Is Fuji S3000 3.2m/pixel output, or 6 m/pixel interpolated output?

Discussion in 'Fuji' started by Peter H, Nov 19, 2003.

  1. Peter H

    Ray Fischer Guest

    That's the key fact. They USE 50% green sensors. They don't use 25%
    green sonsors.
    You are a stupid asshole.

    Again and again people explain that the sensors ARE NOT COMBINED and
    still you repeat your lies and deceptions. Instead you cling to your
    ignorance.
     
    Ray Fischer, Nov 22, 2003
    #21
    1. Advertisements

  2. Are you hoping immaturity will change the sensor count?
     
    George Preddy, Nov 23, 2003
    #22
    1. Advertisements

  3. You can't accept that Bayer has 25/50/25% RGB split, but no matter how much
    you try to deny it, it is true.
     
    George Preddy, Nov 23, 2003
    #23
  4. Peter H

    Bill M Guest

    ]
    Never denied it did, but you can't accept it uses all of them, there
    are no 25% greens orphans. Try a reading comprehension class...you'll
    look less the fool.
     
    Bill M, Nov 23, 2003
    #24
  5. Then why has SOny gone to such great lengths to reduce the green orphan
    problem?
     
    George Preddy, Nov 23, 2003
    #25
  6. Peter H

    Bill M Guest

    Sigh...no green orphans, but you'll never get it....what do you think
    happens to Sony's 25R/25B/25G/25E? Which do you think gets orphaned
    there? Wait, I'll give you the right answer...NONE (just like bayer)
     
    Bill M, Nov 23, 2003
    #26
  7. No orphans, now their are GB pairs with no red info, a much better
    situation. It's about the best you can do given the fundamental design
    problem of a 2D arrary and 3 primary colors.
     
    George Preddy, Nov 23, 2003
    #27
  8. Peter H

    Ray Fischer Guest

    Are you hoping that outright lying will make sensors disappear?
     
    Ray Fischer, Nov 23, 2003
    #28
  9. ....

    Have you read even one of the replies where I or someone else has tried to explain
    to you your misconceptions?

    Again...

    There are no "orphaned" sensors. There are no "GB pairs with no red info".
    You still seem to think that a Bayer chip groups 4 sensors into 1 pixel. *IT
    DOESN"T!* Excuse me for shouting, but you don't seem to be hearing this.
    Each and every sensor provides the basis information for exactly 1 pixel.
    Because that information is less accurate than the ideal (and presumably
    less accurate than the information from the 3 sensors used to provide the
    information for a Foveon pixel), additional information is used from all the
    surrounding pixels (presumably at least 8).

    So it is *NOT* that some pixels are orphaned; there are simply more pixels
    based on Green Sensors than other colors.

    Now, as to why Sony chose to go to 4 different color sensors, I can guess.
    Remember that the "3 primary color" model, while very good, isn't ideal.
    There are gaps in the continuous color spectrum. (Someone with a much
    better knowledge of color sampling gave a very coherent explanation of this
    in a thread here not long ago.) A "4 primary color" model has some small
    theoretical advantage. (The quadralateral covers more of the color circle
    than the triangle.)

    Now, if you bother to actually look at the Sony sensor array, you will see
    that every non-red pixel has neighboring red pixels. Hence every pixel
    is built using information from neighboring pixels of the missing primary
    colors. Obviously, the true color resolution of the Sony sensor is no
    higher than the Bayer, roughly 1/2 that of the actual resolution.

    Now, you are certainly welcome to like your Sigma camera better than
    the alternatives, but please be realistic about the actual strengths and
    weakness of the camera.


