Reading mode: Light Dark

Official Sigma SD-9 Announment

Go to last post
The Bayer CCD needs twice as many photosites to get the same resolution as a scan back or Foveon chip. I am talking in a linear line. The side by side placing of red, green and blue in a square pattern of four pixels can only resolve one line with an approximated color and density. The scan back or Foveon need only one pixel to get a correct color and brightness. The obvious absence of moiré and softness of the assumed Bayer guessing is enough to get an image similar to this Bayer pattern if only half the pixels are there. 500 pixels though still need to be interpolated to an image of 1000 that are seen on the Bayer CCD. The interpolation with software may cause a fuzzy image no better than the Bayer CCD image but there will be no artifacts and the color will be quite pure. Chroma and other lens aberration will not get worse like on the Bayer CCD. If Purity is an unimportant goal as many image colors are subjective at best, than the only real benefit is the image without moiré. When the Foveon sensor gets to be the size of the Bayer CCD and have the same pixel count, the Foveon will totally obliterate any Bayer CCD in its class. It will have twice the resolution and have pure color. Keep in mind that I assume that the other important factors in imagers are on an equal keel.
Rinus of Calgary
 
All DSLR's have an antialiasing filter that reduces the effect of color banding in high-frequency contrast areas (the Kodak DCS series had a removable AA filter, but color fringing got pretty bad when it was removed). This has the double effect of removing the color fringing and also softening the image. This is why native D30/D60/D100 images are all slightly soft straight out of the camera. With the new Foveon sensor, there is no need for this filter, so theoretically, images should be pin-sharp straight off the sensor. This will make for a HUGE increase in file quality and enlargeability.

I'm actually surprised this isn't sited repeatedly why the Foveon sensors really will be enormously better than the current Bayer/AA system. It should stomp current 3mp sensors, and will probably be much better than current 6mp sensors. Remember that AA filter is actually effectively reducing the resolution of that 6mp sensor by blurring the image ever so slightly.

In the end though, it is NOT all about image quality. There are so many factors that go into a camera like this that a small failure at any point in the chain could reduce this camera to the bargain bins. People love the D30/D60 for example, but continually scream bloody murder about the AF system. No one knows how this camera will actually perform. I own a 4x5 camera with a Schneider 110 f/5.6 lens whose images would stomp every single DSLR out there, but I rarely use it because it's such a pain. The image quality is there, the usability is not.

Of course I'm often wrong.

Steve
splutphoto.com
David Barkin said:
I followed this question of 3MP on a number of threads in the Pro
section. Not that I understood them!

However, basically, by eliminating the need for Bayer pattern
interpolation, the 3MP of the Sigma is the equivalent of 6 MP on a
Bayer sensor. In addition you eliminate Bayer patern moire and
artifacts.

If anyone wants they can do a search. These discussions peaked in
July of this year. If anyone wants an analogy, which car is faster?

A v8 with a one speed transmission or a 4 cylinder engine with five
on the floor?

That's how great the difference is.

Dave
Click to expand...
 
Yes Phil answered the question of why be excited over a 3MP camera, it is better than other 3mp cameras. But it is really better then 6MP cameras which are its competition.
Ulysses said:
The mosaic system works. But we need a new way of
constructing the picture,
Click to expand...
Why? Show me an image from a D60 and tell my what is wrong with it and why it NEEDS to be done a different. I am all for new technologies, but I don't like absolutist statements with little basis in fact. Bayer interpolation artifacts become smaller and less releavent as Pixel count increases.
Member said:
as well as a somewhat new thought process
as users as to how we arrived at the image.
Click to expand...
Huh? You are really losing me here. Usually you are pretty clear and logical, but here you are saying users are in need of reprogramming or something. Why should a user even care how the image is assembled?

See the Doc Ulysses, I think you may be coming down with Fo-mania.

Peter
 
Peter Guidry said:
Why? Show me an image from a D60 and tell my what is wrong with it
and why it NEEDS to be done a different. I am all for new
technologies, but I don't like absolutist statements with little
basis in fact. Bayer interpolation artifacts become smaller and
less releavent as Pixel count increases.
Click to expand...
Somebody was saying that the Foveon X3 might handle colors
better. That is, more accurately, et cetera.

And then there's the cost issue. Maybe the manufacturing
costs of the X3 will be "different" than for a CMOS chip in a
D60, or the CCD sensor in some other camera.

I say "different", because I am not knowledgable enough to argue
why the manufacturing costs would be lower, and perhaps they
might even be higher.

But the possibility that the manufacturing costs for the X3
will be "different", that seems quite possible. And that could
be the important thing, in the end, even if the resulting images
are the same.
 
Peter Guidry said:
See the Doc Ulysses, I think you may be coming down with Fo-mania.
Click to expand...
Okay, I know you're saying that with a smirk on your face, but you know me better than that if you've followed my post history. No need to get mean about it. :-)

I'm not saying anything that Phil hasn't already explained himself.

The mosaic/Bayer method is a good one. It's proven. However, you can't get around the fact that that no matter how many pixels you put on the chip, it's still manufacturing a huge portion of its final output. That's okay, and in practical application and the print sizes that most folks produce, it may not matter for some folks.

What the Foveon process brings to the table is a more efficient process, along with potentially faster processing, besides the better resolution than its "mere" 3 megapixels would at first indicate.

Again, this is not new information. Phil has also mentioned the point about re-thinking how we think about what makes up a pixel. I'd tend to agree.

--

Ulysses
 
Hi Rich

What you are saying is true, but only as far as it goes. Here is a camera with a probable street price of $1500. To expensive for the casual user but fairly cheap for a serious photographer.

Ok, I have 4 Sigma lenses for my Nikon. I could use a back-up camera. I buy the Sigma - Do I then blow another $3000 in lenses?

So I understand the dilemma that Sigma is in. They want to sell the camera and they want to sell lenses. But will a serious photographer invest in this camera if they also have to purchase new glass? I don't think so. I certaainly would consider this machine if my only problem was buying the camera.

Get my drift here. How many cameras will they sell if you have to buy a Sigma lens?

Off hand I would think they'd make more in the long run by selling the adapter

Dave
Rich said:
David

While it would be great those of us considering an sd9 since it
would give us access to a much wider choice in affordable used
lenses, it would make no sense for Sigma. I think they want to make
money on the lenses, not the camera. Selling a retrofit mount would
discourage the purchase of their new lenses and would thus equal
lost revenue.
Click to expand...
 
When the X3 was first announced, I was so impressed by the concept behind the sensor, and i have little doubt it will really be that much better than conventional CCD/CMOS design.

But, Sigma? I mean, Sigma, among the third part lenses, is not bad, but I doubt X3 will all of a sudden make Sigma the 'brand of choice' by professionals or even serious amateurs.

Buying an extra body is no big deal, relatively, but buying a new system just for one chip is a big deal. I own 8 Canon lens (all quality lenses, 6 of them are L lenses), and a few Nikkor (left over when I convert over to Canon)... that is the killer. I would love to buy the Sigma for a spin, but what about the lenses?!

If Foevon can get Nikon or Canon on their boat, then that is a whole different story. I would probably prefer a 8MP X3 Canon over the about-to-release, 11MP 1Ds.

ws
 
As I recall, serious photographers (not specifically professionals) average
a major camera purchase about every 5 years -- and that was an old
data point for 35mm film systems. It's probably been more often for digital
cameras. And there are some voids in the market -- Minolta DSLR, Pentax
DSLR, Leica, etc. Seems reasonable that Sigma has a good-sized market
to exploit if the product is really good. I have some Canon lenses around that

I'll give up if I'm sufficiently impressed with the Sigma. There are, of course,

others who have a bigger investment in lenses -- they should just stick with their
system. Can't see the downside of a Sigma camera with a Sigma lens mount.

Another 2 cents.
Darrell
 
Hi Darrell,

Do you mean pro serios photogs change their system or bodies every 5 yrs? body i can understand, lens, system, i cannot! i shot with Nikon for 12 yrs before taking the plunge with Canon...

but anyway, the SD9 is an tempting alternative, and afterall, we consumers, will all benefit from more competition.

ws
Darrell Spreen said:
As I recall, serious photographers (not specifically professionals)
average
a major camera purchase about every 5 years -- and that was an old
data point for 35mm film systems. It's probably been more often
for digital
cameras. And there are some voids in the market -- Minolta DSLR,
Pentax
DSLR, Leica, etc. Seems reasonable that Sigma has a good-sized market
to exploit if the product is really good. I have some Canon lenses
around that
I'll give up if I'm sufficiently impressed with the Sigma. There
are, of course,
others who have a bigger investment in lenses -- they should just
stick with their
system. Can't see the downside of a Sigma camera with a Sigma lens
mount.

Another 2 cents.
Darrell
Click to expand...
--
WSLam
 
The E-Bay phenomenon makes this whole issue of commitment to a lens system much less important. Used EF lenses sell quickly for barely less than they cost new. It's really not a big deal anymore to dump one system for another.

Anyway, at that price point, who says you have to jump? Sigma does make a few cheap, decent lenses; if the Foveon is even half what it's hyped to be, they'll sell every camera they make -- many of them to folks who are hanging on to their Canons and Nikons and just want to try out the latest thing.
WSLam said:
When the X3 was first announced, I was so impressed by the concept
behind the sensor, and i have little doubt it will really be that
much better than conventional CCD/CMOS design.
But, Sigma? I mean, Sigma, among the third part lenses, is not bad,
but I doubt X3 will all of a sudden make Sigma the 'brand of
choice' by professionals or even serious amateurs.
Buying an extra body is no big deal, relatively, but buying a new
system just for one chip is a big deal. I own 8 Canon lens (all
quality lenses, 6 of them are L lenses), and a few Nikkor (left
over when I convert over to Canon)... that is the killer. I would
love to buy the Sigma for a spin, but what about the lenses?!
If Foevon can get Nikon or Canon on their boat, then that is a
whole different story. I would probably prefer a 8MP X3 Canon over
the about-to-release, 11MP 1Ds.

ws
Click to expand...
 
The only people who need to do any rethink are the ones who look at pixel counts and nothing else. I am definitely not one of those. Ever since my first digicam, I noticed images were lacking in detail and still are.

When this was first announced I examined the benefits and even wrote a Bayer and Debayer program to create artifacts at will, so I am quite cognizant of the benefits this technology will bring. I also examined a number of images and came to the conclusion that Bayer loses detail to the extent that it is approxiamate to non bayer of half the pixel size. Even Foveon was using a 2X factor for the detail improvements.

So when 3mp X3 doubled == 6MP bayer. Then there is really not that much to get excited about IMO, when there are already 6MP bayer cameras. It will be neat to see the tech in action warts and all.

Now if they build a 6MP X3 Camera. That would truly be something to get excited about.

Right now there is an almost religious following around this technology, it is merely technology and it will have its own set of tradeoffs.
Ulysses said:
The mosaic/Bayer method is a good one. It's proven. However, you
can't get around the fact that that no matter how many pixels you
put on the chip, it's still manufacturing a huge portion of its
final output. That's okay, and in practical application and the
print sizes that most folks produce, it may not matter for some
folks.
Click to expand...
I fully acknowledge that. I would give X3 at least a 2 x advantage. But why give up that advantage by build your sensor with half the pixels of the competition?
Member said:
What the Foveon process brings to the table is a more efficient
process, along with potentially faster processing,
Click to expand...
This is in the area of Foveons myths. There is no high cost to doing Bayer interpolation. You can buy a Canon A200 for under $200 that does perfect and speedy bayer interpolation.
 
Peter Guidry said:
So when 3mp X3 doubled == 6MP bayer. Then there is really not that
much to get excited about IMO, when there are already 6MP bayer
cameras.
Click to expand...
Let's wait to see the real pictures from the SD-9. Maybe there
are other benefits, beyond simply resolution (i.e. colors).
 
Peter, the problem is that you can't reason with people that have Fo-maina. They don't WANT to know the truth that there will be little difference between a 6MP Bayer and a 3MP X3 and that the 6MP are already in production and becoming reasonably easy to get.

They even attribute things to it that it cannot do such as violating all laws of sampling theory as in not needing an Antialiasing filter (I guess Nyquist Rate no longer applies).

Then there is the claim that it is cheaper to make, even though it is a more complex process than say Canon uses with their CMOS devices.

I guess they also ignore "little" drawbacks like only going to ISO400 as is currently Spec'ed on the SD9. There was a lot of squawking in the Canon SLR forum when the D60 went from ISO1600 of D30 to ISO1000 on the D60. The Fact that people with the D60 are not have a lot of problems that they see caused by Bayer I guess is irrelavent when you have Fo-mania. IF the image is so good with the SD9, why then no ISO800 and ISO1600? -- My guess is that it MIGHT have problems like the Contax did.

The problem for the SD9 is going to be convince people that will actually pay $1800 (we will see how much it gets discounted) for a DSLR Body to buy it. $1800 is a lot of money for somebody that is not serious about photography. If they are serious, then they are more than likely invested in Canon and/or Nikon lenses. The question for many people will be what can the SD9 do that a D60/D100/S2 or the like cannot.

It seems to me, judging from the poster and the topics that they post in, that most of the people with Fo-mania don't have a DSLR and have not seen what a 6MP Bayer can do or where a 6MP Bayer is lacking.

Karl
Peter Guidry said:
Ulysses said:
To paraphrase Phil: Why do you need 6MP to get a good image?
Click to expand...
I don't think anyone is disputing that fully colour pixels capture
more detail mosaic created pixels.

But it doesn't, as many with Fo-mania seem to suggest(I don't mean
you BTW), mean that it is better than ANY amount of mosaic pixels.
There are a lot of indications that 3MP x3, is more or less
equivalent to 6MP bayer.

In which case there are a whole lot of other elements in
consideration, like sensitivity/noise, which is unknown. The
overall execution and usability of the body, which is uknown. In
reality these things will be more important.

Except to the immature owners of each sensor tech which will trade
childish jabs of "I have more pixels, and my pixels are better".

Anyone remember the Feb news break and the sillyness of these
boards, scores of people claiming Bayer cameras were now worthless,
digicams sales would stop while we awaited the revolution. I tried
arguing for reason, but was painted a luddite and obviously a big
stock holder of sony, canon etc..

The thing I find the most amusing, is the inconsistent logic of
many who argued for the X3 Revolution. Many were the same people I
had previously argued against, when they claimed no-one really
needed more than 3MP bayer cameras, anything beyond that was
overkill for the average user printing up to 8x10.

From my perspective it seemed totally inconsistent to one day state
that anything beyond 3MP bayer is overill and the next proclaim it
dead.

I expect Fo-mania will return in full force once there is a review,
any balancing factors will be ignored and the death of bayer will
be proclaimed yet again.

The sigma release will probably start one of the longest running
arguments (6mp bayer vrs 3MP x3) that will never be settled because
many people see what they want to see, especially after they pay
their money. The recent thread with some pretty awful images
purporting to be from an SD9 had many proclaiming they saw "special
qualities", better colour and dynamic range etc, in what were
really complete drek, that my coolpix 950 could have done better.

Also I must have some problem as well, because current 6MP DSLR
images seem to be Fantastic to me, I haven't seen a sample with
obvious artifacts and they have great colour.

Then again I do see artifacts in Fuji Honeycomb images, which the
people on the Fuji forum tell me aren't there, or will dissapear
with printing.

My take on the X3 thing has always been, it is a great idea, that
eventually should dominate, barring significant issues we don't
know about.

But it will probably spend a long time in the niche category until
the technology is perfected.

Peter
Click to expand...
--
Karl
 
Dave the K. said:
Anyway, at that price point, who says you have to jump? Sigma does
make a few cheap, decent lenses; if the Foveon is even half what
it's hyped to be, they'll sell every camera they make -- many of
them to folks who are hanging on to their Canons and Nikons and
just want to try out the latest thing.
WSLam said:
When the X3 was first announced, I was so impressed by the concept
behind the sensor, and i have little doubt it will really be that
much better than conventional CCD/CMOS design.
But, Sigma? I mean, Sigma, among the third part lenses, is not bad,
but I doubt X3 will all of a sudden make Sigma the 'brand of
choice' by professionals or even serious amateurs.
Buying an extra body is no big deal, relatively, but buying a new
system just for one chip is a big deal. I own 8 Canon lens (all
quality lenses, 6 of them are L lenses), and a few Nikkor (left
over when I convert over to Canon)... that is the killer. I would
love to buy the Sigma for a spin, but what about the lenses?!
If Foevon can get Nikon or Canon on their boat, then that is a
whole different story. I would probably prefer a 8MP X3 Canon over
the about-to-release, 11MP 1Ds.

ws
Click to expand...
--Here are some picture from Foveon.
Click to expand...




 
rbadillo said:
Here are some picture from Foveon.
Click to expand...
Yeah? How come the pool table shot is 1800x1800? Is the Foveon a square aspect sensor now?
Bryan
 
Wow, has anyone actually said that the Foveon 3MP will actually outperform a 6MP mosaic sensor? I may have missed those posts.

What will be interesting to see will be the resolution detail and color resolution of a performing camera, assuming that good glass is actually in front of it.
Karl Guttag said:
They don't WANT to know the truth that there will be
little difference between a 6MP Bayer and a 3MP X3 and that the 6MP
are already in production and becoming reasonably easy to get.
Click to expand...
--

Ulysses
 
Darrell Spreen said:
I think most camera makers have announced at least one model more
than 6 months before samples were available for testing.
Click to expand...
Other than Contax, and maybe Silicon Film, please give an example. I'm not talking about availability of cameras. I'm talking about availbility of sample images shot by independent third parties.

-Z-
 
These are the same pictures that have been on the Foveon site, but I don't ever recall them being at 100%. How long have these full-sized samples been floating around cyberspace? Am I living in the past? Are Fritos still a nickel a bag?

Rich
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top