New Sensor technology coming out in the next 18 months?

Started Dec 31, 2012 | Discussions thread
Joseph S Wisniewski Forum Pro • Posts: 34,130
But the company went under...

Barry Fitzgerald wrote:

Joseph S Wisniewski wrote:

He knows he can't show what doesn't exist. So, he has to fight on other fronts.

Foveon sensors are already in production. (it very much does exist)

Need I remind anyone

It only takes more development to overcome the limitations of that design.

No, Barry, it doesn't.

The noise is a characteristic of the differential design and the large off axis matrix coefficients that are needed to cope with the non-colorimetric characteristics of silicon depth penetration. The color accuracy problems, according to the sensor's own creators, would require doubling the layers to fix. That's 4x the noise, because of the math that you need to deal with the differential layers and the high numerical gains in the color processing. Again, things spelled out in their white papers.

Regardless bayer is only the first step in a long history of what will be sensor technology, it has obvious limits itself and cannot be considered to be an ideal solution.

"Ideal" is immaterial. What blocks acceptance of the Foveon sensor is that it's not "sufficiently better" than CFA sensors to attract camera makers.

  • Foveon claimed it would cost less than competing sensors, because it didn't need the organic processing steps that lay down micro lenses and color filter arrays. This did not happen. The deep silicon structures it needs cost more to fabricate than the organics on a CFA sensor. And the first "no organics at all" version used on SD9 failed in the market, and they had to add an organic step to add micro lenses to the version used in SD10.
  • Having equal resolution in all three color channels turned out to only be useful in a limited number of scientific applications. For most photographic purposes, it's inefficient. The human eye has a 30:1 disparity between its most and least populated colors. Color film has about 4:1 disparity in grain density. Foveon's equal density for all colors is a step in the wrong direction. Bayer didn't go far enough with his most popular design, the RGBG pattern with a 2:1 disparity. Some of his other patterns, like the family of pseudorandom patterns in one of his later patents, range from 2.5:1 to 8:1. Fuji adopted a 2.5:1 disparity for their xtrans sensors.
  • Yes, liveview and movie modes were both easier to implement on the Foveon. It's why we used it for Cinci. But Sigma flushed that advantage down the toilet. Camera after camera came out without it: SD 9,10,14,15,1. When they implemented it in the first gen DP models, the company that designed and built those for Sigma didn't even bother to do liveview or movie mode using the Foven's special binning modes, they used the same off-the-shelf row and column skipping they were already using for Bayer sensor cameras. Again, Foven's strongest advantage gets wasted.

Merely a matter of time before these 3 or even 4 layer sensors get better to the point where they are appealing for mainstream use

Again, it isn't.

Dr. Merril said it needs 6 layers to be competitive with CFA sensor color accuracy. Some fairly simple math says that this is going to put it at about a 4 stop disadvantage over a CFA sensor.

-- hide signature --

Rahon Klavanian 1912-2008. Armenian genocide survivor, amazing cook, scrabble master, and loving grandmother. You will be missed. Ciao! Joseph

 Joseph S Wisniewski's gear list:Joseph S Wisniewski's gear list
Canon EOS 5D Mark II Nikon D90 Nikon D2X Nikon D3 Nikon D100 +43 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow