New S2 Photos

[Andrew Martin]

I just printed the wedding pic using my Epson 810/820 on cheap
Jessops photo gloss paper, I think the pic is wonderful.
People who condemn digital photography or pick faults at a picture
because of some minor discrepancy need to sit down and take a
break. 135 format (35mm to the unknowing) is able to resolve more
detail but only if the Developing and Printing are up to scratch.
Most digilabs now scan your negatives with a relatively low
resolution scanner and print at about 300dpi (some will print at
500dpi) therefor the full potential of a 35mm negative cannot be
realised. Just look at the specs of a Fuji Frontier digilab, they
have been superceded but the labs who spent all that money 1-2
years ago to go digital, will not spend more money on a new scanner
to improve the quality of your negative based pix.
My wife said that the quality of the wedding pic is good enough for
anyones album, and Im happy with that.
Best Wishes All
Charles UK

At last - someone on the same wavelength. People will always pick
faults with Fuji. I had the S1 and lots of people complained about
body quality and lack of features etc.. but the printed images were
superior to anything in it's price range. I sold my S1 amd bought a
D60 and changed all of my glass. I had to buy 'L' series glass to
get anything like what I would call acceptable sharp images. I have
an S2 on order an will sell the D60 as soon as it arrives I am
pretty certain it will have better picture quality and the firewire
will be a big bonus for on-site prints etc.

Garry F

I too have a S1 and I am a of the oppinon that for my use it was the best camera to get. It is the print results that are the important isue. For this reason I will be geting the S2 when it copmes out the S1 will be backup plus the wifes camera hehehe.

 

[jdavis37]

Flip a coin.. like the line form an old Dire Straits song... " two men claim to be Jesus... at least one must be wrong " ( or something like that )... point being we have a user manual and now a detailed brochure that do not necessarily agree ( the user manual however does not say the lower resolutions dont support RAW, just doesnt say they do either! ). One or the other is wrong Just not sure which ( though Id put my $$ on the manual being right til I hear otherwise )

jimmy wooten wrote:

You are right, it can be done.

Page 32 of the User manual : there RAW for the highest resolution
again.

But I am praying with you.



Jean

 

[Dick Lau of Calgary]

Anthony, you are great.
It is just too many color and pixel for some.

****

Of course the Nikon D100 can cram more onto a card in RAW mode
because its CCD (or in this case Sony's CCD) doesn't produce as
much data. I think the Fuji SuperCCD simply produces more data
because of its pixel layout. There may not simply be a way to
throw out the data in a RAW format. After all, that's why it's a
RAW format. It's not like you can cut out a few bits of data and
make it perfect.

Or as I once saw in the movie "Amadeus"--

EMPEROR: Well, Herr Mozart! A good effort. Decidedly that. An
excellent effort! You've shown us something quite new today.

[Mozart bows frantically: he is over-excited.]

MOZART: It is new, it is, isn't it, Sire?

EMPEROR: Yes, indeed.

MOZART: So then you like it? You really like it, Your Majesty?

EMPEROR: Of course I do. It's very good. Of course now and then -
just now and then - it gets a touch elaborate.

MOZART: What do you mean, Sire?

EMPEROR: Well, I mean occasionally it seems to have, how shall one
say? [he stops in difficulty; turning to Orsini-Rosenberg] How
shall one say, Director?

ORSINI-ROSENBERG: Too many notes, Your Majesty?

EMPEROR: Exactly. Very well put. Too many notes.

MOZART: I don't understand. There are just as many notes, Majesty,
as are required. Neither more nor less.

EMPEROR: My dear fellow, there are in fact only so many notes the
ear can hear in the course of an evening. I think I'm right in
saying that, aren't I, Court Composer?

SALIERI: Yes! yes! er, on the whole, yes, Majesty.

MOZART: But this is absurd!

EMPEROR: My dear, young man, don't take it too hard. Your work is
ingenious. It's quality work. And there are simply too many notes,
that's all. Cut a few and it will be perfect.

MOZART: Which few did you have in mind, Majesty?

EMPEROR: Well. There it is.

p.s. me and mr davis was addressing the issue of the raw file size
of the s2 being 12 mb and hoping fuji would address this problem.

 

[Anthony32599]

What you say regarding the image size would make sense for a traditional CCD with a traditional layout. As I understand it though, Fuji's layout in the SuperCCD actually generates more information than a standard CCD layout, so you cannot go by the size of the recorded image because the SuperCCD pixels do not map on a 1 to 1 basis.

Anthony

Fiji interpolates the raw file up to a 12mp size, if you check the
S2's manual on page 33 you'll see that the raw file comes out at a
size of 4256x2848, the actual size of the sensor is 3024x2016 so
Fuji interpolates extra pixels into the raw file instead of just
letting it come out just as it should with no interpolation thereby
using less memory and because this raw image has already been
resized if you resize it again in photoshop you'll run the risk of
introducing resizing artifacts into the image.

Dennis D

 

[Adrian Gilbert]

jimmy wooten wrote:

You are right, it can be done.

Page 32 of the User manual : there RAW for the highest resolution
again.

But I am praying with you.



Jean

I handled the S2 this afternoon. Raw is only available at the highest resolution only.I shot 7 files in raw mode continuosly - with the buffer full it took about 40 seconds to flush the buffer.(Sandisk 512 ultra card)

Raw may be a mode suited for specific shots - not a fast moving wedding.
Fantastic enlargements were displayed - 30" x whatever.
Impressive resolution and smooth colour from jpeg files.

S2 uses the same red grid viewfinder screen which isnt as large as the dummy models professed to offer.Still ,nice camera which should sell in the bucketload.........now where was i on that waiting list........

 

[Dennis D]

But if the image comes out of the camera with more pixels than the CCD, then the camera added pixels via intepolation.
Dennis D

Anthony

Fiji interpolates the raw file up to a 12mp size, if you check the
S2's manual on page 33 you'll see that the raw file comes out at a
size of 4256x2848, the actual size of the sensor is 3024x2016 so
Fuji interpolates extra pixels into the raw file instead of just
letting it come out just as it should with no interpolation thereby
using less memory and because this raw image has already been
resized if you resize it again in photoshop you'll run the risk of
introducing resizing artifacts into the image.

Dennis D

--
Dennis D

 

[Moshe Vainer]

i do understand what you are saying, the 12 mb raw file size
(verus a 9mb raw from the d100) is because of the fuju sensor. both
are based on approx 6mb orig captiure.

Nope... The size difference is because fuji decided to use "padding" in their firmware, meaning that they put 12bit of data into 16bits of storage. A standard thing done in computers all the time, unfortunately, not very smart when you are low on storage... This has nothing to do with fuji's sensor, or it's layout...

i am still #3 on the s2 list in atlanta, but my point was one of
$$$ in my original post and my hopes were that the raw file in the
s2 could be given a option of compressed raw so i could determine
if i wanted raw quality with jpeg cost based on my needs and work
in progress.

But that would be either very slow or very expensive.. compromises...

Second,
a) both fuji, and nikon record 6 million pixels in their RAW file
b) both fuji and nikon has 12 bits of information per each recorded
pixel.
c) Nikon packs the bits, so it needs 3 bytes to record 2 pixels
worth of information, while fuji does not pack the bits, therefore
it needs 4 bytes to record 2 pixels worth of information from the
RAW file. Hence the difference of 9mb = 6,000,000 * 12bit = 9 mb
--- vs ---- 12mb = 6,000,000 * 16bit = 12mb.
d) and most important - this has nothing, let me repeat, nothing to
do with fuji's interpolation. The raw file contains info from the
sensor, prior to any interpolation.

Third
Fuji's interpolation is very different from what you might get from
photoshop. There is actually 1.4 times more information in vertical
and horizontal axis in fuji sensor, and 1.4 times less information
in diagonal axis on fuji sensor. There is no way to record this
information in any less than twice the amount of pixels, if you
want to use any standard format (rectangular format). Imagine
square A, turned 45 degrees. Now what size of square B you need so
that square A fits inside square B? - answer - exactly twice the
size.

Rgrds,
Moshe

 

[Moshe Vainer]

Ben, compressed is not always lossy. Sometimes compressed is simply smaller. Take any file on your pc and put it in a zip archive. when you open it, it's the same file, nothing has been lost...

Rgrds,
Moshe

the COMPRESSED RAW of the d100 camera is 4.56 mb.

Hmmm, If something is "Compressed" can it be called "Raw?"

When looking at a Tiff next to a jpg it looks soft.
Is there not More info for the eye to see, more shades
for the eye to define. Take the Tiff and reduce it to 256
colors or black & white and it may look sharper.

Maybe if there were square lense...

Maybe Anthony should explain how 3 rights make a left.

It seems that some should drink tea and cut down on the wineing.

Ben 8)

 

[Moshe Vainer]

Anthony, we are almost in agreement.

see inside:

Moshe--
What I meant by RAW is RAW is that it's the data from the sensor
prior to interpolation,

right,

and thus it makes no sense to beg Fuji for
a compressed RAW, which would mean either loss of data or in-camera
interpolation, thus defeating the purpose of having a RAW file.

wrong, - see my answer to ben, compressed is not always lossy, take Zip compression for example. When you put a file into a zip archive, and then open it, it's the same file, nothing is lost.

So you and I are actually in agreement. Which means we're both
right. Right? Or is that wrong? Do two rights make a wrong? Of
if we're both wrong, then do two rights make a wrong? I do know
that 3 lefts make a right, but that's another story.

Anthony (Good thing my last name is Ho and not Wong. Otherwise,
I'd be accused of doing things the Wong Way.)

, I am sorry about the title, it was a little bit rude of me...

Sorry to jump in with such a statement, but it amazes me how people
do not understand what they talk about.

First of all, fuji's RAW file is 12 MB, not 12MP, Nikon RAW file is
9MB...

Second,
a) both fuji, and nikon record 6 million pixels in their RAW file
b) both fuji and nikon has 12 bits of information per each recorded
pixel.
c) Nikon packs the bits, so it needs 3 bytes to record 2 pixels
worth of information, while fuji does not pack the bits, therefore
it needs 4 bytes to record 2 pixels worth of information from the
RAW file. Hence the difference of 9mb = 6,000,000 * 12bit = 9 mb
--- vs ---- 12mb = 6,000,000 * 16bit = 12mb.
d) and most important - this has nothing, let me repeat, nothing to
do with fuji's interpolation. The raw file contains info from the
sensor, prior to any interpolation.

Third
Fuji's interpolation is very different from what you might get from
photoshop. There is actually 1.4 times more information in vertical
and horizontal axis in fuji sensor, and 1.4 times less information
in diagonal axis on fuji sensor. There is no way to record this
information in any less than twice the amount of pixels, if you
want to use any standard format (rectangular format). Imagine
square A, turned 45 degrees. Now what size of square B you need so
that square A fits inside square B? - answer - exactly twice the
size.

Rgrds,
Moshe

 

[Moshe Vainer]

Dennis, you could not be more wrong than that...

Simple math could prove you that -

See my comments inside:

Sorry to jump in with such a statement, but it amazes me how people
do not understand what they talk about.

First of all, fuji's RAW file is 12 MB, not 12MP, Nikon RAW file is
9MB...

Fiji interpolates the raw file up to a 12mp size, if you check the
S2's manual on page 33 you'll see that the raw file comes out at a
size of 4256x2848, the actual size of the sensor is 3024x2016 so
Fuji interpolates extra pixels into the raw file instead of just
letting it come out just as it should with no interpolation thereby
using less memory and because this raw image has already been
resized if you resize it again in photoshop you'll run the risk of
introducing resizing artifacts into the image.

The interpolation is done in RAW conversion software, RAW itself contains only 6mp.

First of all, fuji's interpolation contains debayering, effectively tripling the file size (the color information has to be taken into account when interpolating), but even if debayering is done later, if you were right, let's see what the minimum file size fuji would need to keep the interpolated file as their raw:

Every sensor produces 12bit worth of information, therefore:

4256x2848x12bit = 145453056 bit / 8 (bit per byte) = 18181632 bytes = 17mb.

That's right, if you are right, fuji raw file should be at least 17mb.

Third
Fuji's interpolation is very different from what you might get from
photoshop.

Actually there is no difference between Photoshops interpolated
image and an image that has been interpolated in the Fuji camera. I
did a test where I took to shots with the S1, one at the 3mp size
and one at the 6mp interpolated size. I resized the 3mp image in
Photoshop to the same exact size as the 6mp image out of the
camera. I studied both images for a good hour to see if there were
any differences, both images looked exactly the same, so I could
see no reason to let the camera do the interpolation and save
memory space also as I mentioned before the artifacts that come
from too much resizing, it's better to do it just once in Photoshop.
Dennis D

Wrong again, the fact that your subjective analysis did not show you any difference does not mean there is none. Take a tiff file, and convert it to least compressed Jpeg. I bet your eyes would not see a difference either, but it will be there. Just check in photoshop how many colors you loose even if you compress at 12 level jpeg compression..

Fuji's algorithm is based on pithagoras theorem, and it has been right for the last several hundreds of years, unless you suddenly proven it wrong.

In a 90 degree triangle, two of the sides are sqrt(2) shorter than the diagonal. And in a rectangular matrix, 45deg diagonal lines has sqrt(2) more density than horizontal and vertical.

Rgrds.
Moshe

 

[Moshe Vainer]

Your concern seems to be a valid way of looking at things but the
proof is not supporting your theory. In the test target resolution
shots it is clearly the large 12 MB RAW file that is by far the
cleanest and better file.

If you could see the raw image before it was interpolated you would
see the difference.

The upsampling in the camera is a very competent
method making you believe the actual CCD is that large.(4256x3024)

You can do the same thing with Photoshop, I've done it before using
the S1's images (see my message above).

Using the tif that was 3024x2016 pixels and shot on "org" cannot be
upsampled in anything and look like the camera RAW file.

Of course not you'd be comparing raw to TIFF's, TIFF's don't
compare to raw in any camera.

Down sampling is a fact of life in this industry and it does no harm but
upsampling is strictly to be avoided as it always causes side effects.

Downsampling does do harm! That's why JPEGs don't look as good as
TIFFs or raw, they've been downsampled.
Dennis D

Downsampling indeed is harmfull, but jpegs are not downsampled, they are compressed, that is somewhat different than downsampling...

 

[Doug Jones]

I handled the S2 this afternoon. Raw is only available at the
highest resolution only.I shot 7 files in raw mode continuosly -
with the buffer full it took about 40 seconds to flush the
buffer.(Sandisk 512 ultra card)
(snip)....

I only came into this thread late and haven't read them all but I gather there is some concern that Raw isn't available for lower resolutions. As far as I know Raw is only available in all pro cameras for the native highest resolution mode.

There has been a lot of discussion in the Fuji Talk forum about what is the "native" resolution of the 6900 (which is very similar to the S1 apart from physical sensor size). It must be remembered that the 6Mp image size of the S1 and 6900 is the true (even if interpolated) native resolution of these cameras. The lower 3Mp res images are post processed in-camera from the 6Mp image to the reduced resolution. There is no way that a 3Mp image can be easily "interpolated" from the Super CCD receptor layout.

This is because the receptors are not in a standard grid arangement like the pixels of a normal image file. They are in a diamond pattern so they are in no way suitable for easy conversion from the 3Mp diamond pattern to the 3Mp regular grid required for a normal image file.

So the Raw file from the S2 has to be one of two things. Either the in-camera conversion of the Super CCD layout to the native regular grid layout of 12Mp (which I doubt because of file sizes). Or it is the straight dump (as in other cameras) of the individual receptor information - still in the Super CCD layout. If the latter then the Raw processing software supplied by Fuji will have to initially - before anything can be displayed on screen - do the same interpolation as the camera does to create a 12Mp "native" regular grid image.

In conclusion I would just emphasize that the native image resolution of Super CCD cameras is the interpolated high resolution one. And the lower resolution image close to the actual receptor count is a reduced resolution version of the higher resolution interpolated image which means that two interpolation steps are required to produce it.
--
Doug Jones
Canberra
http://www.panamagic.com.au

 

[Moshe Vainer]

I do not mean to blow your bubble but the S1 images upsampled to
the large size were indeed pretty much the same if not better that
the camera's large interpolated files. Not so on the S2! The Super
CCD (second generation) is much better when used at the highest
interpolated resolution. The proof is visible in the tests of the
targets.

It could be that the S2 will be able to interpolate better than the
S1 did. we'll have to wait and see, but I really doubt that they
can do any better than any of the programs that were made
especially to do that. After all if Fuji could do a better job of
interpolating than GF, then they would make a fortune if they made
a photoshop plug-in for resizing.

Dennis, Fuji could not produce any plugin to do what they do in-camera, because the original data has to be aligned in a diagonal pattern to record more vertical and horizontal information. In fact, there is no way they can give you 6mp rectangular matrix from 6mp diagonal matrix without loosing information, and you would not want that to happen, would you?

Downsizing does not harm an image if the final size is what you
need for your particular application. It might even give the
impression that it is better as many people are reporting. Any down
size change will cause the rebuilding of pixels but looking at them
in close up is a dumb idea given that you down sampled for a the
valid reason to use it smaller.

My point is that the more you resample the more damage you do to
the image and it doesn't matter if your going down or up it much
better to just do it once. That's why I wished Fuji didn't resample
the raw image as I prefer to resample myself.

They do not resample it, they simply put a diagonal matrix into rectangular one. If you want to do it yourself, take their raw file, and write your own algorithm to do it. You would still end up with twice the size of the file to fit all the information that is there...

Do a simple test - take any image that all it's pixels has different value in them, and try to rotate it losslesly 45 degrees. Tell me what size of the matrix you need to accomodate all (note all) the information that was in the original file. To test yourself, try rotating it back, and comparing to the original. If all the pixels has the same values as were before you rotated it, you have succeded. P.S., if you have used a matrix that has less than twice the original pixels, give a call to Nobel prize commitee, i think they would like to hear about it..

If you resize a raw
image (up or down) out of the S2 then that's the second time the
image has been resampled. They make programs that are specialized
in downsampling images, the reason, because some programs do more
damage in downsizing than others. Does downsampling damage the
image, of course!

Jpegging is not down sampling in
the true meaning, it is compressing!

It's getting rid of information, the same as downsampling does!
Either way the more you do it the more damage you do, that is my
whole point!
Dennis D

 

[Moshe Vainer]

Kevin, Dennis is only part right - the more you resample, the more you loose, that's right, however, Fuji has left the file alone, it contains only the information that came from the CCD, no interpolation is done to the RAW file.

Rinus, Dennis is right! The more you resample an image the more
artifacts will show up. I get what Dennis is saying, Fuji should
have just left the raw file alone instead of interpolating it, so
now anytime you change the size of the raw file (larger or smaller)
you'll just be getting that much more artifacts!

 

[Moshe Vainer]

Michael, you are only part - right:

First, the raw file is just that - raw, everything a camera does is done later, except one thing that is done first - ISO adjustment. ISO is a value that determines the level of amplification on analog level, before the A/D conversion, and therefore before any information is read from the sensor.

So the correct flow IMO is:

1. Apply ISO to the sensor
2. Take the picture.
3. A/D conversion
4. Readout - > RAW file is taken here.
5. Rotate Matrix (what you call interpolation)
6. Debayer (Don't forget debayering)

7. All the rest is probably as you say, except sharpening is probably done after downsampling.

Michael

It's really not 100% raw data because it's been interpolated (the
addition of pixels) the CCD only takes 6mp of data but the camera
interpolates it up to 12mp of data. What he was talking about is he
wished Fuji didn't interpolate the raw file so it wouldn't take up
so much memory. Another thing to think about is that once an image
has been interpolated then if you resize it again later in
Photoshop to fit a certain size for print then you run the risk of
introducing artifacts into the image from too mush resizing. This
feature of interpolating the raw data alone has kept me from buying
the S2
Dennis D

 

[Moshe Vainer]

With only problem being that at 6mp resolution, you would have a 3mp raw file...

hey guys,

i am one happy guy now and there is no question of which camera is
mine if what i read is true and this statement comes from the fuji
home page. it is on page 4 of the s2 brochure and it states and i
sure hope it states it truly:

" High Quality Options"

" The s2 can record at four resolutions, .........blah blah ....
. At each resolution users have a choice of three file
formats:CCD-RAW, TIFF-RGB and JPEG."

will one of you guys please confirm this as i don t believe my
eyes, it is better than i was wishing for. jimmy

 

[Doug Jones]

Michael, you are only part - right:
First, the raw file is just that - raw, everything a camera does is
done later, except one thing that is done first - ISO adjustment.
ISO is a value that determines the level of amplification on analog
level, before the A/D conversion, and therefore before any
information is read from the sensor.

So the correct flow IMO is:

1. Apply ISO to the sensor
2. Take the picture.
3. A/D conversion
4. Readout - > RAW file is taken here.
5. Rotate Matrix (what you call interpolation)
6. Debayer (Don't forget debayering)
7. All the rest is probably as you say, except sharpening is
probably done after downsampling.

Moshe,

I can see we are on the same wavelength as far as the Super CCD layout is concerned. (a bit like preaching to the un-converted??). I replied on the thread below before I read a few more threads (including yours!!). A slightly different "explanation" which is along your lines.
http://www.dpreview.com/forums/read.asp?forum=1020&message=2911074
(6900z and soon S2)
--
Doug Jones
Canberra
http://www.panamagic.com.au

 

[Moshe Vainer]

Hi Doug, I am a 6900er myself, and still undecided on s2 vs the d100...

