Olympus-OM
[Top] [All Lists]

[OM] E1, 4/3 format and Foveon

Subject: [OM] E1, 4/3 format and Foveon
From: Hughes <hi100@xxxxxxxxxxxxx>
Date: Thu, 26 Jun 2003 18:38:35 -0700 (PDT)
C.H.Ling wrote:
>>
The foveon one
http://www.imaging-resource.com/PRODS/SD9/FULLRES/SD9HAWBP1.HTM
The D60 one
http://www.imaging-resource.com/PRODS/D60/FULLRES/D60HMWB.HTM
Before direct comparison the D60 one may need some sharpen as usual.
<<
It sounds like you are trying to compensate for drop off in resolution by 
sharpening! Of course
you also increase noise.

CH also wrote:
>>For the resolution chart look at the comment below you will see most likely
dpreview has wrongly interpreted the resolution on SD9.
http://www.imaging-resource.com/PRODS/SD9/SD9A12.HTM
<<
I think both sites don't have it quite right. Clearly DP reviews claim to a 
resolution beyond
Nyquist is nonsense, but imaging resources claim that Foveon is particularly 
prone to aliasing is
also wrong. Comparing a camera with an anti-alias filter to one without, is not 
a good test of
basic chip technology, although it is fair in a camera comparison. The Foveon 
chip is by design
much less prone to aliasing so works well right up to Nyquist without a filter, 
but is certainly
not immune to aliasing beyonfd Nyquist.  Also the weird color aliasing of Bayer 
is eliminated. Put
another way the Bayer chip will alias much worse (particularly in red and blue) 
without an
anti-alias filter, than the Foveon chip (see previous post for theoretical 
Nyquist comparison) ,or
with an anti-alias filter the Foveon chip needs a much less severe anti-alias 
filter. And yet
another way: for equivalent performance the Bayer chip needs something between 
~2 and ~4 times the
# pixel sites of a Foveon chip depending on colors being imaged.

CH: wrote>>
Anyway the SD9's resolution is better than I have expected (I expect it is
around 4-5MP equ to other DC) it do look similar to D60.
<<

It would be interesting to see these review sites show resolution charts for a 
Beyer sensor
without anti-alias filter for different colors, compared with Foveon.

Dp-Reviews seems to claim that resolving a point (sigproc jargon = "impulse 
response") is better
with Foveon. This maybe partly why they claim resolution beyond Nyquist (?), 
although the
interpretation is false, it is likely true points look better because the 
smearing to adjacent
pixels introduced by Bayer interpolation of non-exsistent color pixels combined 
with anti-alias
filters. 

Regards,
Tim Hughes
TimHughes@xxxxxxxx





< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >


<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz