For more than 100 years, camera and lens makers have been doing signal processing in the analog domain with ever more carefully and cleverly shaped glass – to bend and synchronize light rays.
Now the equivalent thing, and much more, can be done in software, rendering most of that effort moot.
Modern smartphones have the computational abilities of supercomputers, and use them to produce images that rival those from expensive, heavy, bulky cameras – using tiny cheap lenses and sensors.
See, for example:
and
https://research.googleblog.com/2017/11/fused-video-stabilization-on-pixel-2.html
Traditional camera and lens makers need to get on the DSP wagon or be left behind. Soon – time is running out! You don’t want to be the next Kodak.
(Canon, Nikon, Pentax, Olympus, etc…this means you.)
All modern cameras from those manufacturers have DSP chips in them, and they can correct for pincushioning, barrel distortion, chromatic aberration, and vignetting. Canon’s latest mirrorless camera has EIS, and I suspect that’ll be coming to the rest of its camera line, if it hasn’t already.
Regardless, the SLR and point-and-shoot businesses are suffering. I don’t think better technology can save them – at least in the consumer market. A large, dedicated camera will always be theoretically able to do more and do it better than a small multi-function device with a lens on it. But for consumers (who used to be happy with pocket Instamatics and Polaroid One-Steps), the small multi-function device will be more than good enough.
Yes, the traditional players are dipping their toe in the water re DSP, but nothing close to the all-in approach of Apple, Google, Samsung, etc.
Every high-end smartphone captures HD video at 120 fps or more. Some at 4k. How many “cameras” do that?
It’s because they don’t have the computes. Smartphones all have GPUs, and some now have dedicated tensor processors for video and image processing. This is thousands of times more compute than you see in today’s cameras.
These people are making the same kind of mistake DEC did with the Rainbow, and that many other industries make when fundamentally new technologies makes their old expertise largely obsolete – a little adaptation. That’s what Kodak tried to do with digital cameras – now they’re bankrupt.
When a transformation like this comes along, you’ve got to jump in the deep end with all your commitment, not creep in at the edges. Or you get Kodaked.
Large, expensive lenses in fact have little value today – because so much can be done with software correction. Yet this is where all the traditional companies make profits – lenses, not cameras. And it’s where their pride and prestige comes from. They are going to have to swallow that. Or fade away.
I don’t see them doing nearly enough.
Large sensors still have value, but less so as multi-exposure techniques mitigate the noise and dynamic range issues in small sensors. Large sensors will always be better, but outside a few specialized applications, by small margins.