12 million is 12 million is still 12 million

No

I’m an Engineer, but I watch only medical and policies series, Like House, 24 hours, Law & Order, etc…

You're an Engineer? Hmm. Maybe I was wrong. Maybe your the mirror image of Satori.

He's an engineer, you know. He'll probably remind you of that soon.

Leo Terra said:

No
I’m an Engineer, but I watch only medical and policies series, Like
House, 24 hours, Law & Order, etc…

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

Jonathan Kardell said:

In a thought experiment, try this - for each set of photosites,
create a 12mp image and instead of interpolating data, put a black
pixel in the empty space. When you do this, you wind up with a
checkerboard pattern. The two checkerboards from the two sets of
sensors wll be complimentary, rather than identical. Blend them by
letting the illuminated pixel from one checkerboard replace the
black pixel from the other. Viola' you have an image with more
information.

This is my admittedly amateur attempt to understand where the extra
data comes from. (actually an attempt to explain the discrepancy
between experimental results and simple math 6mp reasoning)

Click to expand...

This is a good discriptpon of the S2. We know where the data for the "black" spaces comes from. They sample the 4 surrounding spaces and average to fill in the black space. Thus the data is guessed or created. So the S2 is creating 6.17mp of resolution data, but the net gain is only 2-2.5mp of "appeared" resolution. I do not say actual because I do not think anyone has really tested it or not.

--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Yes, I’m a Mechanical Engineer (of Airplanes) but I don’t trail the vocation, now I’m work w/ Fashion photograph.

well thats because, if we agree to the theory, that the S3 is creating 12.34mp of data for resolution.
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

DillonJames said:

Uschold wrote that in his latest S3 review. Get the facts straight.

Click to expand...

Link?

--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Tks.

Well, try to download and post your conclusions

I made a test some time ago, and The resolution and quality of R pixels image is better then my old Fuji S5000
You can create good files w/ R pixels, try it

Look, it's really pretty simple, but get's confusing because if the numbers.

The S2 has 8 million photosites. These are arranged in a diagonal pattern, Unlike a regular Bayer sensor, Fuji has to do more than a simole demosaic. They have to covert the file so that the file can be displayed on your screeen. It during this process that another "output" pixel is created. Think about an array of diagonally arrayed pixels, imagine they are flexible in that the rows and columns can shift around, but their intersections can't change. There is no way to simply twist this into a vertical/horizonatl grid without adding pixels.

Okay, so we end up with a 12MP file. A file that actaully resolves more detail than an ordinary 6 MP file interpolated up in Photoshop. If you interploate upo a D70 file in Photoshop, or even if you interpolate uo an S2 "6MP file"n Photoshop, you do NOT get more resolved detail.

So, Fuji's interploation is no ordinary interpolation. Next step.

Enter the S3. Th eS3 has an extar set fo tiny pixels, the R. and these create a entirely different file, just like the S2 pixels do. You end up with two 12MP files. That's why the S3 eaws are 25MB! They are literally two 6MP s-ccd files.

These two files are blended, almost like blending layers in Photoshop, except the tiny R pixels are used only for extreme highlight detail. These small pixels are the size of a digicams pixels, and are quite noisy in anything but higher EV values. That's where the extended Dr comes from.

Now, the thing these guys are 'discussing", is whether the R pixels can contribute any additional detail information. Theoretically, this could be possible, because the R pixels are inbetween the S, in a differnt position. It would be like taking a pic with your S2, and then moving the sensor or camera a half pixel to the right, and then exposing another frame.

Either that, or I have no frig-in idea what I'm talkin about!

Bad Pic kev said:

"Ray, the nest time someone asks if you are a god say yes!"

reread whole thread once i got to work. great back to white papers.
in the microlens, so far i have not seen a yes or a no from fuji on
this. a lot of educated people have speculated both ways.

If we keep to my terminalogy. the S2 is a 6mp I12mp. the S3 is a
6mp/I12mp DR12mp/IDR24mp camera? Any takers as to why its slow.
That is a lot of guessing and a double interpolating as well.
Somewhere It just doesnt seem to be linear on this. The appeared
resolution is 8-11mp with most saying 9-10. I would have thought it
would be higher.

I dont mind giving credit when its due, but a IDR of 24mp and a
visual res of only 9mp, doenst seem right. We are talking about 12
million faked pd's with a net gain of 1mp in res? see my point.
With a 25mb file size that can now be understood, but....
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

I will. Leo, check your email.

Leo Terra said:

Tks.

Well, try to download and post your conclusions

I made a test some time ago, and The resolution and quality of R
pixels image is better then my old Fuji S5000
You can create good files w/ R pixels, try it

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

http://db.riskwaters.com/public/showPage.html?page=129730

Sign up for the two week free trial.

Here's the quote if you are too lazy. )

“The Fujifilm FinePix S3 Pro posesses a very progressive sensor design. Instead of any external interpolation method known to my test laboratory, here the interpolated mode provides a real increase of detail information. Using high-end, well-suited lenses, the camera can compete with eight million pixel models of other manufacturers.”

Bad Pic kev said:

DillonJames said:

Uschold wrote that in his latest S3 review. Get the facts straight.

Click to expand...

Link?

--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

Hey, I'm an EAA member!

http://eaa.org/

Leo Terra said:

Yes, I’m a Mechanical Engineer (of Airplanes) but I don’t trail the
vocation, now I’m work w/ Fashion photograph.

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

DillonJames said:

Look, it's really pretty simple, but get's confusing because if the
numbers.

The S2 has 8 million photosites. These are arranged in a diagonal
pattern, Unlike a regular Bayer sensor, Fuji has to do more than a
simole demosaic. They have to covert the file so that the file can
be displayed on your screeen. It during this process that another
"output" pixel is created. Think about an array of diagonally
arrayed pixels, imagine they are flexible in that the rows and
columns can shift around, but their intersections can't change.
There is no way to simply twist this into a vertical/horizonatl
grid without adding pixels.

Okay, so we end up with a 12MP file. A file that actaully resolves
more detail than an ordinary 6 MP file interpolated up in
Photoshop. If you interploate upo a D70 file in Photoshop, or even
if you interpolate uo an S2 "6MP file"n Photoshop, you do NOT get
more resolved detail.

So, Fuji's interploation is no ordinary interpolation. Next step.

Click to expand...

good ok we are on same page, but i assume the 8mp is a typo, because fuji says its a 6

DillonJames said:

Enter the S3. Th eS3 has an extar set fo tiny pixels, the R. and
these create a entirely different file, just like the S2 pixels do.
You end up with two 12MP files. That's why the S3 eaws are 25MB!
They are literally two 6MP s-ccd files.

Click to expand...

no not really if you just add the R pixels the file size is like 13mb, the S2 it was 12.9.

DillonJames said:

These two files are blended, almost like blending layers in
Photoshop, except the tiny R pixels are used only for extreme
highlight detail. These small pixels are the size of a digicams
pixels, and are quite noisy in anything but higher EV values.
That's where the extended Dr comes from.

Click to expand...

right, just the DR

DillonJames said:

Now, the thing these guys are 'discussing", is whether the R pixels
can contribute any additional detail information. Theoretically,
this could be possible, because the R pixels are inbetween the S,
in a differnt position. It would be like taking a pic with your S2,
and then moving the sensor or camera a half pixel to the right, and
then exposing another frame.

Click to expand...

right, i am with you here. I do not agree that it is being used for res. but could be. Data is data, adjust it to fit the S pixel data and bing its there. But can it, will it, and whats the cost of it, I dont know.

DillonJames said:

Either that, or I have no frig-in idea what I'm talkin about!

Click to expand...

Derrel and Gabe will agree here.

Now to readresss this the S2 in I12mp mode is 12.9mp that is 6 real and 6 fake mp sets of data combined to a pic or file. This gives the appearence of 2-2.5mp gain in resolution.

the S2 has 4 res modes 1.4, 3.5, 6.1,and I12.1mp

the S3 has 8 res modes 1.4, 3.5, 6.1,and I12.1mp plus the same modes with the wide DR turned on. It is generally accepted that the DR gain of 1-2 is real and in some measured cases up to 3 and almost 4. Most DSLR have about 8 stops of DR. Thus the S3 has 2. keeping with generally accpted numbers.

My postings have been on a simple note of theory and known facts. we will sat the S pixels are for res, and the r pix for dr. we have 6.17 res pix and 6.17 dr pix. when combined it makes up to a 12.34mp pic that has 6.17res and 10 stops of dr. It will not keed all the s and r data that is duped, so one set would be tossed. then in I12mode, the S3 guesses a complete set of 6.17s and 6.17r pixels. so in max IDR mode we have a total of 24mp. with half being guessed or faked. that is why the file size is so huge.

i may not be 100% right, but i am close
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Bad Pic kev said:

DillonJames said:

Look, it's really pretty simple, but get's confusing because if the
numbers.

The S2 has 8 million photosites. These are arranged in a diagonal
pattern, Unlike a regular Bayer sensor, Fuji has to do more than a
simole demosaic. They have to covert the file so that the file can
be displayed on your screeen. It during this process that another
"output" pixel is created. Think about an array of diagonally
arrayed pixels, imagine they are flexible in that the rows and
columns can shift around, but their intersections can't change.
There is no way to simply twist this into a vertical/horizonatl
grid without adding pixels.

Okay, so we end up with a 12MP file. A file that actaully resolves
more detail than an ordinary 6 MP file interpolated up in
Photoshop. If you interploate upo a D70 file in Photoshop, or even
if you interpolate uo an S2 "6MP file"n Photoshop, you do NOT get
more resolved detail.

So, Fuji's interploation is no ordinary interpolation. Next step.

Click to expand...

good ok we are on same page, but i assume the 8mp is a typo,
because fuji says its a 6

Bad Pic kev said:

Enter the S3. Th eS3 has an extar set fo tiny pixels, the R. and
these create a entirely different file, just like the S2 pixels do.
You end up with two 12MP files. That's why the S3 eaws are 25MB!
They are literally two 6MP s-ccd files.

Click to expand...

no not really if you just add the R pixels the file size is like
13mb, the S2 it was 12.9.

Bad Pic kev said:

These two files are blended, almost like blending layers in
Photoshop, except the tiny R pixels are used only for extreme
highlight detail. These small pixels are the size of a digicams
pixels, and are quite noisy in anything but higher EV values.
That's where the extended Dr comes from.

Click to expand...

right, just the DR

Bad Pic kev said:

Now, the thing these guys are 'discussing", is whether the R pixels
can contribute any additional detail information. Theoretically,
this could be possible, because the R pixels are inbetween the S,
in a differnt position. It would be like taking a pic with your S2,
and then moving the sensor or camera a half pixel to the right, and
then exposing another frame.

Click to expand...

right, i am with you here. I do not agree that it is being used for
res. but could be. Data is data, adjust it to fit the S pixel data
and bing its there. But can it, will it, and whats the cost of it,
I dont know.

Bad Pic kev said:

Either that, or I have no frig-in idea what I'm talkin about!

Click to expand...

Derrel and Gabe will agree here.

Now to readresss this the S2 in I12mp mode is 12.9mp that is 6 real
and 6 fake mp sets of data combined to a pic or file. This gives
the appearence of 2-2.5mp gain in resolution.

the S2 has 4 res modes 1.4, 3.5, 6.1,and I12.1mp
the S3 has 8 res modes 1.4, 3.5, 6.1,and I12.1mp plus the same
modes with the wide DR turned on. It is generally accepted that the
DR gain of 1-2 is real and in some measured cases up to 3 and
almost 4. Most DSLR have about 8 stops of DR. Thus the S3 has 2.
keeping with generally accpted numbers.

My postings have been on a simple note of theory and known facts.
we will sat the S pixels are for res, and the r pix for dr. we have
6.17 res pix and 6.17 dr pix. when combined it makes up to a
12.34mp pic that has 6.17res and 10 stops of dr. It will not keed
all the s and r data that is duped, so one set would be tossed.
then in I12mode, the S3 guesses a complete set of 6.17s and 6.17r
pixels. so in max IDR mode we have a total of 24mp. with half being
guessed or faked. that is why the file size is so huge.

i may not be 100% right, but i am close
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

Jonathan Kardell said:

This is a semantic argument, Gabriele. We do, in fact, have a 12mp
camera, it just resolves like an 8. No reasonable person claims
that it resolves like a D2x. Fuji sacrificed resolution for DR. A
fair trade in my book. Fuji should probably adopt the Intel and
AMD model of "performance ratings". They should probably have
called it the S8 and not told people how many actual photo sites
there were

Click to expand...

But I agree Jonathan, the fact is that in a few posts the S3 was coming out as a 12MP equivalent res. camera. Since the S2 Fuji had a resolution advantage over other 6MP but it is not giganting, actually hard to see in printed photos unless someone is just measurbating. I can see a big jump in resolution from my S2 to my Kodak srl/n but the pixel are more than double and the Kodak has no AA filter (this alone gives an unfair advantage at the cost of ocasional artifacts). If there aren't 30 to 50% more pixels on each of the two axes, on printed photos is hard to tell the difference, even if they are a crop.

That is the reason why cameras like the D2H are still doing a good job and the reason why DR is so important although not everything. I agree that Fuji with the same cost of building the S3 could have built a real 12MP keeping the DR of the S2. Such camera would have been a res. contender of the D2X. Although I consider DR like a third dimension in calculating how the human eye perceive details, and I like DR, under a marketing point of view FUji would have probably done better doubling the # of pixels (and doubling the buffer, and compressing RAWs) and using all of them for resolution. I worked at AMD 7 years and I know how is difficult to take away from the people's head simple to understand things like numbers.

--
Regards
Gabriele
California, CA

Jonathan Kardell said:

You are quite right, you also have a reasonable explanation for the
bump in resolution, the effect of dynamic range on perception.
Thanks for the even tempered response!

Click to expand...

I'm always tempered with tempered people. I'm profoundly un-tempered with unreasonable and dishonest people. It is a pleasure to chat with you jonathan.

--
Regards
Gabriele
California, CA

confused on this one. you quoted me but did not add to it. was that planned?
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

I can't post. Let try this.

Bad Pic kev said:

confused on this one. you quoted me but did not add to it. was that
planned?
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

Click to expand...

--

This info has been brought to you by a 100% genuine S3 user/owner. Beware of substitutes!

Ghosts ghosts go away, please don't come back another day...

Bad Pic kev said:

created. So the S2 is creating 6.17mp of resolution data, but the
net gain is only 2-2.5mp of "appeared" resolution. I do not say
actual because I do not think anyone has really tested it or not.

Click to expand...

Hi Kev, old friend, Nobody ever did a test that directly translated in equivalent pixels but it is easy to make your own table as long as you take a camera as "reference". If you go on the test section of the site you can get how many lines each camera resolve on both axes and starting from that make your own equivalent pixel count. I like that method since Phil has been always very consistent on determining the right extintion point. After a certain point some camera delivery nothing, others mud, others some noise, others artifacts and finally some one imaginary details. It takes an honest man to define what is the highest resolution from the charts and Phil has been coherent with his methods.

You will see a higher number from both the S2 and the S3 when compared to the D100 for example. FOr some cameras like the 16MP Canon and the Kodak 14MP the chart hasn't enough details for the camera so Phil had to take the photo at a higher distance and compensate for that but he still reports the final numbers right.

--
Regards
Gabriele
California, CA

DillonJames said:

I can't post. Let try this.

Click to expand...

now thats funny
--
Kodachrome, they give us those nice bright colours
They give us the greens of summers
Makes you think all the world’s a sunny day, oh yeah
I got a nikon camera, I love to take a photograph

There are complet and seperate images being created. If you convert an S3 raw with s7raw, you can actually cretae two seperate images. One form the S, and obe form the R. Th eR image looks like cra_p, all except the highlights. I also a a few clumps of stuck pixles in my R's, but I only see them in the shadows whe I extracy out the R using s7raw. The stuck pixels don['t matter, since I never see them wherr it counts, in the highlights.