Dolly

This post is a short summary of my thoughts on computational photography and the use of artificial intelligence in photography. Smartphones are now the photographic tool of choice for the vast, vast majority of people— they easily meet, if not surpass, their image creation and sharing needs. There is an entire generation coming up using only smartphone cameras— with no need to even consider traditional cameras. A rather important thing to remember is that the smartphone has not only redefined how we take pictures— it has redefined how we view them as well. Most people do not print images any longer. Most images are viewed on smartphone screens. That changes the requirements for what people want and value in a camera in a huge way. A smartphone camera is best suited for both producing and sharing images with other people. The smartphone changed the entire rules of the game. A camera in your phone is fun, and it is very freeing to walk around taking pics with this instead of lugging a big camera sometimes. Relatively speaking (if you are not making large prints, or do not make destructive edits later), they have good image quality and are getting better with each new model. There is no getting around physics, though— smartphone cameras, and therefore sensors, are tiny. In the beginning, smartphones had a serious disadvantage to deal with: they don't capture enough light. But that's where computational photography comes in. By combining machine learning and algorithms with traditional optical processes, computational photography enhances what is achievable with traditional methods The simple fact is that with this in the smartphone arsenal, they can greatly exceed the physical parameters of the sensor used in that phone. All of the computational photographic "tricks", of course, could be done on a large sensor camera— but all of the traditional camera companies are so far either unable or unwilling to implement these things. Unable is probably a more accurate description, as they don't have the necessary astronomical R&D budget (Billion$) that the phone manufacturers are employing. So the camera companies employ larger sensors, brighter lenses, faster AF (since that has been their modus operandi for twenty years and more)—rather than developing cutting edge software programs for the cameras. The future of photography is most certainly a combination of software and hardware. The innovation from just the past few years of smartphones is a perfect argument for this. For as long as traditional camera companies are unwilling or unable to realize this, smartphone cameras will continue to make major leaps forward in results while cameras make small ones. Camera makers are like a fish out of water when they try to compete with Apple and Google on AI and computational photography. They can't even get the menus right...