Has Foveon future?

Like this?
http://www.xs4all.nl/~bvdwolf/temp/crw_0128_copy.jpg
(2048x3072 pixels 3.29MB)

Bart

 

Different file, but equally blurry, noisey, and low res. You need a
serious standards-adjustment, try these...

http://www.pbase.com/image/23326188
http://www.pbase.com/image/23326190

 

So ... you say that the Foveon chip has a 100% fill factor?
No light falls in between the sensors?
Interesting!

And ... you say that the the detection gives you 100% accuracy
regarding the light sent out from the subject?
Interesting!

The Foveon sensor based camera must perform magic indeed.
It must be able to have knowledge about photons that
never have been detected.
Interesting!


Roland

 

Sorry George. We all have different preferences. I like the smooth,
soft and very accurate look of the Bayer picture and I don't like the
harsh look of the SD9. I think the aliasing artefacts make the poor
girls look like they are not real. Look at the hair - it is not a
nice sight.


Roland

 

George,

After looking at his shots and yours, either you need a new monitor or
new glasses.

Phil

 

Just for the record, it's not my image.

All I did was convert the raw file so graciously donated by Michael Quack,
but this time at the actual size and with a bit of sharpening, things that
'George P' conveniently didn't do in his 'comparison'.

I fear even a new monitor AND new glasses won't help 'George P', a lobotomy
might.

Bart

 

Seems so, Bart. Just returning here after a long absence, so I haven't
been "Georged" until today

Phil

 

In message <bp473f$qur$>,

No one is agreeing with that nonsense. Only you are stupid enough,
after all the facts are presented to you hundreds of times, to still
insist that a partial-color spatial witness has no spatial resolution
value.
--

 

That's fine, that's why they make softening filters (and give you a
sharpness slider). They don't make optical resolution filters for 1.58MP
Canon owners, unfortunately .

 

Mine either, all 3 are pro samples.

 

Does it have to sense every photon in the Universe to know luminance where
it samples?

 

No - but to get "known colors" it has to detect all
photons emitted from the source. It also have to
be noise free.

My point was that all devices are approximations.
All detectors are "guessing", more or less.


Roland

 

You don't get it do you? You cannot soften away aliasing artefacts.
As a matter of fact, you cannot remove them at all.

And, the soft apperence of the anti aliased pictures can be
improved with the "sharpening slider" called USM.


Roland

 

Which is an optically perfect solution, of course. Small amounts of USM
are probably not going to destroy the image with artifact, 3-500%, which is
the starting point for the 10D (the blurriest DSLR the 6MP-interpolated
lot). We'll never really know where the finishing point is, because 10D
owners refuse to post full size images.

 

....

....

And if you actually thought about this point in advance, you got it wrong.
Over and over and over you repeat the same wrong idea. Have you completely
ignored my several (very polite - could that be the problem?) explanations for
you.

Once again...A Bayer sensor does *NOT* use a 2x2 array to produce one
pixel or 4 pixels. This is completely wrong. The 2x2 array is useful only in
describing the layout of R, G, and B sensors. A Bayer sensor produces one
pixel for every R, G, or B sensor. The pixel is the combination of the R, or G, or
B information from the sensor, plus data from at least *EIGHT* surrounding
sensors. Hence there is no "RGGB with one G thrown away" as you so
incorrectly assert. So four nearby RGGB sensors should not and do not
produce 4 identical pixels. Do you get it?


--
Dan (Woj...) dmaster (at) lucent (dot) com

"I believe in the kingdom come / Then all the colours will bleed into one
Bleed into one / But yes, I'm still running
You broke the bonds and you loosed the chains
Carried the cross of my shame / Of my shame / You know I believe it"

 

It is better than aliasing artefacts IMHO.

3-500% USM has no meaning. You must say at what radius.
Personally I have made stunning 10D pictures (originals
downloaded from the net) with just radius 0.5 and 25%.

There are tons of full size 10D pictures on the net.


Roland

 

Addressed many times: no matter how much you shuffle and reshuffle the same
optical data, optical limiting factor is the number of RGB sensors. Yes,
some mathematical Bayer interpolation routines are better than others,
because they get a little closer to that limit.

 

Rearranging all of the pixels in the image based on a computer algortihm
that can't see, is better than an optical result?

 

Quite clearly, yes it is!

http://img.dpreview.com/reviews/SigmaSD9/Samples/Compared/4way/crop_sigma_sd
9.jpg

 

In George's world, 9 lines are the same as 5.


http://img.dpreview.com/reviews/SigmaSD9/Samples/Compared/4way/crop_sigma_sd
9.jpg