Egami (translation here) spotted the next 3 layer sensor patent from Sony. Lately we have so many of these patents that a camera featuring a sensor like that may come in the near future. This would certainly be a quantum leap forward of Sony can do this. Imagine no more Aliasing problems, true color info for every pixel, top class per pixel sharpness….can’t wait to see this!
P.S.: There are two Sigma Foveon (3 layer) 46 Megapixel sensor ompact cameras from Sigma:
DP1 with 28mm lens at Amazon, BHphoto and of course on [shoplink 15083 ebay]eBay[/shoplink].
DP2 with 45mm lens at BHphoto and on [shoplink 15084 ebay]eBay[/shoplink].
More info: Foveon X3 sensor (Source: Wikipedia)
30 Responses
Very interesting, if Sony can get rid of the noise problems at higher ISO with a Foveon type sensor.
Foveon type sensor is the only way to minimize light loss. Its noise performance should be twice or even third times better (i.e. up to 1-1.6 high iso performance boost) when properly implemented.
Maybe in theory…;-) ISO 50 to 400 images are excellent, but did you ever see Sigma >= ISO 800 images? Horrible!
Yes, in theory. Because with bayer filter we loses part of light in each color pixel, there is a possibility with multilayer sensor to avoid these loses.
To quote Wikipedia… “Although the Foveon sensor has greater light gathering ability, the individual layers do not respond as sharply to the respective colours. Thus color-indicating information in the sensor’s raw data requires “aggressive” matrixing (essentially, removal of common-mode signals) to produce color data in a standard color space, which can increase color noise in low-light situations” This is also my understanding from a friend who worked on the technology a few years back. I would have thought this was a fundamental limitation of the technology, which couldn’t be overcome without some major materials development, so I wouldn’t expect to see it again any time soon.
Recently I have seen some good quality ISO 800 images from Foveon. Both the firmware and the RAW converter have been updated, and this has improved the results.
But 800 does seem to be the limit for good results, whereas a NEX can go up to two stops higher.
Sony’s different approach may give better high ISO results than the Foveon. Foveon images at ISO 100, especially from the DP2M, are stunning. Much better than a NEX 5N at ISO 100.
Sounds promising. When will we see about a Foveon-like sensor in a body with flipping or no SLT mirror and on-sensor PDAF?
Sigma dp2 merrill doesn’t have an SLT in its lightpath.
And the SD1M DSLR has a traditional flipping mirror.
I wonder how will a Foveon type sensor work with videos. Will it share the same benefit with stills or it’ll make current sensor issues become worse?
The current Foveon cameras from Sigma are pretty useless at video. This seems to be mainly because they cannot shift the amount of data produced by the sensor (three times as much as from a Bayer)into memory fast enough.
So it is a chip problem rather than a sensor problem. If Sony can use fast enough chips, they could get video from a multilayer sensor. However, there is not much point. Video doesn’t need the extra definition, at least not until we all have 8k screens, maybe in 15 years time.
but wouldn’t it get rid of aliasing issues completely? also without an aliasing filter the image would be sharper
that is interesting news
Just to point out that Sigma offer three cameras, not two. Besides the DP2M and DP1M, there is the SD1M DSLR.
The APS-C Foveon sensor in the Sigma DP2 outresolves the D800e and its rumoured they are working on a FF version which will kill MF digital cameras.
I hope Sony bring out a Foveon type sensor in 2013, but if they do I expect it will 1st be seen in a mobile phone
Let’s not get carried away. The Foveon sensor might record similar or more colour information (accuracy, especially in the red channel is a different matter…), the 36MP Bayer sensor is still able to record much more luminance information. Any comparison (I’ve seen a few) that brings them to equal grounds or even shows a favor for the DP2, is more than likely subject to processing difference and last but not least… lens influences.
The Foveon sensor is the future. Similar to how CCD jumped to 3CCD .. its the next best thing.
The Sigma SD-1 Merril caught my eye but unfortunately there are too many other negative things about the camera to overlook. Lack of lenses that can take advantage of the sensor, uncommon lens mount, lack of RAW software support.
I was thinking of upgrading to the A99 but when I see that they still kept the damned mirror and are still plagued by the same issues associated with a bayer array and AA filter … I can’t justify spending the money. Lets face it – the technology is out there – it has to be adopted though.
So if Sony truly is considering going to a multi-layer stacked sensor design for colour separation – that’s wonderful news for me. They just need to make it in a full frame mirrorless A-Mount system with the largest available dynamic range.
THEN I’ll upgrade.
Nikon have a patent that senses the emotions of the photographer touching the lens and will only fire when they feel just right.
I kid you not!
Many times patents are just to protect an idea or prevent others from using it and never see the light of day.
🙂
Probably a sensor at the shutter trigger to sense if the photographer is worthy of Nikon gear. LOL.
Not FOVEON,
SOny is developing a hybrid organic sensor.
R,G,B photoelectric cell layers are organic films on the Silicon substrate.
Looks likely Sony needs to develop the fourthly layer sensor cells and directly based on the Silicon substrate, to do the built-in PD-AF.
The R,G,B photoelectric cells are with organic films and the visible light spectrum response are much better than the Foveon Si sensor.
Before 2003, Japan NHK led a starting the organic sensor’s development, the item was brought into Fujifilm after 2006. People also saw the flexible organic films sensor’s video camera has ever had an openly demo by Japan NHK’s Year Exhibition 2008.
But, last several years, Fujifilm is obviously ability not equal to one’s ambition. The X-Trans sensor is also some problems.
http://www.nhk.or.jp/strl/open2004/en/tenji/t39.html
http://www.nhk.or.jp/strl/open2005/en/tenji/t03.html
sony make this sensor 🙂
http://bakicirkin.blogspot.com/2012/10/tunnel-sensor-technology.html
Not FOVEON,
Sony is developing a hybrid organic sensor.
R,G,B photoelectric cell layers are organic films on the Silicon substrate.
Looks likely Sony needs to develop the fourthly layer sensor cells and directly based on the Silicon substrate, to do the built-in PD-AF.
The R,G,B photoelectric cells are with organic films and the visible light spectrum response are much better than the Foveon Si sensor.
Before 2003, Japan NHK led a starting the organic sensor’s development, the item was brought into Fujifilm after 2006. People also saw the flexible organic films sensor’s HD video camera has ever had an openly demo by Japan NHK’s Year Exhibition 2008.
But, last several years, Fujifilm is obviously ability not equal to one’s ambition. The X-Trans sensor is also some problems.
http://www.nhk.or.jp/strl/open2004/en/tenji/t39.html
http://www.nhk.or.jp/strl/open2005/en/tenji/t03.html
http://www.printedelectronicsworld.com/articles/strl-develop-organic-image-sensor-00002348.asp?sessionid=1
Please delete the repeated post,
Thanks
Not FOVEON,
Sony is developing a hybrid organic sensor.
R,G,B photoelectric cell layers are organic films on the Silicon substrate.
Looks likely Sony needs to develop the fourthly layer sensor cells and directly based on the Silicon substrate, to do the built-in PD-AF.
The R,G,B photoelectric cells are with organic films and the visible light spectrum response are much better than the Foveon Si sensor.
Before 2003, Japan NHK led a starting the organic sensor’s development, the item was brought into Fujifilm after 2006. People also saw the flexible organic films sensor’s video camera has ever had an openly demo by Japan NHK’s Year Exhibition 2008.
But, last several years, Fujifilm is obviously ability not equal to one’s ambition. The X-Trans sensor is also some problems.
Admin,
Please delete the repeated post,
Thanks
Meh. All of the lofty claims supporting the Foveon design are embellished to some degree or another. Although that’s not to say that the Bayer design is perfect(let alone ideal) either, as its issues are well documented. Let’s face it; the day someone invents the sensor that can detect every color at each photosite(if that’s even possible), stop-gaps like Bayer & Foveon will die a long overdue death, and all of us will be better off for it.
I’m still waiting for a real sensor technology breakthrough.
They need a sensor that can determine light intensity, wavelength (frequency), and direction. THEN you’ll have the perfect camera.
Of course to get such a thing we’ll probably be looking at a quantum computing system that would take a photo by allowing light to “push” particles .. then by measuring the direction and distance they were pushed. Different particles resonating at different wavelengths could be used to determine colour.
But again – its a quantum solution not a digital one.
I always thought that the whole Foveon thing was totally exaggerated and that it had no future. Just like so many people that have never actually used a Foveon sensor. Until I bought a Sigma DP2 Merrill and processed the first photos taken with this sensor. I was blown away by the detail and color accuracy of this sensor. I’ve used many cameras! Including the Nikon D800, but I never saw this kind of per pixel detail. To me, Foveon is the way to go and I can’t wait to get my hands on the first full frame version. The APSC version delivers awesome image quality. Just imagine what a full frame three layer 24mp Foveon sensor can do…
A high resolution full colour sensor with high sensitivity in full frame 35mm would outclass current digital medium format cameras. IMO.
If I’ve been reading correctly, Sony wants to come out with a game-changer. To do that, they need something other than just an incremental advance – and using a 36 megapixel sensor with video that does autofocus is only incremental. Something new and radical in a sensor might do the trick. Also, a reasonably-priced 4k video capability would be a potential game changer (if they can bring the cost of a tv down to 1/3 of current). A sensor that departs from Bayer, but with nearly equal or better iso might do the trick too. On-chip focussing is a must – get rid of the mirrors.
That’s just my opinion – hope you all have a fine day!