on 12/12/02 7:55 AM  michael shaffer wrote:

> I'll have to take a close look at the URL ... but correct me if I'm wrong.
> What you're saying is ... three 12bit channels (or 10bit, or 14) are somehow
> encoded into a single 16bit channel(?) ... and this would ^not^ work if the
> camera captured true 16bit(? ... i.e., three 16bit channels cannot be
> encoded into a single 16bit channel, without throwing away (bits of) info
> ...)

Ummm no - what I was trying to say is that each pixel receptor (or whatever
you want to call them) on the sensor only sees R, G or B depending on the
filter over it. So there's only the need to record a single channel of info
for each position. The other two channels for each pixel are computed later
in processing by looking at that pixel's neighbors.

The only chips so far that gather 3 color info for each receptor are the
Foveons. (And of course the scanning backs.) Some high end backs shift the
chip about in 1 pixel increments to build up 3 color info for each position
in a series of 3 or 4 exposures.

-- 
Jay Busse


===============================================================
GO TO http://www.prodig.org for ~ GUIDELINES ~ un/SUBSCRIBING ~ ITEMS for SALE

Reply via email to