Superior Technology: Foveon X3

Why X3 is Better

To capture the color that other image sensors miss, Foveon X3® direct
image sensors use three layers of pixels embedded in silicon. The
layers are positioned to take advantage of the fact that silicon
absorbs different wavelengths of light to different depths. The bottom
layer records red, the middle layer records green, and the top layer
records blue. Each stack of pixels directly records all of the light at
each point in the image.

Until now, all other image sensors have featured just one layer of
pixels, capturing just one color per point the image. To capture color,
the pixel sensors in CCD and CMOS image sensors are organized in a
grid, or mosaic, resembling a three-color checkerboard. Each pixel is
covered with a filter and records just one color-red, green, or blue.

That approach has inherent drawbacks, no matter how many pixels a
mosaic-based image sensor might contain. Since mosaic-based image
sensors capture only one-third of the color, complex processing is
required to interpolate the color they miss. Interpolation leads to
color artifacts and a loss of image detail. Blur filters must then be
used to reduce color artifacts. The use of blur filters adversely
affects sharpness and resolution of the final image captured.

With its revolutionary process for capturing light, Foveon X3
technology never needs to compromise on quality, so you get sharper
pictures, truer colors, and fewer artifacts. And cameras equipped with
Foveon X3 technology do not have to rely on processing power to fill in
missing colors, reducing hardware requirements, simplifying designs and
minimizing lag time between one shot and the next.

Dollar for dollar, pixel for pixel, nothing compares to Foveon X3
technology.

http://www.foveon.com/article.php?a=69

 

Snore...

 

To capture the color that other image sensors miss, Foveon X3® direct
image sensors use three layers of pixels embedded in silicon. The
layers are positioned to take advantage of the fact that silicon
absorbs different wavelengths of light to different depths. The bottom
layer records red, the middle layer records green, and the top layer
records blue. Each stack of pixels directly records all of the light at
each point in the image.
<<<<<<<<<<<<<<<<<<

Nice theory. Except that it doesn't work that way in real life. The color
separation is poor, and decoding color introduces noise. Foveon has to
interpolate its color vertically, whereas Bayer has to interpolate it
horizontally. Six of one, half a dozen of the other.

Also, you have to store three separate charges at each pixel, so the well
depth is that much smaller and the noise that much larger. Strike two.
That approach has inherent drawbacks, no matter how many pixels a
mosaic-based image sensor might contain. Since mosaic-based image
sensors capture only one-third of the color, complex processing is
required to interpolate the color they miss. Interpolation leads to
color artifacts and a loss of image detail. Blur filters must then be
used to reduce color artifacts. The use of blur filters adversely
affects sharpness and resolution of the final image captured.
<<<<<<<<<<<<<<<<<<<<<

This is lying FUD. Foveon sensors also require antialiasing filters to avoid
Moire and other aliasing artifacts just as Bayer sensors do. Look up the
mathematics of discrete sampling and the Nyquist theorem.
With its revolutionary process for capturing light, Foveon X3
technology never needs to compromise on quality, so you get sharper
pictures, truer colors, and fewer artifacts.
<<<<<<<<<<<<<<<<

Snake oil. In real life you get bogus snap-to-grid detail, poor color
reproduction, and more artifacts, since none of the Foveon cameras have the
mathematically required low-pass filter.
And cameras equipped with
Foveon X3 technology do not have to rely on processing power to fill in
missing colors, reducing hardware requirements, simplifying designs and
minimizing lag time between one shot and the next.
<<<<<<<<<<<<<<<<<<<<

Now this is simply a lie. Foveon cameras require three times the data
storage and processing capacity of Bayer cameras (of the same pixel count),
since they have to store three values for each pixel in the RAW format and
have to perform complex processing to decode the color.

David J. Littleboy
Tokyo, Japan

 

yeah, must be why its doing so well

piss off you dickhead!

 

Er, ah, gents: this is the same kook who's been posting and cross
posting here under a variety of names, some made up, and some spoofed.

This was not from G. Burnore. Groups adjusted.

Please help and do the same.

 

Ouch! That'll sure drive him away.

 

Yes, a particularly vicious attack! ;-)

 

No, it is doing poorly due to a camera related marketing choice.

Foveon licensed their sensor technology exclusively to Sigma. Sigma took
a $1,000 sensor, wrapped a $300 camera around it tried to sell it for $1,500.

A Foveon sensor will outperform a Bayer type sensor in the same application.
That doesn't mean that a Bayer type sensor will not work well if properly
constructed in a properly constructed camera with properly designed software.

My late HP 2 megapixel digital camera produced high quality A4 (8x10's for
those of you who think that way). It was difficult to use with a long delay
between pressing the shutter and the exposure. But the picture quality
was excelent. An autopsy revealed that parts of the camera were made by
Konica.

Modern digital cameras with Bayer sensors produce good results and so would
a Foveon based camera if it had been properly designed, had really good
lenses and decent software.

Quite possibly a Foveon based Nikon or Canon full frame camera would
have taken the market by storm and blown away the competition. Since
they never had a chance to produce one, they used what they could get
and have produced excelent cameras without them.

When you have technology to sell, choosing your "strategic partner" is
far more important than people think.

Geoff.

 

The Sigma cameras are far better than you're implying. Perhaps Foveon
licensed their technology to Sigma because no other company was interested?

So "would" a Foveon based camera? So "does" a Foveon ased camera. Have you
ever seen any images produced by the Sigma cameras? Doesn't sound like it
to me. Sigma makes some very, very good lenses and the Photo Pro software
is very good. Your implication that the lenses and software are not up to
par is wrong.

Canon has invested too much in their Bayer technology to start with
Foveon's. Hasn't hurt their standing in the market. Why fix something that
isn't broken?

Again, perhaps there was no other choice. Foveon needed an entrance into
the market and they took what they could get.

 

Canon has had a patent on it's "foveon like" sensor for some time. It's
single layer pixels can be sensitive to r, g or b at any given time. They
then cycle this sensitivety at extreme high speed during the exposure.

Sounds cool, but when or if we see it?????

 

Based from what I've seen at the various posting sites, the Foveon
works well. Better than the POS Bayer sensor does, anyways. For first
generation, it does very well. They are having trouble scaling it up,
I hear.

Fuji has a new sensor coming out, (as in five years or so) based on
organic sensors, that capture all the color information as well. Don't
know if it will need anti-aliasing, etc, but at least it does not guess
about the missing color, and then fill it in with crayon, as the Bayer
does. Perhaps it will solve some other problems digital has as
well, such as bit depth, and loss of fine detail. Yeah, those sky's
sure are noise free, right? Just like the ones I paint with my
watercolor set!

Can't wait for good digital, but for now, slide film, scanned, is the
way to go for me.

 

[troll target groups removed from followups]

**** off, Bowtie.

 

May we live in interesting times! As far as photography goes, we are.

 

Actually, it is an analog of how color film works. Color
film also has three layers, each of which responds to one
color. And there was a B/W film with a very wide dynamic
range. It was made up of three layers with widely different
ASA ratings.

 

Gary wrote:

NOTHING ...

Gary Burnore posts from databasix; not from gmail. This is some netkook
just trying to start trouble.

 

The Foveon sensor had problems with poor colour (not really
that much difference in colour response between layers, so getting
enough saturation requires large negative values in the colour matrix),
inaccurate colour (a special filter is needed to get human-eye-like
colour matching), and non-uniformity of response across the sensor.
If all these were somehow fixed, *then* a Foveon sensor might have a
chance of outperforming a Bayer sensor at the same pixel count.

But Foveon sensors have remained at 3.4 Mpixels max, while Bayer sensors
of 6, 8, or more Mpixels are now the standard in DSLRs. So Foveon
cannot come close to the existing Bayers in resolution. That would also
have to be addressed to make a Foveon sensor competitive. Because of
the extra complexity, the Foveon sensor must be more difficult to scale
to higher density than a Bayer sensor is (either CCD or CMOS).

Sigma added to the woes by omitting an antialiasing filter and providing
a camera body that couldn't provide anything but RAW output. These
could have been fixed by a more competent camera vendor, but the sensor
problems would remain.

And if all of the sensor problems above were somehow dealt with.

Dave

 

So with sigma getting hold of it it was pretty well bloody doomed from the
word Go then.

 

I don't suppose that was the fault of the body - more the fault of software
designers that were clueless.

 

In message <>,

You have a very strange understanding of what a digital image is. A
digital image is a discreet-point representation of a continuous analog
focal plane. As such, the analog original is always undersampled,
regardless of whether the system uses a bayer CFA, a foveon, or some
imaginary sensor that that stores the wavelength of every photon that
strikes each pixel. A 12 MP bayer sensor doesn't "fill in with a
crayon" any more than any 3MP sensor of the same size does, regardless
of the type. What it boils down to is resolution.
--

 

In message <>,

Bayer has three layers, too. They're staggered.

Film grain in the three layers are not aligned like the Foveon, and
aren't subject to aliasing as in the Sigma implementation.

High-DR can be done with digital and color filters, although the color
response will be different in the different ranges.
--