Fuji S7000

Discussion in 'Digital Photography' started by Mattearoadie, Feb 3, 2004.

  1. Mattearoadie

    Mattearoadie Guest

    what does 12.3 milllion recorded pixels mean vs
    6.3 million effective pixels?

    i am looking at the fuji S7000

    any feedback
     
    Mattearoadie, Feb 3, 2004
    #1
    1. Advertising

  2. Mattearoadie

    Bill Guest

    The S7000 uses a 6MP sensor, but it can be set to save 12MP size files. The
    camera's internal firmware uses the captured data and creates a 12MP file if
    you want to do that. You also have the option to save in 6MP, 3MP or smaller
    sizes as well.

    Bill

    "Mattearoadie" <> wrote in message
    news:...
    > what does 12.3 milllion recorded pixels mean vs
    > 6.3 million effective pixels?
    >
    > i am looking at the fuji S7000
    >
    > any feedback
     
    Bill, Feb 3, 2004
    #2
    1. Advertising

  3. "Mattearoadie" <> wrote in message
    news:...
    > what does 12.3 milllion recorded pixels mean vs
    > 6.3 million effective pixels?
    >
    > i am looking at the fuji S7000


    It means the camera normally builds 6M full color pixels interpolated from
    1.5M Red sensors, 3M Green sensors, and 1.5M Blue sensors. Optionally, it
    will build a 12M pixel image from the same number of sensors.

    There is normally no optical benefit whatsoever to interpolation, but Fuji's
    double-interpolation (the 12M pixel output) does enjoy a little bit of
    optical benefit, because the brick pattern layout of Fuji's SuperCCD will
    slightly underperform when compared to other 6MP cameras unless it is
    oversampled. This is because images are ultimately viewed in a conventional
    grid pattern, and some info is lost during the
    brick-pattern-->conventional-pattern coordinate conversion unless an
    oversamplimg method is employed.

    It's a long story but, in a nut shell, Fuji claims they use a brick pattern
    because it slightly increases vertical and horzontal resolution compared to
    a conventional grid pattern. This is true. It is also true (the part they
    "forget" to tell you) that a brick pattern decreases diagonal resolution by
    a corresponding amount, so the net gain in overall resolution is exactly
    zero.

    Then why do it? Because vertical and horizontal resolution is all that is
    usually tested.

    So what are you left with when you buy into Fuji SuperCCD? A 6MP image with
    slightly lower resoultion that a conventional 6MP camera, or, a 12MP image
    with the slightly better vertical and horizontal resolution than a 6MP
    camera, and correspondingly worse diagonal resolution than a 6MP
    camera--resulting overall resolution: 6MP. If all a reviewer tests is V/H
    res, the Fuji will magically outperform its 6MP competitors by a little.
    The benefit to you is nothing, the cost is a need to generate 12MP worth of
    bandwidth in order to have a picture with the same overall resolution as a c
    onventional 6MP camera.
     
    George Preddy, Feb 3, 2004
    #3
  4. Mattearoadie

    Paul H. Guest

    "Mattearoadie" <> wrote in message
    news:...
    > what does 12.3 milllion recorded pixels mean vs
    > 6.3 million effective pixels?
    >
    > i am looking at the fuji S7000
    >
    > any feedback


    I'd look at some more sample pictures before buying the S7000. From what
    I've seen ( www.steves-digicams.com, www.dcresource.com, et al.) the S7000
    has a much noisier output than many other cameras in its class.

    I'm not trying to pick a fight, by any means: Generally speaking, I like
    Fuji products and my two brothers both own Fuji digital cameras (the 2800
    and 3800) and they are quite happy with them. However, I don't think the
    S7000 is quite up to Fuji's standards.
     
    Paul H., Feb 3, 2004
    #4
  5. Mattearoadie

    chevron Guest

    "George Preddy" <> wrote in message
    news:bvn3rl$m7d$...
    >
    > "Mattearoadie" <> wrote in message
    > news:...
    > > what does 12.3 milllion recorded pixels mean vs
    > > 6.3 million effective pixels?
    > >
    > > i am looking at the fuji S7000

    >
    > It means the camera normally builds 6M full color pixels interpolated from
    > 1.5M Red sensors, 3M Green sensors, and 1.5M Blue sensors. Optionally, it
    > will build a 12M pixel image from the same number of sensors.
    >
    > There is normally no optical benefit whatsoever to interpolation, but

    Fuji's
    > double-interpolation (the 12M pixel output) does enjoy a little bit of
    > optical benefit, because the brick pattern layout of Fuji's SuperCCD will
    > slightly underperform when compared to other 6MP cameras unless it is
    > oversampled. This is because images are ultimately viewed in a

    conventional
    > grid pattern, and some info is lost during the
    > brick-pattern-->conventional-pattern coordinate conversion unless an
    > oversamplimg method is employed.
    >
    > It's a long story but, in a nut shell, Fuji claims they use a brick

    pattern
    > because it slightly increases vertical and horzontal resolution compared

    to
    > a conventional grid pattern. This is true. It is also true (the part

    they
    > "forget" to tell you) that a brick pattern decreases diagonal resolution

    by
    > a corresponding amount, so the net gain in overall resolution is exactly
    > zero.
    >
    > Then why do it? Because vertical and horizontal resolution is all that is
    > usually tested.
    >
    > So what are you left with when you buy into Fuji SuperCCD? A 6MP image

    with
    > slightly lower resoultion that a conventional 6MP camera, or, a 12MP image
    > with the slightly better vertical and horizontal resolution than a 6MP
    > camera, and correspondingly worse diagonal resolution than a 6MP
    > camera--resulting overall resolution: 6MP. If all a reviewer tests is V/H
    > res, the Fuji will magically outperform its 6MP competitors by a little.
    > The benefit to you is nothing, the cost is a need to generate 12MP worth

    of
    > bandwidth in order to have a picture with the same overall resolution as a

    c
    > onventional 6MP camera.
    >
    >




    Don't forget the extremely aggressive Jpeg compression that Fuji insist upon
    using! (one of my main reasons for not considering any Fuji 'prosumer'
    model)
     
    chevron, Feb 3, 2004
    #5
  6. "Paul H." <> wrote in message
    news:...
    >
    > "Mattearoadie" <> wrote in message
    > news:...
    > > what does 12.3 milllion recorded pixels mean vs
    > > 6.3 million effective pixels?
    > >
    > > i am looking at the fuji S7000
    > >
    > > any feedback

    >
    > I'd look at some more sample pictures before buying the S7000. From what
    > I've seen ( www.steves-digicams.com, www.dcresource.com, et al.) the S7000
    > has a much noisier output than many other cameras in its class.
    >
    > I'm not trying to pick a fight, by any means: Generally speaking, I like
    > Fuji products and my two brothers both own Fuji digital cameras (the 2800
    > and 3800) and they are quite happy with them. However, I don't think the
    > S7000 is quite up to Fuji's standards.


    I never said it wasn't a competitive 6MP camera. But it is a shame Fuji
    seems to ruin what their engineers come up with, by relentlesly trying to
    trick people at the store counter. Fuji is a dishonest
    company--deliberately, intentionally, and repeatedly.

    But I do agree that Fuji's true 6MP sensors (not to be confused with their
    3MP sensors which they advertise as 6MP, like the F700) are notably noisier,
    pixel for pixel, than their 3MP sensors were. Same thing happened to Sony,
    btw. Seems that we're finally bumping into technology limits that many have
    been saying are right around the corner. Cramming more pixels into the same
    real estate is starting to degrade image quality as much as it helps.
     
    George Preddy, Feb 3, 2004
    #6
  7. "George Preddy" <> wrote in message
    news:bvnt1s$m5m$...
    > Cramming more pixels into the same
    > real estate is starting to degrade image quality as much as it helps.
    >


    So how many pixels are Foveon cramming into their real estate? 3.43 or 10.3?
     
    Braindead Preddy, Feb 3, 2004
    #7
  8. Mattearoadie

    R2D2 Guest

    "Braindead Preddy" <> wrote in
    news:bvnvso$d8i$:

    >
    > "George Preddy" <> wrote in message
    > news:bvnt1s$m5m$...
    >> Cramming more pixels into the same
    >> real estate is starting to degrade image quality as much as it helps.
    >>

    >
    > So how many pixels are Foveon cramming into their real estate? 3.43 or
    > 10.3?


    According to DP Review, the SD9 gives better resolution than other 3MP
    cameras, but not even close to what a true 10mp would give.
     
    R2D2, Feb 3, 2004
    #8
  9. "R2D2" <> wrote in message
    news:Xns948445B2D78A8R2D2starwarsjedi@130.133.1.4...
    > "Braindead Preddy" <> wrote in
    > news:bvnvso$d8i$:
    >
    > >
    > > "George Preddy" <> wrote in message
    > > news:bvnt1s$m5m$...
    > >> Cramming more pixels into the same
    > >> real estate is starting to degrade image quality as much as it helps.
    > >>

    > >
    > > So how many pixels are Foveon cramming into their real estate? 3.43 or
    > > 10.3?

    >
    > According to DP Review, the SD9 gives better resolution than other 3MP
    > cameras, but not even close to what a true 10mp would give.


    Actually they rate it higher than all 6MP Bayers. But that's for a B&W
    target, which is meaningless. In color tests, the SD9 blows the 1Ds away...

    http://www.outbackphoto.com/artofraw/raw_05/essay.html

    ....but how could it not blow the $8000 Canon away, with 25% more RGB sets?
     
    George Preddy, Feb 3, 2004
    #9
  10. "George Preddy" <> wrote in message
    news:bvo1mh$nr6$...
    >
    > Actually they rate it higher than all 6MP Bayers.


    No they don't. That's a blatant lie.

    The closest they come to that is saying that it matched the D60 in
    resolution, which is a way out of date camera.

    http://www.dpreview.com/reviews/sigmasd9/page24.asp
     
    Darren Sawyer, Feb 3, 2004
    #10
  11. "Darren Sawyer" <> wrote in message
    news:bvo52t$im9$...
    > "George Preddy" <> wrote in message
    > news:bvo1mh$nr6$...
    > >
    > > Actually they rate it higher than all 6MP Bayers.

    >
    > No they don't. That's a blatant lie.
    >
    > The closest they come to that is saying that it matched the D60 in
    > resolution, which is a way out of date camera.


    The 10D uses the same 90's vintage sensor. It's made using '486 fabrication
    technology.

    "Again the SD9's X3 sensor achieving amazing levels of detail and sharpness,
    from this comparison the EOS-D60 doesn't seem to have any resolution
    advantage, indeed in some places the SD9 manages to resolve detail which is
    either blurred or pixel bridged on the D60 image."

    Phil is being nice to his Canon cash cow, his pics show the SD9 has 2-3X the
    resolution as the qualdruply priced per sensor Canon. Look at that boat
    roof, amazing how bad the Canon is by comparision...

    http://www.dpreview.com/reviews/sigmasd9/page18.asp
     
    George Preddy, Feb 3, 2004
    #11
  12. "George Preddy" <> wrote in message
    news:bvo38u$oc4$...
    >
    > The 10D uses the same 90's vintage sensor. It's made using '486

    fabrication
    > technology.
    >


    Again, I think this is quite incorrect. The 10D and D60 have quite different
    sensors. The 10Ds is a significant improvement over the D60, though the
    pixel count is the same.

    > "Again the SD9's X3 sensor achieving amazing levels of detail and

    sharpness,
    > from this comparison the EOS-D60 doesn't seem to have any resolution
    > advantage, indeed in some places the SD9 manages to resolve detail which

    is
    > either blurred or pixel bridged on the D60 image."
    >


    This is just alias artifacts though. You can quite clearly see it in the
    close up crops.


    > Phil is being nice to his Canon cash cow, his pics show the SD9 has 2-3X

    the
    > resolution as the qualdruply priced per sensor Canon. Look at that boat
    > roof, amazing how bad the Canon is by comparision...
    >


    Nice to his cash cow? It's a very objective review, but he does seem to miss
    the fact that alias artifacts aren't image detail. If anything he is being
    kind to the Sigma.

    For you to claim 2-3 time resolution is completely ludricrous, the images
    don't bear that out at all.

