Panasonic announces new sensor technology: color prism on chip

[Solarflare]

Many people probably know the 3ccd or 3cmos video cameras that have a prism with two chroitic mirrors in them, splitting the light beam into 3 different parts.

Foveon has tried to realize the same thing with 3 layers of pixels on a single sensor; unfortunately this technology has its big disadvantages - very poor high ISO performance, because the lowest of the three pixel layers only receives very little light after the higher layers have processed the light.

Now Panasonic has announced they developed the technology to implement split prisms for every pixel individually:

Panasonic to double sensor?s sensitivity with unique ?micro color splitters? | Photo Rumors

This means:
- no light is lost (like through the usual Bayer colorfilters, or alternative colorfilters like Fuji X-Trans etc)
- no layers of pixels darken the light before processing (like with the Foveon technology)
- no need for an expensive, large, heavy and extremely precise prism

The disadvantages are: 3 pixels per final pixel, one for each color channel. So the glorious 36 Megapixel D800 would be reduced to a mere 12 Megapixel design. However, each of these pixels would be of maximum, ultimate quality, free of any losses thanks to demosaicing, just like with Foveon X3 sensors, or 3ccd/3cmos video cameras, or like with the Leica M Monochrom. And no light would be lost at all, too, allowing drastically better high ISO performance than with the Foveon technology.

All in all I think this technology has the potential to realize what Foveon only promised.

 

[Mully]

Time will tell