If this is not false advertisement, what is?

[fjohn]

So, you decided the obvious thing to do to a thread that is already
too long and drifting too much was to throw in your own off topic
contribution.

Your comment is definitely a case of the pot calling the kettle black.

I guess you were in a better mood when you posted your "Badgers???" comment. (Which I thought was funny BTW.)

--
[fjohn]

 

[Marco van den Hout]

Actual Foveon quote:

"The Foveon X3 PRO 10M direct image sensor has a total of 10.2
million red, green, and blue pixels that are organized into three
layers (2268 x 1512 x 3 layers)"

Well, their 'pixels' (if they say 10.2) aren't 'red, green and blue' but red, green, OR blue.

You may call the 10D's 6 megapixel claim deceptive since in each of those pixels either r, g or b are true color values. But Sigma claiming 10.2 megapixels is even more ridiculous in my eyes. SD9/10 is 3.4 megapixel in color and D10 is 6.3 megapixel in color, but there's different techniques used in how to get to those pixels and that should be the message.
But I won't buy a camera based solely on how many megapixels it produces...

 

[Alatar]

Alatar, thanks for taking the bait.
Since we now agree that "Basing quality on output attributes is
ludicrous", how do you justify your previous statement that "I
still claim that the Foveon sensor has 3.4 mega pixels;"?

Simple. Because the sensor measures 3.4 million incoming rays of light. It just measures them properly (for human vision) by doing it at three different wavelengths, instead of only partially by measuring a single wavelength. That's why I also said that the bayer sensors measure "one-third pixels", which may be a bit perjorative, but I can't think of a better way to describe it at the moment.

Note that measuring 3.4 million incoming rays of light is only enough to make it a 3.4 megapixel sensor . In order to be a 3.4 megapixel camera, the rest of the system has to retain all the measured information all the way through it's processing (which the SD9/10 does, at least in high-res mode).

 

[Just Looking]

But I won't buy a camera based solely on how many megapixels it produces...

Of course not. Fuji S2 produces 12M and Sigma SD9/10 produces 3.4M.

A much more reasonable number to compare would be how many "somethings" (as Joe puts it) they measure, not how many they "produce": Fuji 6 Million, Sigma 10 Million.

I wouldn't buy a camera based on that number, either, but at least it's useful data that compares SuperCCD and X3 in the same terms as normal Canon and Nikon sensors.

j

 

[Just Looking]

Yes, Alatar, good point.

So the Sigma is 3.4 MP just as soon as the industry accepts your change of definition to knock everyone else's numbers down by a factor of three. Good luck with this campaign to change to the world.

j

Alatar, thanks for taking the bait.
Since we now agree that "Basing quality on output attributes is
ludicrous", how do you justify your previous statement that "I
still claim that the Foveon sensor has 3.4 mega pixels;"?

Simple. Because the sensor measures 3.4 million incoming rays of
light. It just measures them properly (for human vision) by doing
it at three different wavelengths, instead of only partially by
measuring a single wavelength. That's why I also said that the
bayer sensors measure "one-third pixels", which may be a bit
perjorative, but I can't think of a better way to describe it at
the moment.

Note that measuring 3.4 million incoming rays of light is only
enough to make it a 3.4 megapixel sensor . In order to be a 3.4
megapixel camera, the rest of the system has to retain all the
measured information all the way through it's processing (which the
SD9/10 does, at least in high-res mode).

 

[Alatar]

I don't know if you are old enough to remember the days when the stereo manufacturers started a similar marketing game talking about output power. There was RMS power, True power, Instantaneous Peak-to-Peak Power and all kinds of nonsense. So much so that everyone stopped paying attention to the claims, and started paying attention to quality. The same thing happened with low-end scanners a few years back.

All these guys are doing is contributing to the general pollution of meaningful information -- although I do understand why Sigma has to use the 10.2 figure. I predict that it will eventually all become meaningless and we will start to see stuff like "peak enhanced equivalent resolution".

At that point maybe we can all go back to looking at quality again instead of marketing hype.

So the Sigma is 3.4 MP just as soon as the industry accepts your
change of definition to knock everyone else's numbers down by a
factor of three. Good luck with this campaign to change to the
world.

j

Alatar, thanks for taking the bait.
Since we now agree that "Basing quality on output attributes is
ludicrous", how do you justify your previous statement that "I
still claim that the Foveon sensor has 3.4 mega pixels;"?

Simple. Because the sensor measures 3.4 million incoming rays of
light. It just measures them properly (for human vision) by doing
it at three different wavelengths, instead of only partially by
measuring a single wavelength. That's why I also said that the
bayer sensors measure "one-third pixels", which may be a bit
perjorative, but I can't think of a better way to describe it at
the moment.

Note that measuring 3.4 million incoming rays of light is only
enough to make it a 3.4 megapixel sensor . In order to be a 3.4
megapixel camera, the rest of the system has to retain all the
measured information all the way through it's processing (which the
SD9/10 does, at least in high-res mode).

 

[Phil Askey]

Yes, and I've listed this in our camera database for over a year now, I call it "sensor photo detectors". Compare:

http://www.dpreview.com/reviews/specs/Fujifilm/

http://www.dpreview.com/reviews/specs/Sigma/

http://www.dpreview.com/reviews/specs/Canon/

Of course not. Fuji S2 produces 12M and Sigma SD9/10 produces 3.4M.
A much more reasonable number to compare would be how many
"somethings" (as Joe puts it) they measure, not how many they
"produce": Fuji 6 Million, Sigma 10 Million.

I wouldn't buy a camera based on that number, either, but at least
it's useful data that compares SuperCCD and X3 in the same terms as
normal Canon and Nikon sensors.

j

--
Phil Askey
Editor / Owner, dpreview.com

 

[Roland Karlsson]

The camera then has the possibility to detect color
information at each pixel. This increases the color depth,
but not the resolution.

This not only increases the color depth - it also increases the
sharpness.

Please tell me how you from the SD9 you can get
a resolution that is higher than 3.5 Mpixel.

Roland

 

[Andy61557]

I just was a cold, rainy Monday morning. I was hating my work, and the fall, and the 'daylight time saving', and the coming winter. And I see this 'great news', 10MP Fofeon sensor. I say, WOW, this is interesting, this must be great. Foveon is ultra-sharp, even if Canon has no plans for it. I read on and... what a bummer! You know the rest. I just lost it. Sorry about that.

But I'm glad you had so much fun! Smell you all later, when I dump my Canon for Sigma. And don't get too relaxed, it's still raining in New York...

 

[Just Looking]

I don't know if you are old enough to remember the days when the
stereo manufacturers started a similar marketing game talking about
output power. There was RMS power, True power, Instantaneous
Peak-to-Peak Power and all kinds of nonsense. So much so that
everyone stopped paying attention to the claims, and started paying
attention to quality. The same thing happened with low-end scanners
a few years back.

All these guys are doing is contributing to the general pollution
of meaningful information -- although I do understand why Sigma has
to use the 10.2 figure. I predict that it will eventually all
become meaningless and we will start to see stuff like "peak
enhanced equivalent resolution".

At that point maybe we can all go back to looking at quality again
instead of marketing hype.

Agreed. That's why I like to help fuel these arguments. It should eventually make it clear that "megapixels" is about as meaningless as "brake horsepower".

j

 

[Alatar]

I just was a cold, rainy Monday morning. I was hating my work, and
the fall, and the 'daylight time saving', and the coming winter.
And I see this 'great news', 10MP Fofeon sensor. I say, WOW, this
is interesting, this must be great. Foveon is ultra-sharp, even if
Canon has no plans for it. I read on and... what a bummer! You know
the rest. I just lost it. Sorry about that.
But I'm glad you had so much fun! Smell you all later, when I dump
my Canon for Sigma. And don't get too relaxed, it's still raining
in New York...

As one of the posts above said -- or maybe it was in the Canon thread you started -- it only counts as a master troll if you reach 150 posts.

Whoops, my mistake: you've posted thrice in the thread (you're only allowed twice), so you've lost your bid for mastership already. ;-)

But congrats anyway and thanks -- it was fun....

 

[Mike Fried]

Mike, good analysis of different things that one might count as
pixels. Your introduction of the term "Sensor Sites" is
interesting, but differs from what Phil called a "photosite" when
he said "the X3 sensor used in the SD9 would be called '10.3
million photosites'."
http://www.dpreview.com/news/0203/02030602foveonx3notation.asp

I think I first saw the term "Sensor Site" used in Fuji marketing for their SR cameras. In my mind, Photo Site = Sensor Site. The difference between a site and a sensor is the difference between a physical location (region in space) and an object which sits there. The idea here is that more than one sensor can be in the same physical location, and with the anti-aliasing filter and microlenses on most CFA digital cameras, this is close to true (some portion of light from the same region goes to each of the 4 neighboring sensors). For the SR sensor, one sensor is inside the other. For the Foveon, 3 sensors are stacked.

By the way, have you noticed that Foveon and Sigma have over a year
of experience trying to teach people the subtle distinctions about
all this? It seems that maybe they're just about giving up and
going with the flow in saying "10.2 MP" because it's the only way
to keep people from trivializing their message to "3.4 MP".

It's hard to explain without the pictures of the scattered red, green, and blue channels.

Pathetic.

It's unfortunate that marketers are not as technically savvy as engineers. But these are not easy concepts to get. We're comparing apples and oranges, and the industry has set aside a single number to summarize it: Pixel Count.

Sensor/Pixel count on a Camera is like the clock rate on a CPU. For people who didn't get that terminology, I just said: MP = MHz/GHz. As technology advances, it becomes less and less relevent. Eventually, everyone will have a 30MP camera with a 30GHz CPU, 30GB RAM, and 30TB HD/DVD and a built-in 30Gb/s wireless network. Eventually, it will cost $30. Ultimately, batteries will last 30 Hours per 1 hour charge, and the issues will have more to do with content than the limitations of technology.

Why buy now? Because we're not there yet! And we won't ever get there unless people invest the money to take us there. We need to invest that money by buying technologies that provide us value. How do we know what products to buy? Well, Phil reviews them, and people buy them, and they share their feedback in the forums. It's a pretty simple equation.

You may think it's pathetic, but most people would probably skip reading the details of my post for not being able to understand what it is that I was talking about. They might be great photographers, but this technology is more than just photography. So, if Sigma/Foveon want to market their number 10.5/3.5 next to Canon's 11/11 or Fuji's 6/12, they will have to go the next round of explaining what they mean by that number.
-Mike

 

[Brett DiMichele]

Someday with much practice he will move up to Jedi Troll and then in many
years Master Jedi Troll, Yes, Yes!

I sense much anger and hate in him.. Hate leads to Anger. Anger leads to
the Dark Side!

Hahaha

 

[Joseph S Wisniewski]

As one of the posts above said -- or maybe it was in the Canon
thread you started -- it only counts as a master troll if you reach
150 posts.

Whoops, my mistake: you've posted thrice in the thread (you're only
allowed twice), so you've lost your bid for mastership already. ;-)

I don't think we should count that. A troll posts to his own thread to keep it going, to fan the flames. Andy's comment seems more like a deliberate attempt to disqualify himself.

If he doesn't want to be branded a "master troll", it would seem to be even more fun to do so.

--
Ciao!

Joe

http://www.swissarmyfork.com

 

[Brian Slater]

As one of the posts above said -- or maybe it was in the Canon
thread you started -- it only counts as a master troll if you reach
150 posts.

Whoops, my mistake: you've posted thrice in the thread (you're only
allowed twice), so you've lost your bid for mastership already. ;-)

I don't think we should count that. A troll posts to his own thread
to keep it going, to fan the flames. Andy's comment seems more like
a deliberate attempt to disqualify himself.

If he doesn't want to be branded a "master troll", it would seem to
be even more fun to do so.

--
Ciao!

Joe

I think we need a new category. As well as Master Trolls, we have those that spur them on. We could call them Master Baiters?

 

[Joseph S Wisniewski]

I don't know if you are old enough to remember the days when the
stereo manufacturers started a similar marketing game talking about
output power. There was RMS power, True power, Instantaneous
Peak-to-Peak Power and all kinds of nonsense. So much so that
everyone stopped paying attention to the claims, and started paying
attention to quality.

As an aside, those days are back, if you try to shop for car stereos. I've seen stereo systems claiming 1600 watts. That's the peak power of a 10% pulse train. You can't systain 1600 watts in a car, the alternator is typically only 1200 watts. (and most of that goes into running the lighting, climate control blowers, engine control systems, etc).

The same thing happened with low-end scanners
a few years back.

All these guys are doing is contributing to the general pollution
of meaningful information -- although I do understand why Sigma has
to use the 10.2 figure. I predict that it will eventually all
become meaningless and we will start to see stuff like "peak
enhanced equivalent resolution".

At that point maybe we can all go back to looking at quality again
instead of marketing hype.

Agreed. That's why I like to help fuel these arguments. It should
eventually make it clear that "megapixels" is about as meaningless
as "brake horsepower".

While looking for information on "back CCDs", I ran into literally thousands of papers online that attempt ot analyze the MTF functions of sensors, just like they do with film and lenses. It's a little difficult, there are excellent papers discussing techniques to compensate for aliasing, etc. But it is doable. Of course, it took years for someone to round up enough stereo manufacturers to form the IHF, I don't know how we're going to get camera manufacturers into a rating program.

--
Ciao!

Joe

http://www.swissarmyfork.com

 

[Joseph S Wisniewski]

I think we need a new category. As well as Master Trolls, we have
those that spur them on. We could call them Master Baiters?

Where are those computer destroying solar flares, when you really need them?

Where, oh where?



--
Ciao!

Joe

http://www.swissarmyfork.com

 

[Joseph S Wisniewski]

Well, OK, some interpretation: that page is about as it says in
the subtitle, "A discussion of the ambiguous meanings of pixel and
megapixel," not a justification for a uniquely correct answer. If
I can ever get a clear definition of "pixel" from you, I'd like to
compare and contrast it to these, and see what shakes out.

Hey j, I think I've finally got the clear definition of "pixel". And it doesn't even involve buyng a $220 book (or getting my copy back from the guy who borrowed it).

It's from RLG, the "Research Libraries Group", and they seem to be a pretty authorative outfit, founded by Columbia, Harvard, and Yale Universities and The New York Public Library and claiming a membership of "over 160 universities, national libraries, archives, historical societies, and other institutions".

http://www.rlg.org/rlg.html

This is part of their series of "Guides to Quality in Visual Resource Imaging". Those are frighteningly well researched, quote ISO standards, technical papers, you name it.

http://www.rlg.org/visguides/visguide3.html

I've quoted the entire section 2 "Basic Terminology". It pretty clearly defines a "pixel" as a spatial entity, and color as an attribute of a pixel.

First, three sentences, out of context

"The pixel dimensions of the image are its width and height in pixels. The density of the pixels, i.e., the number of pixels per unit length on the document, is the spatial sampling frequency, and it may differ for each axis".

"The value of each pixel represents the brightness or color of the original object..."

Then the entire "Basic Terminology" section, with those three sentences in context.

--- begin quote ---

2.0 Some Basic Terminology

Digital images are composed of discrete picture elements, or pixels, that are usually arranged in a rectangular matrix or array. Each pixel represents a sample of the intensity of light reflected or transmitted by a small region of the original object. The location of each pixel is described by a rectangular coordinate system in which the origin is normally chosen as the upper left corner of the array and the pixels are numbered left-to-right and top-to-bottom, with the upper left pixel numbered (0,0).

It is convenient to think of each pixel as being rectangular and as representing an average value of the original object's reflected or transmitted light intensity within that rectangle. In actuality, the sensors in most digital image capture devices do not "see" small rectangular regions of an object, but rather convert light from overlapping nonrectangular regions to create an output image.

A document or another object is converted into a digital image through a periodic sampling process. The pixel dimensions of the image are its width and height in pixels. The density of the pixels, i.e., the number of pixels per unit length on the document, is the spatial sampling frequency, and it may differ for each axis.

The value of each pixel represents the brightness or color of the original object, and the number of values that a pixel may assume is the number of quantization levels. If the illumination is uniform, the values of the pixels in a gray-scale image correspond to reflectance or transmittance values of the original. The values of the pixels in a color image correspond to the relative values of reflectance or transmittance in differing regions of the spectrum, normally in the red, green, and blue regions. A gray-scale image may be thought of as occupying a single plane, while a color image may be thought of as occupying three or more parallel planes.

A bitonal image is an image with a single bit devoted to each pixel and, therefore, only two levels—black and white. A gray-scale image may be converted to a bitonal image through a thresholding process in which all gray levels at or below a threshold value are converted to 0-black—and all levels above the threshold are converted to 1-white. The threshold value may be chosen to be uniform throughout the image, that is, "global thresholding" or it may be regionally adapted based on local features, or "adaptive thresholding." Although many high-contrast documents may be converted to bitonal image form and remain useful for general reading, most other objects of value to historians and researchers should probably not be. Too much information is lost during thresholding.

These concepts are illustrated in fig.1, which contains a gray-scale image of the printed word "Gray" and a color image of the printed word "Color." A portion of the gray-scale image has been enlarged to display the individual pixels. A bitonal image has been created from the enlarged section to illustrate the consequent loss of information caused by thresholding. As may be observed, the darker "r" remains recognizable, but the lighter "a" is rendered poorly. The color image is also displayed as three gray-scale images in its red, green, and blue planes. Note that the pixels corresponding to the color red are lighter in the red plane image and similarly for the green and blue plane images.

Figure 1. A gray-scale image with a portion enlarged, a bitonal image of the same portion, and a color image with its three color planes shown separately.

--- end quote ---

--
Ciao!

Joe

http://www.swissarmyfork.com

 

[Aal]

I don't have neither 10d nor sd 9.
However I did some print tests with the samples.

I donwnloaded dpreview's samples from all dslr's except 1ds. I enlarged them all to a2 size 240 dpi with the help of interpolation software, all using the same parameters, and used the descent epson large format printer to print them.
From the results, I believe foveon is 10 mega pixels.

From my observation. Foveon successfully pulled the detail out from random textures (i.e. foliage or grass in a far distance), from where bayer has to guess, and covers the weakness with antialiassing.

And from sd10, the color is much better. I guess I need to save some more money for SD10.

----
Aal

 

[Alatar]

Well, he didn't really post thrice.....

I was just trying to get him to post again in protest, at which point he really would have posted thrice.

Didn't work though.... :-(

Guess he's smarter than you'ld guess from his original post ;-)

As one of the posts above said -- or maybe it was in the Canon
thread you started -- it only counts as a master troll if you reach
150 posts.

Whoops, my mistake: you've posted thrice in the thread (you're only
allowed twice), so you've lost your bid for mastership already. ;-)

I don't think we should count that. A troll posts to his own thread
to keep it going, to fan the flames. Andy's comment seems more like
a deliberate attempt to disqualify himself.

If he doesn't want to be branded a "master troll", it would seem to
be even more fun to do so.

--
Ciao!

Joe

http://www.swissarmyfork.com