Sony patent shows a new curved sensor technology.
Egami (Translation here) spotted a new Sony patent describing a curved sensor technology. Having a curved sensor allows an increase in the lens performance and reduces the size of the lens. Egami speculates that this may be ideal for the future Full Frame NEX-9 camera. But of course, if you bring a camera with a sensor like this current A or E mount lenses wouldn’t work. I highly doubt the NEX-9 will have a sensor like this.
Andrejew
8 months ago |Yes the a and e mount lenses would work! With an adapter! Like canon fd lenses on A-Mount does.
There is an nice comparison on Flickr for all new cameras against the rx1 around photokina:
http://m.flickr.com/#/photos/39900657@N08/8023739331/
Daemonius
8 months ago |Nope it wouldnt cause conventional lens cant work with such curved sensor.
LifeStoryImages.com
8 months ago |“Apples and oranges” difference between adapting for different mounts (varying the back focal length) vs. compensating for a curved image plane. A lens is designed either with or without field curvature. For Alpha/NEX lenses, image quality would be shot trying to “adapt.” (If the lens is designed with a curved plane in the first place, yes indeed, there are inherent design advantages.)
Even if you used an adaptor, it would have to be very near the image plane. i.e., the existing space between the rear of the lens and the image plane.
A quick (5 minute) test I just ran confirms this; you would not be able to effectively correct (or add) field curvature w/o changing focal length and field of an existing lens whilst maintaining the same optical performance.
Maximus
8 months ago |Is this graphic showing the whole sensor or just one Pixel? If its just one Pixel i don’t see a reason why A or E mount lenses shouldn’t work.
Steve Jones
8 months ago |The translation clearly chows it’s the image plane that is curved, so it’s the whole sensor surface. As others have noted, the most likely use of such a technology is in devices with fixed lenses as otherwise it would require a whole new family of interchangeable lenses with very precise exit optics. That, frankly, is never going to happen so I think the applicability of this patent to A and E mount systems is pretty well zero. Any optically corrected adapter, if it’s possible at all, would reduce quality and rather negate the whole point of the deign to reduce the lens size.
Varn
8 months ago |FF RX2 ILC with dedicated compact Lenses???
decato
8 months ago |That is the most likely scenario I guess. I don’t see this being used in an interchangeable lens system anytime soon.
Like the organic sensor, these patents may not translate to products most of the time . At least not yet.
雜草
8 months ago |These type of lenses are probably cellphone lenses.
Dave Lively
8 months ago |+1
Long focal lengths would require less curve than short ones making a curved sensor much less useful in a camera with interchangeable lenses. It would make the most sense for a camera with a fixed, wide angle prime lens like cell phones have.
Cell phone sensors may not cost a lot individually but when you look at the volume they are a very important part of the sensor market. Sony probably spends a lot of R&D on them.
ChenAlan
8 months ago |+1
Imagine if they could make cellphone sensors a bit “bigger” using this technology.
Then again… we could extend this to maybe Cybershots as well. Imagine an RX100 with an effective APS-C sensor instead
Sky_walker
8 months ago |So far they don’t even make EXMOR-R “bigger” and there’s not even half as much problems with that as there is with making APS-C sized sensor curved. So I don’t see it happening.
Freddo
8 months ago |Highly probable. The curved sensor would simplify the design of the lens array (field curvature = no problem). Maybe they are even looking at a solution where the sensor is glued directly to the rear lens?
Dirk
8 months ago |More than highly probable.
All lenses have field curvatures; wide angle lenses have larger curvatures (smaller radii) than telephoto lenses. A flat sensor/film has zero curvature = infinite radius.
Soft corners and distortion are more difficult to eliminate from WA lenses because their curvatures differ most from that of the sensor/film. If the sensor curvature is halfway between the curvatures of the telephoto and wide angle lenses, the telephoto lens would be as easy to make as before, but the wide angle lens would be (much) easier.
Time to start saving up for new systems…
passer-by
8 months ago |It can be made compatible with current lenses (flat sensor) if they can make it flexible (flat to curved to flat in real time). But I don’t think the technology is there yet.
Sky_walker
8 months ago |Yep. We’re not even close to that. Probably the first step will be in liquid lenses engineering. Then they’ll think about anything more.
Boooe
8 months ago |or they can use positive lens unit on top of sensor to make surface look flat to work with older flat-filed lenses.
Simon_P
8 months ago |Good for DC and mobile phone, so that the lens size can reduce, and sensor surface can increase
terrance hounsell
8 months ago |This makes alot of sense and corrects for the fundamental physical defects inherient in lenses (as we currently know them). Of course this has been done with film several times, EG: my AGFA Clack has a curved film plane whereby they were trying to improve the performance of a simple (usually 1,2, or 3 element) lens. Pinhole cameras also do this EG: my 6x18cm Dragon II has an extremely curve film plane to compensate for the light fall off inherient in pinholes. This is the sensor configuration of the future, all digital cameras will be this way as soon as manfacturing details are worked out. Current A or E mount lenses will not work without a corrector optic in the adapter because the focal plane is effectively farther away from the nodal point of the lens and therefore a negative distance is required to correct it and this is not possible with an adapter alone. Same reason that some film lenses will not focus on some digital cameras at infinity without optics in the adapter, To learn more google “back focus distance” or try Wiki.
Boooe
8 months ago |negative distance?
the “corrector” is simple positive lens unit.
Rooru S.
8 months ago |they’re after making ultra compact good quality lenses/sensor combos. I expect seeing this kind of sensor together with the other technology Sony patented some time ago about Liquid lenses to have a camera with the human eye resolution in a few years for sure.
JD
8 months ago |Question is : how much size reduction is that?
