The big claim to fame of Foveon is greatly reduced red and blue aliasing and
the aliasing is
consistent across the colors, so less objectionable. This also then does not
require as much
anti-alias filter. The camera resolution is then the same independent of color,
unlike Bayer where
the red and blue resolve only half that of the green. As you allude to ,the
downside is the Si
absorbtion "filters" are not nearly as sharp as discrete on chip filters , so
the reconstruction
algorithms to seperate colors, adds noise over the inherent sensor site noise.
This at least partly negates the theoretical huge signal to noise ratio
advantage, of not throwing
away more that 2/3 of the photons, that occurs in a conventional Bayer sensor.
You can think of
the mathematical color seperation partially requiring the differencing and
scaling of combinations
of three large numbers, so the errors (noise) in the large numbers gets
magnified by the
differencing operation.
Tim Hughes
--- Winsor Crosby <wincros@xxxxxxxxxxx> wrote:
>
> I have read just the opposite even on the foveon forum sites that the
> color is not so great. I think that may be because of the thin
> filtration required to get to the bottom sensor. I don't know the
> order of the filters, but to explain what I mean suppose you have top
> to bottom, green, blue, and red. If the filtration is efficient you
> would not expect any red light reaching the bottom. So it must be a
> little green, blue, and red mixed with white light that gets through.
> Probably harder to guess the true color based on that.
>
> Certainly the review on dpreview indicates that the real resolution
> of the 3.5 MP sensor is about equivalent to an 8MP Bayer pattern
> sensor. Yeah, I think the Sigma camera is just a so so camera. It
> will be interesting to see whether there is any progress in the new
> Sigma camera and its sensor. There is a teaser up, but no details.
>
> The other thing that no one seems to talk about is noise. The last
> foveon sensor had a pixel pitch substantially larger than the pitch
> in a 6MP APS-C sensor, yet the noise level was not much different
> than that in a D100. You would think that it should be much better.
> It probably is related to the small amount of light pushing through
> the upper filters to the buried pixels, but I don't know for sure.
>
> I sort of think that the Foveon sensor is not the huge breakthrough
> that enthusiasts think. Otherwise version 2 of the sensor would have
> been to market a couple of years ago and the big guys would be
> scrambling because of losing sales to Sigma.
>
>
>
>
> Winsor
> Long Beach, CA
> USA
>
>
> On Sep 13, 2006, at 10:28 AM, khen lim wrote:
>
> >
> > I myself am unsure if the Foveon sensor is going to be the right
> > way to go.
> > Granted that it produces very beautiful colours - a trifle bit
> > exaggerated
> > but still nice, IMHO - there were issues with it.
> >
> > K.
> >
> > On 13/09/06, John Hermanson <omtech1@xxxxxxxxxxx> wrote:
> >>
> >>
> >> As I recall though, any camera(s) with the foeven sensor (Sigma?) got
> >> only mediocre test results. Am I missing something here?
> >>
> >> ___________________________________
> >> John Hermanson
> >> Camtech Photo Services, Inc.
> >> 21 South Lane, Huntington NY 11743
> >> 631-424-2121 | Olympus OM Service since 1977
> >> http://www.zuiko.com | omtech1 AT verizon.net
> >>
> >>
> >> Moose wrote:
> >>> khen lim wrote:
> >>>
> >>>> John
> >>>>
> >>>> I will be frank with you. Olympus HAS NOT FINALISED/CONFIRMED
> >>>> actual
> >> Mpix
> >>>> count and neither have they FULLY CONFIRMED/SIGNED OFF the
> >>>> choice of
> >> image
> >>>> sensor. There are "several" E-3s available at Shinju-ku under
> >>>> cloak and
> >>>> dagger. It's now up to them to drag it to the eleventh hour,
> >>>> then hold
> >> their
> >>>> breath and make the decision they will have to live for the rest of
> >> their
> >>>> lives. In and amongst the selection of E3s is one that has a
> >>>> 10.2Mpixphoto
> >>>> sensor. None of them have a Kodak sensor. There are surprises in
> >>>> there.
> >> I
> >>>> hope they make the right choice; the one I have in mind because
> >>>> that
> >> sensor
> >>>> will produce absolutely stunning natural colours.
> >>>>
> >>>
> >>> I'll toss my guess into the ring that the 'stunning natural colours'
> >>> entry is a Foveon.
> >>>
> >>> I prefer AG's alternative, though, as I think a 4/3 development
> >>> using
> >>> the technology of the Fuji sensor system in the F30 could be
> >>> stunning.
> >>>
> >>> Moose
> >>>
> >>
> >>
> >>
> >> ==============================================
> >> List usage info: http://www.zuikoholic.com
> >> List nannies: olympusadmin@xxxxxxxxxx
> >> ==============================================
> >>
> >
> >
> >
> > --
> > Khen Lim
> > XIOS Network Solutions
> > IBM Business Partner
> > +60 +16 528 6010 / 016 528 6010
> >
> >
> > ==============================================
> > List usage info: http://www.zuikoholic.com
> > List nannies: olympusadmin@xxxxxxxxxx
> > ==============================================
>
>
> ==============================================
> List usage info: http://www.zuikoholic.com
> List nannies: olympusadmin@xxxxxxxxxx
> ==============================================
>
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|