al lehman Loc: San jose, ca.

I've heard rumors that Canon may have filed for a new curved sensor patent. Has anybody heard of this or familiar with this potential sensor technology?

rmalarz Loc: Tempe, Arizona

Two companies are already producing curved sensors. So, why not?
--Bob

Architect1776 Loc: In my mind

Yes, and they have patented several lenses for use with the curved sensor.

https://www.canonwatch.com/canon-researching-curved-sensors-dual-pixel-af-patent-application-suggests/

https://www.canonwatch.com/canon-might-work-on-curved-sensors-lens-patent-suggests/

https://www.canonwatch.com/yet-another-canon-patent-curved-sensor-design/

https://www.canonwatch.com/canon-patent-curved-sensor-design-suppress-vignetting/

https://www.canonwatch.com/canon-patent-electrically-bending-image-sensor/

And many more.

wdross Loc: Castle Rock, Colorado

They have. There are even pictures of it. This would require a whole new lens design for it to function the way that I think they are trying to go. Older lenses would work, but not like the newer lenses would work. It would cut aberrations and distortions in the corners and curve the depth of field. It will be interesting to see where they actually go with this. It should also make lens design easier.

R.G. Loc: Scotland

Unless you can make the curvature variable you are limited to one specific focal length - i.e. no zoom lenses, just one prime.

rmalarz Loc: Tempe, Arizona

Good point. So, here's a suggestion. Since they are trying to emulate the curvature of the eye, why not make the lens variable by changing the curvature. All we need is a flexible, perfect clear optical component.
--Bob

R.G. Loc: Scotland

They have been experimenting with liquid lenses (like wot our eyes have ) but there are problems they haven't managed to resolve (last I heard).

Architect1776 Loc: In my mind

Canon CAN make it variable.
Look at my previous post and links.

Architect1776 Loc: In my mind

Canon has been experimenting with this since the 80's.

wdross Loc: Castle Rock, Colorado

Not totally true. The plane of focus becomes more variable. Where most camera lenses are designed to create a flat plane of focus that coincides with the sensor plane, theirs will take a curved plane of focus from what is in front of the camera and project that curved plane on to the curved sensor. A zoom can be made to accommodate this design. Which means that for group shots one would have the corner people move closer to the camera and would distort them less. Also there would be less vignetting with a curved plane of focus and curved sensor. It would be my worry of what one would do for a filter.

R.G. Loc: Scotland

I'll take your word for it . The last I heard was that they didn't have a working zoom for curved sensors. And while a liquid lens allows for fast focusing, that doesn't mean that it allows different focal lengths with a fixed curvature sensor. But that may be out of date.....

R.G. Loc: Scotland

There's curvature and then there's curvature. I suspect that a high level of precision is needed to maintain the sharpness across the sensor. Just bending it with a piezo crystal may not be precise enough - unless they've come up with something clever......

al lehman Loc: San jose, ca.

awesome, Thanks!

al lehman Loc: San jose, ca.

The precision on the hubble space telescope glass is incredible and it works under tremendous environmental variations with heat and cold. Similar technology.

larryepage Loc: North Texas area

Actually, at least one patent was already awarded to them in this area. After seeing your post, I did some research, reading some articles in the popular domain and downloading a couple of academic ones. Keeping in mind that many times, companies in the business of intellectual property will patent a technology not nearly so much because they want to actually produce it as because they want to make sure that no one else makes it. (Or if they do, they have to pay to license the rights to do so.)

Once you dig a little bit, you learn a couple of things. The first is that these sensors are not made as curved sensors. They are made as flat sensors, then "deformed," forcing them to assume the curved shape. The second is that they are produced on silicon (or whatever semiconductor material) substrate that is significantly thinner than normal sensors in order to make them easier to curve. Unfortunately, silicon is like glass, so that means that it becomes more fragile as it gets thinner. In fact, while work has been done on "full frame" size sensors, the most recent articles indicate that current work is being focused on significantly smaller sensors for applications like drone cameras.

There was an old joke about the difference between being "involved" and being "committed." The parallel had to do with breakfast and how the chicken was involved in breakfast (the eggs), while the hog is committed to breakfast (the ham). The reason this matters here is that there is no reasonable way to adapt planar image lenses to curved sensor cameras nor curved image lenses to flat sensor cameras.

There seems to be a consensus that this is not a near-term development for general photography, but that the focus will be to equip miniature drones and other surveillance equipment. I wouldn't be surprised if it were militarized, classified, and kept from the commercial marketplace altogether, at least for some time.