    > http://www.dpreview.com/reviews/sigmasd9/page18.asp
    >


    This is a much better review is you think dpreview is biased.

    http://www.imaging-resource.com/PRODS/SD9/SD9A12.HTM
     
    Darren Sawyer, Feb 3, 2004
    #12
  13. Mattearoadie

    R2D2 Guest

    "George Preddy" <> wrote in
    news:bvo38u$oc4$:

    >
    > "Darren Sawyer" <> wrote in message
    > news:bvo52t$im9$...
    >> "George Preddy" <> wrote in message
    >> news:bvo1mh$nr6$...
    >> >
    >> > Actually they rate it higher than all 6MP Bayers.

    >>
    >> No they don't. That's a blatant lie.
    >>
    >> The closest they come to that is saying that it matched the D60 in
    >> resolution, which is a way out of date camera.

    >
    > The 10D uses the same 90's vintage sensor. It's made using '486
    > fabrication technology.
    >
    > "Again the SD9's X3 sensor achieving amazing levels of detail and
    > sharpness, from this comparison the EOS-D60 doesn't seem to have any
    > resolution advantage, indeed in some places the SD9 manages to resolve
    > detail which is either blurred or pixel bridged on the D60 image."
    >
    > Phil is being nice to his Canon cash cow, his pics show the SD9 has
    > 2-3X the resolution as the qualdruply priced per sensor Canon. Look
    > at that boat roof, amazing how bad the Canon is by comparision...
    >
    > http://www.dpreview.com/reviews/sigmasd9/page18.asp


    This comparison is made to the D30, not D60.
     
    R2D2, Feb 3, 2004
    #13
  14. "Darren Sawyer" <> wrote in message
    news:bvo6ep$jq3$...
    > "George Preddy" <> wrote in message
    > news:bvo38u$oc4$...
    > >
    > > The 10D uses the same 90's vintage sensor. It's made using '486

    > fabrication
    > > technology.
    > >

    >
    > Again, I think this is quite incorrect. The 10D and D60 have quite

    different
    > sensors. The 10Ds is a significant improvement over the D60, though the
    > pixel count is the same.


    The 10D moved Canon's fab process up to late 486 standards. The Foveon is
    manufactured using the same fab technology as Pentium 4's, but the Foveon
    has more transistors.

    > > Phil is being nice to his Canon cash cow, his pics show the SD9 has 2-3X

    > the
    > > resolution as the qualdruply priced per sensor Canon. Look at that boat
    > > roof, amazing how bad the Canon is by comparision...
    > >

    >
    > Nice to his cash cow? It's a very objective review, but he does seem to

    miss
    > the fact that alias artifacts aren't image detail. If anything he is being
    > kind to the Sigma.
    >
    > For you to claim 2-3 time resolution is completely ludricrous, the images
    > don't bear that out at all.


    Look at the 6th set from the top and tell me how many roof tines you count?
    http://img.dpreview.com/reviews/SigmaSD9/Samples/Compared/outdoor/outdoor_sd9up_d60.jpg

    In this set, what's the number where the 10D blurs...
    http://www.outbackphoto.com/artofraw/raw_05/essay.html
    (I know the whole 10D image is very blurry, but I mean really really very
    blurry)

    Take a look at the roof detail in the building on the right in the back
    ground in the full image, the Canon is completely blind to the construction
    netting on top, displaying well less than half the resolution of the SD9.
     
    George Preddy, Feb 3, 2004
    #14
  15. "R2D2" <> wrote in message
    news:Xns948453BB8BB2CR2D2starwarsjedi@130.133.1.4...
    > "George Preddy" <> wrote in
    > news:bvo38u$oc4$:


    > > Phil is being nice to his Canon cash cow, his pics show the SD9 has
    > > 2-3X the resolution as the qualdruply priced per sensor Canon. Look
    > > at that boat roof, amazing how bad the Canon is by comparision...
    > >
    > > http://www.dpreview.com/reviews/sigmasd9/page18.asp

    >
    > This comparison is made to the D30, not D60.


    The comparision is to the D60, but the SD9 was moved 50% farther away from
    the boats in these resolution tests, due to Phil's misconception the a crop
    factor is actually optical magnification. Its not, just as digital zoom is
    not optical magnification and provides no resolution benefit whatsoever.
    The test is substantially skewed in favor of the D60 and it gets completely
    clobbered.
     
    George Preddy, Feb 3, 2004
    #15
  16. Mattearoadie

    Dan Pidcock Guest

    R2D2 <> wrote in message news:<Xns948445B2D78A8R2D2starwarsjedi@130.133.1.4>...
    > "Braindead Preddy" <> wrote in
    > news:bvnvso$d8i$:
    > > "George Preddy" <> wrote in message
    > > news:bvnt1s$m5m$...
    > >> Cramming more pixels into the same
    > >> real estate is starting to degrade image quality as much as it helps.

    > >
    > > So how many pixels are Foveon cramming into their real estate? 3.43 or
    > > 10.3?

    >
    > According to DP Review, the SD9 gives better resolution than other 3MP
    > cameras, but not even close to what a true 10mp would give.


    But it's not the numbers that matter. A bayer 10MP might cost twice
    the price of the foveon sensor. And theoretically for the same sized
    sensor it would have worse noise as the pixels are smaller. I wish
    people would stop going on about number of pixels and look at image
    quality instead.

    Dan
     
    Dan Pidcock, Feb 3, 2004
    #16
  17. "Dan Pidcock" <> wrote in message
    news:...

    > But it's not the numbers that matter. A bayer 10MP might cost twice
    > the price of the foveon sensor. And theoretically for the same sized
    > sensor it would have worse noise as the pixels are smaller. I wish
    > people would stop going on about number of pixels and look at image
    > quality instead.


    There is an exact comparison available. The Kodak 14n and the Sigma SD9/10
    both produce the exact same number of recorded output pixels, 13.72MP. And
    both interpolate to the exact same degree to do it, 4 recorded output pixels
    for each RGB set. The SD9 beats the 14n up pretty badly on color resolution
    and image quality, but perhaps the 14n iwasn't the best Bayer design.
     
    George Preddy, Feb 3, 2004
    #17
  18. "R2D2" <> wrote in message
    news:Xns948445B2D78A8R2D2starwarsjedi@130.133.1.4...
    > "Braindead Preddy" <> wrote in
    > news:bvnvso$d8i$:
    >
    > >
    > > "George Preddy" <> wrote in message
    > > news:bvnt1s$m5m$...
    > >> Cramming more pixels into the same
    > >> real estate is starting to degrade image quality as much as it

    helps.
    > >>

    > >
    > > So how many pixels are Foveon cramming into their real estate?

    3.43 or
    > > 10.3?

    >
    > According to DP Review, the SD9 gives better resolution than other

    3MP
    > cameras, but not even close to what a true 10mp would give.


    The review on the site also notes that the SD9 outputs resolution on a
    par with 6 MP D60:

    "Once again the SD9 shows that its images enlarge very well and are a
    match for the EOS-D60, at this stage I think it would be fair to say
    that the SD9 has the capability to resolve at least as much resolution
    as the six megapixel EOS-D60, this is not something I thought I would
    be writing in this review."
     
    Peter A. Stavrakoglou, Feb 4, 2004
    #18
  19. "George Preddy" <> wrote in message
    news:bvp54h$pjd$...
    >
    > "Dan Pidcock" <> wrote in message
    > news:...
    >
    >
    > There is an exact comparison available. The Kodak 14n and the Sigma

    SD9/10
    > both produce the exact same number of recorded output pixels, 13.72MP.

    And
    > both interpolate to the exact same degree to do it, 4 recorded output

    pixels
    > for each RGB set. The SD9 beats the 14n up pretty badly on color

    resolution
    > and image quality, but perhaps the 14n iwasn't the best Bayer design.
    >
    >



    Thats not even close to an exact comparison. The 14N is a 14Mp Bayer.
    The SD9 is a 3.4 MP Foveon.

    If you upscale the Foveon images in this way you will exaggerate the jaggies
    and alias artifacts by a factor of 4. The 14N image will be much better.


    Why did you not respond to my previous post?