Daemonius
8 months ago |Not that much and price for it would be ability to use ONLY lens for that specific sensor. Regular lens are designed for FLAT sensor or film.
butch
8 months ago |for anything other then a prime lens then would need an adjustable flexible sensor by vacuum behind the sensor maybe and a square or round sensor would make that easier
terrance hounsell
8 months ago |Remember the Contas RTS III which had the RTV (Real Time Vacuum) system. Kyocera decided that it should develop a system to hold the film as flat as possible by using a vacuum system to suck the film to an ultra flat ceramic pressure plate. This is accomplished by using a electro-magnetic rubber suction diaphram placed behind the pressure plate. The pressure plate had three vertical slits, each of which has a small hole through it. The diaphram created a vacuum by sucking air through the holes thus vacuuming the film to conform to the pressure plate. This system was fast enough to allow the camera to have a reliable 5 fps motor drive. I am sure a simialr technology could be used for a soft membrane type digital sensor.
JMi
8 months ago |I guess this type of sensor could be developed ad hoc for the manufacture of a fixed zoom on a Point & Shot. The fact of not having to compensate for the field curvature will produce better quality in the rest of the optics.
What I have a problem is that the sensor must be curved in all directions, that is to say concave. That is much more difficult to achieve.
Lonnie Utah
8 months ago |Eh. You could have an adapter that adjusted the focal plane of a traditional lens to the curved sensor.
But it’s most likely a cell phone sensor.
LifeStoryImages.com
8 months ago |Absolutely not, for alpha/nex systems. Image quality would be shot. Either a lens is designed with or without field curvature, and even if you used an adaptor, and effective one would have to be very near the image plane.
Yes, cell phones.
PhotoNut
8 months ago |Such a curved sensor would only work in cameras like the RX1 or cellphones with fixed focal length lenses as otherwise the curvature would need to be varied based on the optical design of which ever lens is attached to the camera.
LifeStoryImages.com
8 months ago |I know what you’re saying, but it’s not that a zoom lens couldn’t have a curved image plane. Having a fixed lens/sensor combo is the key.
TimberWolf
8 months ago |Ahh… so we’re finally catching up with eyeballs in this area
John
8 months ago |Hey, we have flat sensors now with just about all lenses having curved image planes. So I bet existing lenses would work with such a sensor, in fact they may work even better unless they have little to no field curvature.
Most likely this technique will be useful for a class of lenses for compact large sensor cameras and associated lenses where small lenses are of high importance as well and high IQ off center.
They amount they are bending the sensor is really small (though large compared to field of curvature) and is exaggerated in the patent images.
Henry
8 months ago |Take that RX100 sensor, put it in HX200 body, slap a 24-300mm lens in there, instant bestseller!
JOhn
8 months ago |You won’t want fresh air coming anywhere near that sort of design because dust is going to collect quite nicely on the bottom half of that curve as time goes by. What about anti-dust technology? Well, that’ll shake the dust from the top of the curve to the bottom.
LifeStoryImages.com
8 months ago |“When activating the anti-dust shake, hold the camera in a downward-pointing position….”
lukas
8 months ago |im not an expert in sensor tech… so… maybe im wrong about this, but
i missunderstand things? or dont all sensors get a curved image and then by software, all currrent photo engines compensate about this?
i remember a most in use test for quality image in sensor its to look at the sharp of the edges, am i right or wrong?
this technology can allow to bypass that software processing? i think…
much like the new sensor (with a new array of colors) skip a must-do software procesing AA?
today pictures are too washed, with AA filters, demosaiced, curver lense compesation (someone please tell me if im wrong or rigth about it), jpeg engine, etc….
maybe this is a tech to more clean image, a free path from sensor to file, that dont need all the procesing they get today.
Booe
8 months ago |No, most current sensors are flat and most of current lenses produce flat images (except for special ones or very simple lenses)
It is natural for lens to produce non-flat image, optics designers have to make it flat to match flat sensors.
Curved sensor is not about higher image quality – it is about the same image quality with number of lens elements reduced by about a factor of two in case of WA lenses (cause optics designers may keep lens’ “natural” field curvature and angular response issues are less a problem with curved sensor)
Note that with curved sensors a non-retrofocus fisheye might become real in future
LifeStoryImages.com
8 months ago |No, and, yes. You are correct that some post-processing algorithms take out residual lens aberration, including residual field curvature, astigmatism, coma, spherical, distortion, and chromatics. What is being talked about here is field curvature that is WAY beyond the depth of focus and would not be correctable in post.
David DeBar
8 months ago |It not really such a new idea. Minox did it decades ago:
They had the film held in a curve. I recall an IBM publication where they had film in bowl shaped sections for use with small cameras. The idea was to reduce the cost of the lens. If the patent office does it’s job this patent application will not be granted, because of prior art. That does not mean the patent will not be granted. Sadly, the PTO has frequently granted multiple patents for the same invention.
http://en.wikipedia.org/wiki/Minox
LifeStoryImages.com
8 months ago |The patent is not for “curved focal plane.”
David DeBar
8 months ago |True LifeStory, it just a curved image sensor. I’m not sure if that has a similar effect or what it does? Also it’s done in a solid state device rather than film. That could make it patentable. Somebody needs to post the claims if we are to understand this.
juan
8 months ago |im seriouly thinking about moving to nikon, sony is killing me with all these changes i have $5,000 worth of lens and if sony discontinues its alpha line and lens to replace with this curve sensor im going to be really mad,,,,,,,,,,,,,, THEY NEED TO FOCUS MORE ON THERE NOISE ISSUE SMH 1600iso looks like crap on all alpha lines i dont wanna hear “oh its passable at 1600iso” crap