Reading mode: Light Dark

Fovean

Go to last post
PHGN said:
How many people here disagree with the labeling of Fovean resolution?
Click to expand...
I think calling it a 10MP camera is stretching the truth to breaking point, yes, but that's largely because the marketplace is conditioned to the Bayer point of view and is confusing pixels in terms of the sensor with pixels in terms of the image file. Perhaps we should start quoting two values for each camera, MPixels for the number of pixels in the file (and the 2D spatial resolution of the sensor) and MSensors for the number of monochrome sensors used to produce it. So for example, for the EOS-300D we could say: 6.3MP/6.3MS while for the SD10 we could say 3.5MP/10.2MS.

I don't think that's going to help Foveon in the consumer market place though, like MHz in the PC market, the digicam market has been conditioned to think that MP is the be all and end all of a camera - image quality and aesthetics be damned! Like the PC market, there are only a small fraction of customers that currently see through the fallacy of this, and that's the problem Foveon is going to have to tackle. The big question for Foveon is how does the consumer shortlisting based on specs do that? Do they look at the MP rating, the output image resolution or both? And if it's both, how many are going to notice the discrepency between Bayer and Foveon sensors, and what will the reaction and preference be?
 
I don't doubt that it has good resolution... In fact if it weren't for the other flaws of the original I would probably already own one! However, the point is claiming 10 MP is wrong because it isn't really a 10 MP file.
Scott Greiff said:
http://www.dpreview.com/reviews/sigmasd9/page20.asp

Even in a comparison with a Canon 1Ds, the SD-10 holds its own:



While the Foveon may not make show 10M pixels in black and white
resolution, the current crop of 6M (or 11M) pixel bayer sensor
cameras can't claim to offer 6M pixels in color resolution.
  • ScottAG
Pete Perry said:
This Foveon will never achieve 10 Million Pixels in black and white
or color so with everything being considered equal I would say the
Foveon is the one not being honest.
Scott Greiff said:
Pete Perry said:
Standard Sensors are called 6 MP because they're monochromatic so
it captures everything in Black and White... This basically means
that the base image captured is 6 MP so it isn't really deceitful!
Click to expand...
It wouldn't be considered deceitful if the world was monochromatic.
  • ScottAG
--
http://www.greiff.org/scottag
http://www.pbase.com/scottag
Click to expand...
--
'The only real currency in this bankrupt world is what we share
with each other when we're being uncool.' -- Cameron Crowe
Click to expand...
--
http://www.greiff.org/scottag
http://www.pbase.com/scottag
Click to expand...
--

'The only real currency in this bankrupt world is what we share with each other when we're being uncool.' -- Cameron Crowe
 
All of the other sensors are Monochromatic (look it up) and the color patterns are recorded by the CFA and interpolated back in to the image! In short this is why Foveon can claim to be the worlds only color sensor.
Jimmy Chen said:
Jimmy Chen said:
Standard Sensors are called 6 MP because they're monochromatic so
it captures everything in Black and White... This basically means
that the base image captured is 6 MP so it isn't really deceitful!
Click to expand...
Are you talking about a Bayer sensor, or a true monochrome sensor?
I do not know of a Bayer sensor that do not use a CFA.
Jimmy Chen said:
The Foveon Sensor is deceitful because at no point will you get a
10 MP image from the camera! The truth is in raw resolution the
camera fell apart on the test charts because there weren't enough
pixels to clearly define the smaller areas.
Click to expand...
Agreed, that you cannot get a 10.2mp bw image out of a Foveon
sensor, you can only get a 3.4mp BW RAW image. But then, nor would
you get a 6mp full RGB and/or BW image out of a Bayer RAW file
either.

So why it is okay for Bayer sensors to state it is 6mp, when it
only captures 1.5mp worth of full RGB, with 0.5mp extra G ?

Maybe I wasnt paying attention in my 1st grade math classes, but I
can only count a max of 2mp worth of RAW data here for any given
color.

BTW, just how truly BW would that 6mp Bayer RAW image look w/o
interpolation?

--
jc
Click to expand...
--

'The only real currency in this bankrupt world is what we share with each other when we're being uncool.' -- Cameron Crowe
 
Kendall Helmstetter Gelner said:
Joseph S. Wisniewski said:
Kendall Helmstetter Gelner wrote:
And that's where my "megasomethings" rating comes in handy. It
doesn't matter if the
"somethings" are stacked 3 deep (as in a Foveon), if there's two of
them to a pixel (as in the Fuji), or if they're spread out in a 3
color pattern (lieka Bayer) or a 4 color pattern (like a Sony).
Click to expand...
I like the Megasomethings rating (I read that before as well), but
I think the name could be improved as it sounds a bit wishy-washy.
Click to expand...
I had also thrown Pods (Pieces of Data) and MegaPods on the table. For soem strange reason, Megasomethings gained a little more acceptance.
Kendall Helmstetter Gelner said:
That's why instead I tried to make the reader think along the lines
of output quality as best I could...

But you're right about needing something like an MTF indicator. It
sounds a lot more impressive as well!
Click to expand...
Have you looked at the Printapixel thread yet?

--
Ciao!

Joe

http://www.swissarmyfork.com
 
Member said:
All of the other sensors are Monochromatic (look it up) and the
color patterns are recorded by the CFA and interpolated back in to
the image! In short this is why Foveon can claim to be the worlds
only color sensor.
Click to expand...
... and then from the interpolation image then can you get a BW image. Since you brought up RAW data, then please keep the scope as RAW data since interlopation hasnt take place yet. So in other words no Bayer sensor can generate a BW RAW data file either.

--
jc
 
Pete Perry said:
I don't doubt that it has good resolution... In fact if it weren't
for the other flaws of the original I would probably already own
one! However, the point is claiming 10 MP is wrong because it
isn't really a 10 MP file.
Click to expand...
And the Bayer crowd claiming their resolution figures even-though the fake it, is any less wrong?

Sigma should have provided a 10MP upsampled option in camera similar to what Fuji does. You guys would have been happy then.
  • ScottAG
--
http://www.greiff.org/scottag
http://www.pbase.com/scottag
 
If the 3mp x3 is to be referred to as 10mp,and the x refers to the number of layers like I understand it, the new naming makes it a 30mp (10x3). So it seem logical to refer that sensor to a 10 :3 rather than just 10mp

After all is the 8mp Sony equals to the 8mp Canon D1 Mark2? That’s where the naming is wrong and throwing pixel numbers without reference is not accurate at all with either Bayer, layering or multishot. This industry needs to seriously think about a standard referencing for sensors and have those posted on packaging, enforcing it by law if necessary.
 
benoit said:
If the 3mp x3 is to be referred to as 10mp,and the x refers to the
number of layers like I understand it, the new naming makes it a
30mp (10x3). So it seem logical to refer that sensor to a 10 :3
rather than just 10mp
Click to expand...
What are you talking about?

The SD9 or SD10 is 10.2 MP (3.4 MP Red + 3.4 MP Green + 3.4 MP Blue)

The x530 is 4.5 MP (1.5 MP Red + 1.5 MP Green + 1.5 MP Blue)

The total count and the possibility of spelling out the different color counts is perfectly compatible with industry standards and practice.

Why is this so hard to understand?

j
benoit said:
After all is the 8mp Sony equals to the 8mp Canon D1 Mark2? That’s
where the naming is wrong and throwing pixel numbers without
reference is not accurate at all with either Bayer, layering or
multishot. This industry needs to seriously think about a standard
referencing for sensors and have those posted on packaging,
enforcing it by law if necessary.
Click to expand...
 
As you know some digital backs use not 3 but 4 shots or 4 colors. If Foveon was to ad a secondary green would that make the 3 layers so called 10mp sensor a 13.6mp 4 layers sensor? Would it really be 25%better?

This said I’ve shot with multi exposure backs that would go to 12 and 18mp when combining all channels (the capture). With the advantage of using the same exact software and work flow there is no way I would compare a combined 3 shots of 12mp with an 11mp single shot back.

