If the new camera makes 12 MP photos at 1:1:1, then what should be next?

Started 1 week ago | Discussions thread
Johan Borg Senior Member • Posts: 2,664
Re: I'm wrong but...

Tom Schum wrote:

FDecker wrote:

May I add, the data is not redundant. We need 3 color values and we get 3 values from the Foveon layers which as you correctly say must be decomposed by a linear matrix calculation.

From each pixel location we get two values. From adjacent locations with different color filters on top, we can interpolate or somehow mathematically guess three values per pixel location. I guess this is the linear matrix calculation you are talking about.

This sensor would be a hybrid between Foveon and Bayer. Since there are only two layers, it is likely going to exhibit lower noise than any Foveon. I think and maybe I am mistaken that the lowest layer gets the fewest photons and is the noisiest. This two layer design would remove that noisy layer.

Do we all share in this idea? If so let somebody in this thread volunteer to pay the cost of filing for a patent and I'll be happy to share in the royalties for no charge!

If I didn't bother patenting the X-Trans pattern, I think I'll pass on this one


Now, while the lowest layer is the noisiest, there's no guarantee that calculating the value it would have had if it was measured directly is going to be any cleaner. What may help, though, is that by using a filter, the cutoff between Green and Red is going to be much more controlled than from silicon, so the red value is actual red instead of Foveon "red". Noise in Foveon photos is exaggerated by the strong transformations that take place to generate a proper RGB representation of the data.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow