not going to stand the test of time particularly well. Memories that are captured on smartphone cameras do not display well in larger formats or at higher resolutions.
Digital picture frames, desktop backgrounds, TV screensavers, virtual reality - and of course, printed images - all depend on having an image quality that smartphones simply cannot deliver in their current state.
When I look back on a memory, I want to be able to see the details and be pulled into the scene, not caught off-guard by mushy details, HDR artifacts and an exaggerated tone mapping.
Because the smartphone camera plays such a significant role in our lives, it should absolutely be the best it can be. The good news is that many of the issues outlined above can be addressed with improved hardware, clever software, and a little out-of-the-box creative thinking.
The innovations we are working on at Glass Imaging can be stacked with the last fifteen years of advances in computational photography to deliver a camera experience which will finally allow you to leave your camera bag behind. But this time, without compromise.