    For a rational summary of the SD9, try :

    http://www.imaging-resource.com/PRODS/SD9/SD9A12.HTM
     
    Darren Sawyer, Feb 4, 2004
    #19
  20. "Peter A. Stavrakoglou" <> wrote in message
    news:3fYTb.26387$...
    > "R2D2" <> wrote in message
    > news:Xns948445B2D78A8R2D2starwarsjedi@130.133.1.4...
    > > "Braindead Preddy" <> wrote in
    > > news:bvnvso$d8i$:
    > >
    > > >
    > > > "George Preddy" <> wrote in message
    > > > news:bvnt1s$m5m$...
    > > >> Cramming more pixels into the same
    > > >> real estate is starting to degrade image quality as much as it

    > helps.
    > > >>
    > > >
    > > > So how many pixels are Foveon cramming into their real estate?

    > 3.43 or
    > > > 10.3?

    > >
    > > According to DP Review, the SD9 gives better resolution than other

    > 3MP
    > > cameras, but not even close to what a true 10mp would give.

    >
    > The review on the site also notes that the SD9 outputs resolution on a
    > par with 6 MP D60:
    >
    > "Once again the SD9 shows that its images enlarge very well and are a
    > match for the EOS-D60, at this stage I think it would be fair to say
    > that the SD9 has the capability to resolve at least as much resolution
    > as the six megapixel EOS-D60, this is not something I thought I would
    > be writing in this review."


    "At least "as much. Phil is being tactful, the pics clearly speak for
    themselves and if he makes one wrong move Canon takes their ball and goes
    home.

    Also, one has to be careful to specifically talk color or B&W resolution.
    In B&W, a 6MP Bayer is about as good as 10.3MP Foveon, because the Foveon,
    being a full color sensor, always combines 3 black sensors to make black.
    1/3rd spectrum Bayer sensors can indicate Black by themselves. In fact,
    full Black (no light at all) is the only time a Bayer enjoys true 6MP
    resolution. But in color, the 10.3MP Foveon has a bigger advantage than
    even the 10.3/6 ratio, because Bayer throws aways 25% of its color
    capability for cheap scalability. IOWs, instead of making a custom pattern
    for every camera with 33/33/33% RGB, it works with simple RGGB 2x2 scalable
    units. You have to double a primary to build a cheap 2 dimensional sensor
    with holes for 3 primary color pegs. 25% loss of color resolution is the
    "price" you pay for "cheap" stuff.

    Bottom line...

    B&W resolution - 10.3MP Foveon = (roughly) 7MP Bayer
    Color resolution - 10.3MP Foveon = (roughly) 14MP Bayer**

    ** Note: The current 14MP Bayer is full frame, this dramatically lowers
    resolving power in exchange for FOV. You can concentrate all the sensors on
    the same FOV using a stronger zoom, but then you no longer can enjoy any
    benefit from having a (way expensive) full frame DSLR vs a cropper.
    Conversely, you can dillute the SD9's much higher sensor density with a
    wider lens, and enjoy the same wide FOV as the full frame 14MP Bayer, at the
    same low resolution. Take your pick.
     
    George Preddy, Feb 4, 2004
    #20
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. MBlaster

    Fuji S7000

    MBlaster, Aug 1, 2003, in forum: Digital Photography
    Replies:
    13
    Views:
    791
    RDKirk
    Aug 9, 2003
  2. Opentoe

    Fuji Finepix S7000?

    Opentoe, Aug 9, 2003, in forum: Digital Photography
    Replies:
    2
    Views:
    479
    Steve Michelson
    Aug 10, 2003
  3. Opentoe

    Fuji Finepix S7000

    Opentoe, Aug 20, 2003, in forum: Digital Photography
    Replies:
    1
    Views:
    359
    Trevor S
    Aug 20, 2003
  4. Willie C

    Fuji Finepix S602, S5000, S7000

    Willie C, Aug 30, 2003, in forum: Digital Photography
    Replies:
    4
    Views:
    484
  5. Paul D

    Fuji S5500 v Fuji S7000

    Paul D, Jul 2, 2005, in forum: Digital Photography
    Replies:
    6
    Views:
    528
    Stewy
    Jul 5, 2005
Loading...

Share This Page