In any case, i think what people are confused about, is that d100 produces 9mb raw files, while s2 gives 12mb raw files. This causes people to jump into conclusions regarding "interpolation" in fuji raw files, without stopping to do a simple math, that would easily reveal how wrong they are. After all 12/9 is 4/3, and not 2/1 ... They are confused because fuji choose to pad 12bit of info into 2bytes..

Reminds me how important math education is, even if you are not going to be a mathematician...

Rgrds,
Moshe

P.S., read above my Nobel Prize message.., I think this was the best way so far i came up with to explain myself..

Michael, you are only part - right:
First, the raw file is just that - raw, everything a camera does is
done later, except one thing that is done first - ISO adjustment.
ISO is a value that determines the level of amplification on analog
level, before the A/D conversion, and therefore before any
information is read from the sensor.

So the correct flow IMO is:

1. Apply ISO to the sensor
2. Take the picture.
3. A/D conversion
4. Readout - > RAW file is taken here.
5. Rotate Matrix (what you call interpolation)
6. Debayer (Don't forget debayering)
7. All the rest is probably as you say, except sharpening is
probably done after downsampling.

Moshe,
I can see we are on the same wavelength as far as the Super CCD
layout is concerned. (a bit like preaching to the un-converted??).
I replied on the thread below before I read a few more threads
(including yours!!). A slightly different "explanation" which is
along your lines.
http://www.dpreview.com/forums/read.asp?forum=1020&message=2911074
(6900z and soon S2)
--
Doug Jones
Canberra
http://www.panamagic.com.au

 

[Doug Jones]

Hi Doug, I am a 6900er myself, and still undecided on s2 vs the
d100...

In any case, i think what people are confused about, is that d100
produces 9mb raw files, while s2 gives 12mb raw files. This causes
people to jump into conclusions regarding "interpolation" in fuji
raw files, without stopping to do a simple math, that would easily
reveal how wrong they are. After all 12/9 is 4/3, and not 2/1
... They are confused because fuji choose to pad 12bit of info
into 2bytes..

Reminds me how important math education is, even if you are not
going to be a mathematician...

Yes, fortunately (or perhaps un-fortunately because it dates me!!) most of my maths education was in the 60's. I am yet to be convinced that my children received anywhere near the same quality maths education as I did.

But when it comes to the S2 - I have printed the images of the same subject from all 3, S2, D100 and D60 at 20x30 inches on my Epson 7500 and the S2 just blows the others away IMHO. If you are into large format printing (as I am) then the S2 is what you are waiting for. None of the other 5-6Mp (including "pro") cameras were enough of an improvement over the 6900 to make an upgrade worthwhile. The S2 is the level I have been waiting for. 35mm film is "dead".
--
Doug Jones
Canberra
http://www.panamagic.com.au

 

[Moshe Vainer]

Hi Doug, I am a 6900er myself, and still undecided on s2 vs the
d100...

In any case, i think what people are confused about, is that d100
produces 9mb raw files, while s2 gives 12mb raw files. This causes
people to jump into conclusions regarding "interpolation" in fuji
raw files, without stopping to do a simple math, that would easily
reveal how wrong they are. After all 12/9 is 4/3, and not 2/1
... They are confused because fuji choose to pad 12bit of info
into 2bytes..

Reminds me how important math education is, even if you are not
going to be a mathematician...

Yes, fortunately (or perhaps un-fortunately because it dates me!!)
most of my maths education was in the 60's. I am yet to be
convinced that my children received anywhere near the same quality
maths education as I did.

Sigh...

But when it comes to the S2 - I have printed the images of the same
subject from all 3, S2, D100 and D60 at 20x30 inches on my Epson
7500 and the S2 just blows the others away IMHO. If you are into
large format printing (as I am) then the S2 is what you are waiting
for. None of the other 5-6Mp (including "pro") cameras were enough
of an improvement over the 6900 to make an upgrade worthwhile. The
S2 is the level I have been waiting for. 35mm film is "dead".

Well, the difference between us (except age, as i don't think i was even planned in the 60s , is that i don't print that large, i actually am more interested in high iso performance (photographing mostly my son in motion/natural light). That's why although i acknowledge the definite edge S2 has in resolution, i am still sitting on the fence between it and the D100. Both are good performers in that regard, i'll just wait and see for more samples to show up...

--
Doug Jones
Canberra
http://www.panamagic.com.au