Give people a chance to compare cameras with accurate complete and detailed info not just MP count. Education can’t hurt. From there it’s their problem to look at the details and compare or to let the guy at Best Buy (who does not even know what pixel pitch is)make a n easy sale on his MP speach Every time Foveon is in the news this forum picks up on the same debate,. That makes it obvious that the industry standard is not that clear even to people with a certain awareness such as readers of this forum. I read what I see, sorry . Compatibility is not the issue here it’s more about readability.

In a way I find profiling and pixel count are understood very little.

If you believe I think little about the Foveon concept, it would be a serious mistake. Actually, I believe most people who, like me can’t make sense at calling a 3.4mp x3 a 10mp camera have a very high interest in that technology. Yes it has 10mp, no, it does not compare with a 10mp Bayer.

Noticed recently when Phil was stating his desapointment in the never ending MP war? Foveon is not sending the right message but they have a good product.
 
I always new you are a Foveon fan at heart Phil. I bet you have a 9 and 10 in your closet and carry it around as your preferred camera with tape over the identifiers!!!!

Rick
Phil Askey said:
Foveon's additional photodiodes constitute towards a higher
resolution image, they deliver an image with more resolution.

The Fujifilm additional photodiode DOES NOT.

The Foveon sensor can distinguish the color of a single pixel on
its array without being influenced by a neighbouring pixel, a Bayer
sensor can not.
PalmsWestPhoto said:
Phil

so you agree that the new S3 is a 12MPixel camera ?

As for stretching the truth i wouldnt put it like that. When these
ratings where made nobody thought that one day there might be a
small company with a sensor that doesnt fit the description. Once
used for years it couldnt be easily changed.

--
Michael Salzlechner
http://www.PalmsWestPhoto.com
Click to expand...
--
Phil Askey
Editor / Owner, dpreview.com
Click to expand...
--
It's a tough job, living in Hawaii, but someone has to do it!!!

Powered By Sigma..........Empowered By FOVEON..........

http://www.lightreflection.com
http://www.silveroaksranch.com
http://www.pbase.com/rickdecker
http://www.pbase.com/sigmasd9/user_home
 
Here is that same image when the SD9 is sized UP to match the 1DS instead of resizing the 1DS image down. Look at the artifacting in the SD10 image... pretty bad aliazing, and the resolution is pretty bad. I have found that many people misrepresent the SD9 by resizing the competing camera down to 3mp which throws away data, instead of sizing up the SD9/10 image as one would do for printing.



and here is a comparison between 6mp cameras upsizing the SD9 and all other cameras to the 12mp output size of the S2.

http://www.pbase.com/image/25960316/original

There is more evidence all around that the SD9/10 are not higher res than any other 6mp sensor DSLR.

and here is a res charts from Phil himself:



regards,
Sean
 
Pete Perry said:
Scott Greiff said:
http://www.dpreview.com/reviews/sigmasd9/page20.asp

Even in a comparison with a Canon 1Ds, the SD-10 holds its own:



While the Foveon may not make show 10M pixels in black and white
resolution, the current crop of 6M (or 11M) pixel bayer sensor
cameras can't claim to offer 6M pixels in color resolution.
  • ScottAG
Pete Perry said:
This Foveon will never achieve 10 Million Pixels in black and white
or color so with everything being considered equal I would say the
Foveon is the one not being honest.
Scott Greiff said:
Pete Perry said:
Standard Sensors are called 6 MP because they're monochromatic so
it captures everything in Black and White... This basically means
that the base image captured is 6 MP so it isn't really deceitful!
Click to expand...
It wouldn't be considered deceitful if the world was monochromatic.
  • ScottAG
--
http://www.greiff.org/scottag
http://www.pbase.com/scottag
Click to expand...
--
'The only real currency in this bankrupt world is what we share
with each other when we're being uncool.' -- Cameron Crowe
Click to expand...
--
http://www.greiff.org/scottag
http://www.pbase.com/scottag
Click to expand...
--
'The only real currency in this bankrupt world is what we share
with each other when we're being uncool.' -- Cameron Crowe
Click to expand...
 
Scott Greiff said:
Pete Perry said:
I don't doubt that it has good resolution... In fact if it weren't
for the other flaws of the original I would probably already own
one! However, the point is claiming 10 MP is wrong because it
isn't really a 10 MP file.
Click to expand...
And the Bayer crowd claiming their resolution figures even-though
the fake it, is any less wrong?

Sigma should have provided a 10MP upsampled option in camera
similar to what Fuji does. You guys would have been happy then.
Click to expand...
At least Fuji gives some measureable benefit to the resulting file... if Sigma had made the SD9/10 output a 10.2mp file people would be outraged as they would be horrible! Sigma/foveon are outright lying claiming 4.5 and 10.2 mp!
 
Kendall Helmstetter Gelner said:
PHGN said:
How many people here disagree with the labeling of Fovean resolution?

The case against:
Click to expand...
Consider these in the case for:

1) RAW file size. If I take a raw image from the new polaroid
camera, and a raw image from a 4MP camera, the raw image sizes will
be the same (actually the Polaroid would be larger). For a
concrete example, an SD10 RAW file is around 10MB and a 10D RAW
file is around 6MB. You certainly would like to know about how
much storage you might need for a new camera, and shouldn't the
rating help decide that?
Click to expand...
Well the S3 raw file is HUGE. Can only fit 20 of them on a 512mb card. So that is not that accurate either.
Kendall Helmstetter Gelner said:
2) Output. If the output quality for prints and web use is
identical between the Polaroid 4.5MP camera and a 4-5MP camera,
than shouldn't the rating be the same? Should not the MP rating
relate to the maximum potential quality of the image produced, and
nothing else?
Click to expand...
Sounds good to me. Then the camera can stand on its own against other 4.5 mp output cameras. Foveon/sigma should have done this in the beginning. However, if the Sigmas output a 10.2 mp image file, then people would REALLY be dissapointed when they saw a 1DS output file! So they would do poorly in this way as well. But if they want to call their camera a 10.2 mp camera, they better output a 10.2mp file and compete at that level, the sigmas fail at that level, so it is no wonder they do not do it that way.
Kendall Helmstetter Gelner said:
How is a user being cheated if the output is of the same quallty
between two similarily rated cameras?
Click to expand...
Because they are not. There is no way in the world a 10.2mp output file can match the 10mp output file from a 1ds. SO the user is really getting cheated.
See here:



SOrry, but if I got results like this from a 10.2mp camera, I would be furious!

regards,
Sean
 
Yep and actual images back that up:





Seems the res charts are not that far off.

Regards,
Sean
Peter G said:
Hmmm, that looks like the 3.4MP x3 is barely equivalent to a 6MP
bayer.

I know you want to use a Red on Blue res chart right?
Brett DiMichele said:
Why does everyone confuse Resolution with Megapixels?

A Megapixel is an output dimension and nothing more than that.
Just because an image may be 3.45 Megapixels "Dimensionally" does not
mean it can not carry 10.2 Photosites worth of collected data which
makes
it easy to interpolate up to a double size without degredation.

If Foveon is wrong for the way they label thier sensors then all of
the
manufacturers using the Bayer Sensor Topology are likewise guilty of
perpetrating consumer fraud because thier sensors do not even capture
true full color data on the entire surface of the Sensor.

What good does a "Megapixel" do you when a camera with less "MP" has
equal output to sensors that boast twice the "MP" ?

And please quit crying "Zealot" just because you don't have a firm
grasp
on technology.. Because you are a Bayer Zealot when you act that way!

--



http://www.westol.com/~brettd/sd10/
Click to expand...
--
http://www.trytel.com/~pguidry/vacation.html
Click to expand...
 
Just Looking said:
Peter,

You poke fun at people who would use a color target to show the
advantages of X3, but you choose a black-and-white target to show
the advantages of Bayer. Can you see your bias and double-standard
there?
Click to expand...
Nope, real world images bear the res charts out. Like I have asked you before Just looking, yet you always skirt the issue.... just how often is pure red next to pure blue at a high frequency pattern in nature? Almost never. In fact if there is any green in the colors at all (most of the time there is), then the bayer sensor does pretty darn well.
Just Looking said:
The Sigma users have found an interesting thing: the SD9 and SD10
make much better black-and-white photos than 6 MP Bayer cameras do.
Surprising, perhaps? Contradictory to your res chart data perhaps?
Not really. The real world has a lot of color detail, and the
Bayer sensors, because they under-sample color, alias that detail
into false luminance detail, which gives the image a typical
Bayer-mushy look. With X3, you get better luminance detail, on
real-world scenes.
Click to expand...
Again you are all words, yet here we are continuously giving actual images as examples and you claim they are just hogwash because her highness said so. Well sorry, but the only area that the Foveon is clearly superior to 6mp bayer sensors is in red/blue res charts, which are much more unrealistic in real life than black and white lines. After all white is made up of all colors, so it should also be a disadvantage to bayer, right? I would say that black and white shades are much more common in nature at high frequencies than pure red/blue. Yet here is what the SD9 sees for black and white charts notice the bad aliasing as well on the SD9 sample around the numbers:


Just Looking said:
Do I have a standardized chart to prove this? No. Go take some
pictures. Or take those shots of the boat stacks and convert to
black-and-white (ideally, go get the raw files and do it right) and
see how that blue swoosh on red stack looks when compared in
black-and-white. Lots of other such detail, too...
Click to expand...
You never have any images to prove anything. I guess we just have to trust your flawed logic. Remember you are the one that claimed that downsizing does not get rid of noise, and sizing up does not affect noise negatively, that aliasing doe not get sampled out when downsizing or get more pronounced with upsizing, yet I proved all of that with samples, and you still refuse to admit it. Where is your credibility? Peter's samples above show it too. Looks like bad moire patterns on teh circular patterns above for the Sigma, no? I guess everyone viewing that is just imagining it because your logic refutes it.

Regards,
Sean
 
Yep, that is why I call the SD9/10 to be "Comparable" to a 6mp bayer. Foveon needs to stop being "comparable" and WOW the world with a 6mp or 8mp sensor... I would buy one!!!

Regards,
Sean
Peter G said:
Erik Magnuson said:
The problem is that neither 3.4 nor 10.3 is really a satisfactory
answer. And the answer that is (i.e, somewhere around 6 - 7) has
mainly subjective observation to support it.
Click to expand...
I really think this statement should be supportable by those who
own either camera. ( I own neither ).

I label anyone who pushes either end number ( 3.4 or or 10.2 ) as
an extremist ( Fanboy, zealot, or whatever ). All the real world
comparisons thus far show it to be in the 6-7MP equiv range.
Click to expand...
 
benoit said:
As you know some digital backs use not 3 but 4 shots or 4 colors.
If Foveon was to ad a secondary green would that make the 3 layers
so called 10mp sensor a 13.6mp 4 layers sensor? Would it really be
25%better?
Click to expand...
Probably not 25% better; just 25% more pixels.

Pixels is a count, not a goodness.
benoit said:
This said I’ve shot with multi exposure backs that would go to 12
and 18mp when combining all channels (the capture). With the
advantage of using the same exact software and work flow there is
no way I would compare a combined 3 shots of 12mp with an 11mp
single shot back.
Click to expand...
Not sure what you mean. Sounds to me like you are comparing them, and noting their differences. Makes sense to do that, if you interpret "compare" in the broad sense, or substitute "compare and contrast".
benoit said:
Give people a chance to compare cameras with accurate complete and
detailed info not just MP count. Education can’t hurt.
Click to expand...
Absolutely!
benoit said:
... Yes it has 10mp, no, it does not compare with a 10mp Bayer.
Click to expand...
I think you mean it is not "equivalent to" a 10mp Bayer. Still, it can and should be compared. Look at all those comparisons with 1Ds that people are throwing around. Are you saying they shouldn't do that?

j
 
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