602 questions

[Bob Israelian]

I am just starting to investigate the Fuji 602 and have a few questions that i hope someone here can help me with. I have been looking at Nikon 5000, Dimage 7 and Olympus E 20 cameras for features and 5 megapixel capabilities. I have seen or read that the Fuji systems use a special kind of chip or algorythim or something that doubles the megapixels so that the 6 MP is really about 3 but looks and acts like a 6 mp? Is this true and if so is it for real or more wishful thinking. I know it has been around for awhile and hope it is as good as they say it is. Any explanation that is not too technical is appreciated.
Do photos really compare?

If it is truly performing like 6 megapixels then the next question: How much is it in US dollars.

It looks great in photos and I like the design. It seems loaded with features like the other top end systems.--Bob

 

[Jim Gilliland]

I am just starting to investigate the Fuji 602 and have a few
questions that i hope someone here can help me with. I have been
looking at Nikon 5000, Dimage 7 and Olympus E 20 cameras for
features and 5 megapixel capabilities. I have seen or read that the
Fuji systems use a special kind of chip or algorythim or something
that doubles the megapixels so that the 6 MP is really about 3 but
looks and acts like a 6 mp? Is this true and if so is it for real
or more wishful thinking. I know it has been around for awhile and
hope it is as good as they say it is. Any explanation that is not
too technical is appreciated.
Do photos really compare?
If it is truly performing like 6 megapixels then the next question:
How much is it in US dollars.
It looks great in photos and I like the design. It seems loaded
with features like the other top end systems.

OK, here we go again

Let me first suggest that you read the many other threads on the topics that you raise here. All of these points have been addressed in great detail in recent threads in this very forum. There are a couple of lengthy threads on resolution and on the SuperCCD. And there are a couple on S602 pricing.

So I'll give you some answers.

The S602 is a 3 megapixel camera. It stores its output in a 6mp file, but its native resolution is 3mp. Don't expect more than that. Don't buy into the marketing hype. Fuji engineers have done a nice job, but they're still using a 3mp CCD.

Equally important - don't let resolution become your most important criterion for selecting a camera. There are other factors that are more important (Geez, didn't a just write this a minute ago in another thread?)

The US price hasn't yet been announced, but it will almost certainly come in under $1000 list.

I agree with you that it appears to offer a winning combination of performance and features. Keep in mind that it won't be available until June.

 

[Aaron_6900]

OK, here we go again

Let me first suggest that you read the many other threads on the
topics that you raise here. All of these points have been
addressed in great detail in recent threads in this very forum.
There are a couple of lengthy threads on resolution and on the
SuperCCD. And there are a couple on S602 pricing.

So I'll give you some answers.

The S602 is a 3 megapixel camera. It stores its output in a 6mp
file, but its native resolution is 3mp. Don't expect more than
that. Don't buy into the marketing hype. Fuji engineers have done
a nice job, but they're still using a 3mp CCD.

INCORRECT

It is a 3 mp camera, but it's CCD is like no other on the market. The sensors are octogonal shaped rather than the traditional square shape of other manufacturers. This makes for practically no space in-between the sensors (due to octagon shape). This is an engineering solution to a design problem/inadequacy inherint to the traditional CCD. You can increase the sensor count on a tiny CCD as much as you want, but if the method in which they are placed and the algorithm is flawed then that DOES NOT guarantee superior images. Fuji is not pumping smoke and mirrors. This is NOT like the interpollation tricks used in scanner design. The fact is that there design works better than the method used by other manufacturers and it often seems to bother other users that may have purchased cameras with hyped up pixel count and inferior lenses.......helps them justify the excess cash layed down for an inferior design. Bottom line is superCCD works and works very well. There is no denying it. In addition ALL CAMERAS use a form of manipulation in establishing an image made from a CCD's light capturing sensors. Fuji just does it better. They have won a number of design awards for SuperCCD technology........not from hyped marketing tatics as suggested by some. Use the 6900 and 602's 6mp setting for pictures you may need to make large prints from. Use 3 mp setting for up to 8x10 size and save storage space. I will put ANY one of the 6900's 6 mp shots up against the current crop of 4 and 5 mp cameras out there. It will match up to them shot for shot........in a compentent set of hands and fingers on the shutter. It will also BEAT them in color accuracy and vibrancy.

Equally important - don't let resolution become your most important
criterion for selecting a camera. There are other factors that are
more important (Geez, didn't a just write this a minute ago in
another thread?)

Agreed. There is more to a quality t picture than just CCD megapixel numbers. Lens optics and the method/algorithm used to produce the image means more.

The US price hasn't yet been announced, but it will almost
certainly come in under $1000 list.

I agree with you that it appears to offer a winning combination of
performance and features. Keep in mind that it won't be available
until June.

 

[Aaron_6900]

Also.

SuperCCD design truly allows it to capture more (lost) information than traditional square sensor designs due to that octagonal honeycomb arrangement. Traditional square designs lose light information in-between the sensors (space) because they are further apart. SuperCCD sensors are much closer together on the CCD and therefore capture more light/image. Not smoke and mirrors here.

Super CCD at 3.1 mp captures more information than traditional 3.1 mp(insert brandname here) CCD.

So if we were to use "x" as a component of CCD information in terms of a quantitative amount and + as a component of the amount of information lost by the space between the sensors' arrangment on the CCD chip then:

Traditional 3.1 mp CCD = x amount of information.

SuperCCD 3.1 mp CCD = x++. (More like 4-5 mp level in terms of information QUANTITY.

 

[Rosco70263]

See the following thread:

http://www.dpreview.com/forums/read.asp?forum=1012&message=2176106

I agree with both of the above opinions, if that is possible. The CCD is octagonal but it is a 3MP CCD. Interpolation can be done in camera as well as by using fractal software with similar outcome.

The bottom line is a 3MP camera is capable of producing images large enough for "most". Not all. It depends on the size print you want to product. If you'll never print larger than 8x10 anything more is overkill.

Conversely higher end cameras get more megapixels and more features. So your not just paying more for the MP CCD.

Hope this helps.

OK, here we go again

Let me first suggest that you read the many other threads on the
topics that you raise here. All of these points have been
addressed in great detail in recent threads in this very forum.
There are a couple of lengthy threads on resolution and on the
SuperCCD. And there are a couple on S602 pricing.

So I'll give you some answers.

The S602 is a 3 megapixel camera. It stores its output in a 6mp
file, but its native resolution is 3mp. Don't expect more than
that. Don't buy into the marketing hype. Fuji engineers have done
a nice job, but they're still using a 3mp CCD.

INCORRECT
It is a 3 mp camera, but it's CCD is like no other on the market.
The sensors are octogonal shaped rather than the traditional square
shape of other manufacturers. This makes for practically no space
in-between the sensors (due to octagon shape). This is an
engineering solution to a design problem/inadequacy inherint to the
traditional CCD. You can increase the sensor count on a tiny CCD as
much as you want, but if the method in which they are placed and
the algorithm is flawed then that DOES NOT guarantee superior
images. Fuji is not pumping smoke and mirrors. This is NOT like the
interpollation tricks used in scanner design. The fact is that
there design works better than the method used by other
manufacturers and it often seems to bother other users that may
have purchased cameras with hyped up pixel count and inferior
lenses.......helps them justify the excess cash layed down for an
inferior design. Bottom line is superCCD works and works very well.
There is no denying it. In addition ALL CAMERAS use a form of
manipulation in establishing an image made from a CCD's light
capturing sensors. Fuji just does it better. They have won a
number of design awards for SuperCCD technology........not from
hyped marketing tatics as suggested by some. Use the 6900 and 602's
6mp setting for pictures you may need to make large prints from.
Use 3 mp setting for up to 8x10 size and save storage space. I
will put ANY one of the 6900's 6 mp shots up against the current
crop of 4 and 5 mp cameras out there. It will match up to them shot
for shot........in a compentent set of hands and fingers on the
shutter. It will also BEAT them in color accuracy and vibrancy.

Equally important - don't let resolution become your most important
criterion for selecting a camera. There are other factors that are
more important (Geez, didn't a just write this a minute ago in
another thread?)

Agreed. There is more to a quality t picture than just CCD
megapixel numbers. Lens optics and the method/algorithm used to
produce the image means more.

The US price hasn't yet been announced, but it will almost
certainly come in under $1000 list.

I agree with you that it appears to offer a winning combination of
performance and features. Keep in mind that it won't be available
until June.

 

[ianR]

Before buying the 6900 I printed of dozens and dozens of images from the 5 cameras I was considering. With the 6900 there were some I thought were great, some I thought were ok and some I wasn't sure about. Then it dawned on me that I could tell which ones were the 6mp because the file size was different. So I checked and in nearly every case the ones I liked were the 6mp images.

This to me is finally all that matters: they look better on print. Curiously, I'm not sure I like them so much on the monitor.

Of all the pictures I looked at there was one I thought looked really bad. It's the one of St. Pauls Cathedral in Phil's gallery. Just look at all the shadows around the superstructure of the bridge. Just look at all the blocky dark colour.

I printed it off at about A4 (I just dragged it onto the page in Paint Shop Pro) just to get an idea of what a terrible looking image printed like. Try it. It looks amazing when printed!

Also, resolution as we call it, is comprised of a number of factors. A very important factor is contrast. Something I think you'll see with the 6900 is excellent contrast and tonal quality. The shadows, the dappled light,is beautifully defined, and this does seem to add to perceived resolution too.

All CCDs throw away a lot of resolution. Fuji claims that the CCD is about 130% more efficient than its rivals. This has been backed up mathematicaly by people like Moshe. Therefore, if this is true, that means at the very least it has the 'equivalent' resolution of a 4mp camera. Not actual resolution please note, but when compared against a number of 4mp cameras it seems to produce similar results. Some of the 5mp cameras too. The Sony generally seems to produce more resolution, but I'm not convinced at all the Dimage7 does. Also, the Sony seems to produce a flatter image.

In the final analysis, download some samples, print them off for yourself, and let your eyes be the judge.
regards
Ian

 

[ianR]

Yes Aaron, and while we are on the subject, it's a bigger sensor too. It captures more light than other 3mp cameras. It's just a little smaller than the Sony 5mp sensor but not much. So also the image doesn't have to be expanded as much as other CCds so you get a better image that way too.

All pixels are not created equal. The D30 has a 3mp sensor 12X larger than the average Sony 3mp CCD.
regards
Ian

Also.

SuperCCD design truly allows it to capture more (lost) information
than traditional square sensor designs due to that octagonal
honeycomb arrangement. Traditional square designs lose light
information in-between the sensors (space) because they are further
apart. SuperCCD sensors are much closer together on the CCD and
therefore capture more light/image. Not smoke and mirrors here.

Super CCD at 3.1 mp captures more information than traditional 3.1
mp(insert brandname here) CCD.

So if we were to use "x" as a component of CCD information in terms
of a quantitative amount and + as a component of the amount of
information lost by the space between the sensors' arrangment on
the CCD chip then:

Traditional 3.1 mp CCD = x amount of information.
SuperCCD 3.1 mp CCD = x++. (More like 4-5 mp level in terms of
information QUANTITY.

 

[Peter G]

It has not been mathematically backed up by anyone, please site the link?

Here is a link into the recent war over this that shows images from the reviews (on this site) from the Fuji 6mp output, and 3mp,4mp, and 5mp Sonys. Judge for yourself how good the 6mp output is.

http://www.dpreview.com/forums/read.asp?forum=1012&message=2160163

Here is a quote from the 6900 review:
http://www.dpreview.com/reviews/fuji6900z/page13.asp

Next question - does the 6900Z show more resolution in its native six megapixel mode than a typical 3 megapixel image interpolated in Photoshop? Not really, bear in mind that we've already seen that the 6900Z's lens is sharper than the 995 so some detail will 'stand out' a little more. That said the 6900Z isn't really able to render any finer detail than a good 3 megapixel digital camera. Its in-camera 'processing' is definetly very sophisticated and can generate very pleasing six megapixel images, however they still really only have 3 megapixels of detail.

The only honest statement is that the Fuji produces a better interpolated image than any other 3mp camera.

I am very interested in the 602 as it looks to be one of the best 3mp cameras on the market, but I wouldn't compare it to a true 5mp camera like the Sony 707, but I expect it to be priced lower.

Peter

Before buying the 6900 I printed of dozens and dozens of images
from the 5 cameras I was considering. With the 6900 there were some
I thought were great, some I thought were ok and some I wasn't sure
about. Then it dawned on me that I could tell which ones were the
6mp because the file size was different. So I checked and in nearly
every case the ones I liked were the 6mp images.
This to me is finally all that matters: they look better on print.
Curiously, I'm not sure I like them so much on the monitor.
Of all the pictures I looked at there was one I thought looked
really bad. It's the one of St. Pauls Cathedral in Phil's gallery.
Just look at all the shadows around the superstructure of the
bridge. Just look at all the blocky dark colour.
I printed it off at about A4 (I just dragged it onto the page in
Paint Shop Pro) just to get an idea of what a terrible looking
image printed like. Try it. It looks amazing when printed!
Also, resolution as we call it, is comprised of a number of
factors. A very important factor is contrast. Something I think
you'll see with the 6900 is excellent contrast and tonal quality.
The shadows, the dappled light,is beautifully defined, and this
does seem to add to perceived resolution too.
All CCDs throw away a lot of resolution. Fuji claims that the CCD
is about 130% more efficient than its rivals. This has been backed
up mathematicaly by people like Moshe. Therefore, if this is true,
that means at the very least it has the 'equivalent' resolution of
a 4mp camera. Not actual resolution please note, but when compared
against a number of 4mp cameras it seems to produce similar
results. Some of the 5mp cameras too. The Sony generally seems to
produce more resolution, but I'm not convinced at all the Dimage7
does. Also, the Sony seems to produce a flatter image.
In the final analysis, download some samples, print them off for
yourself, and let your eyes be the judge.
regards
Ian

 

[Aaron_6900]

Sites to condider reading more about how SuperCCD differs.

http://home.fujifilm.com/products/digital/pdf/sccd.pdf
(huge PDF file)

http://home.fujifilm.com/products/digital/sccd/review.html

http://www.fujifilm.co.uk/fujidc/ccd/
(view the interactive movie)

 

[Jim Gilliland]

OK, here we go again

Let me first suggest that you read the many other threads on the
topics that you raise here. All of these points have been
addressed in great detail in recent threads in this very forum.
There are a couple of lengthy threads on resolution and on the
SuperCCD. And there are a couple on S602 pricing.

So I'll give you some answers.

The S602 is a 3 megapixel camera. It stores its output in a 6mp
file, but its native resolution is 3mp. Don't expect more than
that. Don't buy into the marketing hype. Fuji engineers have done
a nice job, but they're still using a 3mp CCD.

INCORRECT
It is a 3 mp camera, but it's CCD is like no other on the market.
The sensors are octogonal shaped rather than the traditional square
shape of other manufacturers. This makes for practically no space
in-between the sensors (due to octagon shape). This is an
engineering solution to a design problem/inadequacy inherint to.....

Boy, you sure digested Fuji's marketing BS well!

I stand by what I said - the Fuji is a 3mp camera. For further discussion, join the lengthy thread that we've had going all this week on this topic. Make sure you read all the posts there already before jumping in.

 

[Jim Gilliland]

Ian (and all), go read the thread on the SuperCCD from earlier this week. Moshe and I and many others discussed this at length. I'm not going to rehash it all here except to say that it is mathematically impossible for a 3mp sensor to produce 4mp worth of information.

http://www.dpreview.com/forums/read.asp?forum=1012&message=2153499

Please, let's not restart all the BS now - go read the thread. We've already put this issue to bed.

 

[Jim Gilliland]

Sites to condider reading more about how SuperCCD differs.

http://home.fujifilm.com/products/digital/pdf/sccd.pdf
(huge PDF file)

http://home.fujifilm.com/products/digital/sccd/review.html

http://www.fujifilm.co.uk/fujidc/ccd/
(view the interactive movie)

Thanks, Aaron, I've read all of this material. Fuji has done some very interesting things with their design work - I give their engineers a lot of credit. Furthermore, their marketing writers have done some nice work ask well - making it sound like the Fuji designers are not only good engineers, but good magicians as well.

I'll give Fuji credit for great design work, but I won't give them credit for violating the laws of information science. A 3mp CCD cannot produce more than 3mp of information. Never, not under any circumstances.

If you read the Fuji marketing fluff carefully, though, it isn't too difficult to see what they HAVE accomplished.

1) They've designed the CCD so that each pixel is larger and can gather more light. That allows them to achieve better ISO ratings without increasing noise

2) They've used an interpolation algorithm that allows them to maximize the amount of information that is retained from the 3mp CCD.

3) They've used some additional clever algorithms to combine pixels from the CCD in such a way as to allow them to capture video and do some other neat tricks.

But none of that involves anything beyond 3 megapixels.

Go back and read the lengthy thread that we've already had on this topic. And please don't feel that anyone is "slamming" your camera by calling it a 3mp camera. It's simply the truth, and it's NOT a drawback.

 

[Mofongo]

Aaron_6900 wrote:
Bottom line is superCCD works and works very well.
There is no denying it. In addition ALL CAMERAS use a form of
manipulation in establishing an image made from a CCD's light
capturing sensors. Fuji just does it better. They have won a
number of design awards for SuperCCD technology........not from
hyped marketing tatics as suggested by some. Use the 6900 and 602's
6mp setting for pictures you may need to make large prints from.

I agree here. The only way to see for yourself is compare prints. Large prints are the bottom line for shooting in the 6mp mode. I don't care what they look like on the moniter. If the prints look better than a standard 3mp CCD camera than that's all that matters isn't it? Of course it is. People arn't going to look at a beautiful enlargement and question the reason why it looks so good. There not gonna care and neither do I that it was taken with a Super CCD that used non traditional methods.

There are many things in history that were accomplished even though naysayers said they were "mathematically impossible" The Great Pyramids comes to mind...

The end result is what counts, bottom line is large prints look better...everything else is just mathematical mumbo jumo. I'm glad Fuji is capable of thinking outside the box...

Bob--www.pbase.com/mofongo'The most beautiful sunsets are made by cloudy skies.'

 

[John Glover]

But none of that involves anything beyond 3 megapixels.

Go back and read the lengthy thread that we've already had on this
topic. And please don't feel that anyone is "slamming" your camera
by calling it a 3mp camera. It's simply the truth, and it's NOT a
drawback.

I wonder if what is confusing the issue here is what exactly is everyone talking about with regard to the information collected and output by the CCD sensor. You are absolutley correct in stating a 3MP CCD cannot gather more than 3MP of data. But, in the algorithm that process and output the 6MP file, isn't the final data file generated composed of 6MP of information, but about half of that is processed information....information not gathered by the CCD itself but generated by the camera's processing software. So, at least my take on it, is you have two issues here, the actual raw information gathered by the CCD, which is 3MP's worth, and the actual data generated by the camera in its output file which is 6MP's worth, a portion of which is interpolated data.

Everyone seems to get so defensive when it comes to this technolgy, but like so many other posters have stated, all CCD's interpolate some degree....in fact the Nikon D1x does it in one dimension (vertical I believe) to up the pixel count on it's data files. It is not a bad thing and when done correctly, it works beautifully. The bottom line is, if you are happy with you prints and your images, then enjoy them! Don;t worry about the technology and what others think of it....if it works for you, then so be it!

And have fun with you 6900.

--John http://www.pbase.com/jglover

 

[ianR]

Yes Jim. But What is 3mp of information worth? Is the 3mp from a Canon D30 the same as the 3mp from an average sensor when its sensor is 12x the size?

If the SuperCCD is doing a better job than the 4mp brigade and it's giving relatively equal images because it's more efficient, then why not accept, as many reviewers do that the SuperCCD gives the 'equivalent' resolution of about a 4.5mp camera. We're not claiming that it gives a true 6mp equivalent here.

Perhaps we should be argueing why some other cameras seem to give less than their true resolution if that's a more acceptable way of putting it.

One other thing, all this is going on in camera before JPEG compression, so there's a tiny gain in quality there.
regards
Ian

Ian (and all), go read the thread on the SuperCCD from earlier this
week. Moshe and I and many others discussed this at length. I'm
not going to rehash it all here except to say that it is
mathematically impossible for a 3mp sensor to produce 4mp worth of
information.

http://www.dpreview.com/forums/read.asp?forum=1012&message=2153499

Please, let's not restart all the BS now - go read the thread.
We've already put this issue to bed.

 

[E'Mahn]

I have read all the posts in these two threads and I have found it frequently entertaining and more than occasionally educational. I am continually amazed at the collective enthusiasm, intelligence and knowledge of the DP Review forum communities.

In these two threads, it is often difficult to separate fact from well conceived fiction. I am certainly no expert on the underlying technical issue, but I do pride myself as a fair analyst of argument and a logical thinker. From my perspective, it seems that it is all coming down to semantics. (It seems many arguments do.)

Jim's main assertion is that a 3 mp sensor can't produce more than 3 mp of information. This is a hard argument to deny. In fact I would guess that this statement really isn't as strong as it could be. I would guess that it would be true, given the fallibility of all things human, that a 3 mp sensor can produce no more than 3 mp of information. Good information anyway. The reality is that information is unavoidably lost or distorted at several stages of the picture taking process - info is lost/distorted as light is transmitted to the lens from the object being captured, the lens itself loses/distorts information, the photosensor loses/distorts information, the never perfect processing algorithms lose/distort information, etc. In the end, you end up with something less than what you were trying to capture with your camera. Although we can minimize the loss and distortion, it's always there to some degree, and it will never completely go away.

So I think in reality what the problem reduces to, is how accurate is each camera? That is, how little error did the camera introduce to the process? The best camera is the one that screws things up the least! The flaw in most of the arguments here, imho, is that people are trying to state how a given camera (the 602/6900) increases the information over another camera, when I think what they really mean is that it loses_less information than a comparable 3mp camera. Is it possible that a 4mp camera produces less good information than a well-engineered 3mp camera. Of course it's possible. This assertion is as difficult to deny as Jim's assertion.

I think viewed from this perspective, the arguments here fall mostly into alignment. Everyone seems to agree that the Fuji camera does a pretty nice job and better than most 3mp cameras. But ultimately, it all comes down to each person's subjective assessment to determine which camera is the best.

I hope this makes sense to someone. It does to me, but then it's very late.

Thanks for all the informative posts.

 

[Jim Gilliland]

Yes Jim. But What is 3mp of information worth? Is the 3mp from a
Canon D30 the same as the 3mp from an average sensor when its
sensor is 12x the size?

Other than light gathering ability, size is irrelevant. As long as the size of the CCD is matched to the image size produced by the lens, the number of pixels in the CCD decides the maximum resolution of the camera. It doesn't matter whether the CCD is an inch across or a centimeter as long as the number of pixels is the same. The larger one should be capable of better ISO, but not better resolution.

If the SuperCCD is doing a better job than the 4mp brigade and it's
giving relatively equal images because it's more efficient, then
why not accept, as many reviewers do that the SuperCCD gives the
'equivalent' resolution of about a 4.5mp camera. We're not claiming
that it gives a true 6mp equivalent here.

No one said anything about 6mp. I haven't seen any reviewer say that the SuperCCD gives the equivalent of a 4.5mp camera. I've seen some imply that the Fuji approaches a higher resolution, but it's never been that specific in the examples that I've seen.

Perhaps we should be argueing why some other cameras seem to give
less than their true resolution if that's a more acceptable way of
putting it.

That would be an interesting discussion to have, and it might shed some light on the benefits of the SuperCCD in a more honest fashion.

 

[Jim Gilliland]

I wonder if what is confusing the issue here is what exactly is
everyone talking about with regard to the information collected and
output by the CCD sensor. You are absolutley correct in stating a
3MP CCD cannot gather more than 3MP of data. But, in the algorithm
that process and output the 6MP file, isn't the final data file
generated composed of 6MP of information, but about half of that is
processed information....information not gathered by the CCD itself
but generated by the camera's processing software

Data that is produced by the interpolation process is not "information" in the scientific sense of the word - and that's exactly the definition that matters to us if we're going to talk about resolution. The file produced by the Fuji contains 6mp of data that represents the same 3mp worth of information that was gathered by the CCD in the first place. The transformation that takes place in the camera's LSI chip reformats the information into a usable image format, but it does not - and cannot - add information.

 

[Jim Gilliland]

E'Mahn, I think you have summed up the discussion beautifully.

You are 100% correct in asserting that the camera must be looked at as a system, and I would extend your idea to point out that the resolution issue that we have been discussing is only one issue among many that determine a camera's worth, and a rather small one at at that - at least once we reach the 3mp, 4mp, 5mp level of performance.

For the sake of the science, I will point out that "information", as I have been discussing it, is what is gathered by the CCD regardless of any flaws in the optical system before it. My discussion has been focusing on the process that takes place starting at the CCD and ending at the image data before JPEG compression. That's where most of the confusion lies, so that's what's been under discussion.

Fuji's marketing fluff is doing just what it is supposed to - it's attempting to convey the benefits of their CCD to a generally non-scientific public. And since the marketing writers aren't scientists themselves, they sometimes get a bit carried away in hyperbole.

But the stuff they've written isn't overly inaccurate. The best thing to do is this - start with these premises as mathematically proven truths:

1) A 3mp CCD cannot gather more than 3mp worth of information, and
2) No algorithm can expand 3mp of image information into more than that.

Then reread the Fuji information. You'll see where they have lapsed into hyperbole, but you'll also be able to read between the lines a bit to understand the real benefits of the system and to filter out the fictional ones.

Here are the real benefits of the SuperCCD:

1) The geometric arrangement of the individual cells allows Fuji to make the cells larger than their competitors. This allows them to gather more light, providing higher ISOs and less noise.

2) The organization of the data collected by the CCD (which is directly related to the geometry of the sensors themselves) allows Fuji to use an image processing algorithm in their LSI chip that loses less of the information than their competitors. However, to achieve this, the algorithm must reorganize the data into a larger image format, thereby requiring the use of a larger filesize to store the image. This feature allows the Fuji to produce better resolution than any other 3mp camera on the market.

3) The organization of the data collected by the CCD also allows Fuji to manipulate that data to provide some other useful capabilities, including 30fps video, very high ISO images at lower resolutions, and fast multishot performance.

Now that's a fairly honest description of the benefits of the SuperCCD, and it doesn't in any way detract from the value of the camera.

I have read all the posts in these two threads and I have found it
frequently entertaining and more than occasionally educational. I
am continually amazed at the collective enthusiasm, intelligence
and knowledge of the DP Review forum communities.

In these two threads, it is often difficult to separate fact from
well conceived fiction. I am certainly no expert on the underlying
technical issue, but I do pride myself as a fair analyst of
argument and a logical thinker. From my perspective, it seems that
it is all coming down to semantics. (It seems many arguments do.)

Jim's main assertion is that a 3 mp sensor can't produce more than
3 mp of information. This is a hard argument to deny. In fact I
would guess that this statement really isn't as strong as it could
be. I would guess that it would be true, given the fallibility of
all things human, that a 3 mp sensor can produce no more than 3
mp of information. Good information anyway. The reality is that
information is unavoidably lost or distorted at several stages of
the picture taking process - info is lost/distorted as light is
transmitted to the lens from the object being captured, the lens
itself loses/distorts information, the photosensor loses/distorts
information, the never perfect processing algorithms lose/distort
information, etc. In the end, you end up with something less than
what you were trying to capture with your camera. Although we can
minimize the loss and distortion, it's always there to some degree,
and it will never completely go away.

So I think in reality what the problem reduces to, is how accurate
is each camera? That is, how little error did the camera introduce
to the process? The best camera is the one that screws things up
the least! The flaw in most of the arguments here, imho, is that
people are trying to state how a given camera (the 602/6900)
increases the information over another camera, when I think what
they really mean is that it loses_less information than a
comparable 3mp camera. Is it possible that a 4mp camera produces
less good information than a well-engineered 3mp camera. Of
course it's possible. This assertion is as difficult to deny as
Jim's assertion.

I think viewed from this perspective, the arguments here fall
mostly into alignment. Everyone seems to agree that the Fuji
camera does a pretty nice job and better than most 3mp cameras.
But ultimately, it all comes down to each person's subjective
assessment to determine which camera is the best.

I hope this makes sense to someone. It does to me, but then it's
very late.

Thanks for all the informative posts.

 

[Aaron_6900]

I do not think that anybody here is arguing the point that 3 mp is 3 mp. I think we all agree on that.

The argument is wether or not the SuperCCD is superior in it's method of capturing and reproducing that quantity of pixels.

The question remains........Should there be an additonal spec added to the overhyped megapixel rating? Is the process of determining of that number flawed? I think Fuji has proven that there is more to simply a count of sensors.

I know my AMD Athlon XP 2000+ chip only runs at a 1.66 ghz clock cycle but it outperforms the Pentium 4 2.0 ghz chip from Intel. Not smoke and mirrors.......hard core tested. Many other elements of the design attribute to that superiority.

As Ian put it.....the issue here is that Fuji seems to be giving us a hell of alot more information from their 3.1 mp spec than the competition's 3.1 mp spec.

So MY OPINION gives credence to this. I feel the 3.1 mp spec from most other manufacturers is flawed at best and inferior to the spec level of information given to me by the SuperCCD.

We can argue all we want.

The fact remains that many reviewers feel it performs as a 4 - 5 mp camera. Cnet considers it a true Prosumer model.

The funniest part of this whole argument is this:

We all see the many posts by forum members comparing the 6900's shots to the Sony 707, Nikon 5000, Canon G2, etc.

Why ? Somebody tell me why we are comparing a 3.1 ccd to 4 and 5 mp units?

I paid 699 USD for my 6900 and can not touch a 707 or 5000 for less than 999 USD here. Pretty impresive there Fuji!!! Your in a class all by yourself as proven by the many comparisons of your inferior 3.1 mp chip to the latest 4 and 5 mp offerings. (And your colors remain superior all along)

Come on!