    --
    Dan (Woj...) dmaster (at) lucent (dot) com

    "These are the days of miracle and wonder / This is the long distance call
    The way the camera follows us in slo-mo / The way we look to us all
    The way we look to a distant constellation / That's dying in a corner of the sky
    These are the days of miracle and wonder / And don't cry baby, don't cry"
     
    Dan Wojciechowski, Nov 26, 2003
    #29
  10. Yes, you are mistaken. Over-using and re-over-using the same red and blue
    data does not provide more optical complimentary matches for the extra green
    sensors. This is an extremely simple concept. Have another look at the raw
    data...
    http://www.pbase.com/image/23420444

    Or are you saying there is no optical downside to capturing more of one
    color in a zero sum game? Maybe you'd like to explain how well the sensor
    would resolve color with 90% green, or 99%? Why not 100%, all orphans?
     
    George Preddy, Nov 26, 2003
    #30
  11. Peter H

    Bill M Guest

    Nice contribution there Mike. I understand how both sensors work, from
    your post in another thread, you're as stupid as George...go learn how
    Bayer works then come back when you're educated enough to say
    something intelligent. Better yet, prove the 25% green orphans smart
    ass! Draw a ascii diagram of a Bayer sensor and mark the orphans! Of
    course, you're a sock puppet, so you'll just do like your mentor and
    babble incoherently...
     
    Bill M, Nov 27, 2003
    #31
  12. Peter H

    Bill M Guest

    '

    And this will somehow prove tthat something that doesn't exist does?
    It doesn't matter what course I take, Bayer sensors won't suddenly
    ophan 25% of their green pixels. I can go read a fairytale, but a
    unicorn won't be on my lawn in the morning...
     
    Bill M, Nov 30, 2003
    #32
  13. [/QUOTE]

    Suggesting that an all-green sensor, minus 1 red pixel and 1 blue pixel, is
    not as good in full color as the ideal 33/33/33% mix is being an idiot now?
    Hardly. Your ideal that more green is always better is silly, a child could
    tell you that, without a course in information theory.

    Why do you think using the 1 blue pixel and the same 1 red sensor in all of
    6M color interpolated output pixels is better than 2M R, 2M B, and 2M G?
    You'll need to anwer the question at some point.
     
    George Preddy, Nov 30, 2003
    #33
  14. A fairy tale won't be as helpful as a course in information theory.
     
    George Preddy, Nov 30, 2003
    #34
  15. Ah, I know the answer to that. To reconstruct the color information, you
    need to have 3 sensors that have at least 3 linearly independent frequency
    responses. Having 2 with magenta filters and one of some other color
    wouldn't work in theory. R, G and B are probably selected because they are
    nicely spred out across the visivle sprectrum. But they are not the only
    possibilities and at least one nikon digital camera uses cyan, magenta,
    green and yellow.
     
    Gherry Bender, Nov 30, 2003
    #35
  16. The Bayer view, here, is that you don't need to sense 3 colors, only one.
    Just green, and the more the better. Obviously it's a silly point of view,
    but that's what they've been told so they repeat it.
     
    George Preddy, Nov 30, 2003
    #36
  17. Peter H

    Bill M Guest

    You 2 are adept at dodging and using smokescreens...sometime try
    proving a point w/real data instead of bullshit. I've been involved in
    the IT field for 20+ years, I don't need any courses to know when
    someone's trying to blow smoke up my ass...

    Now, which one of you two wants to show us on a bayer layout where the
    25% green orphans are (that was the original assertion I was
    challenging)? Or will you continue babbling and dodging...bet you both
    dodge again, because they don't exist...
     
    Bill M, Nov 30, 2003
    #37
  18. Peter H

    Bill M Guest

    Jesus, this has been answered 3 or 4 times...firstly, show me where I
    suggested a all green sensor. Secondly, I'm discussing bayer and your
    assrtions about it, which you fail to address. You're nothing but
    smoke & mirrors George. Third, look around the group, others have
    explained all this to you...I could type my explanation, but you'll
    just shift into some other BS, so I'll save my fingers some wear &
    tear...
     
    Bill M, Nov 30, 2003
    #38
  19. If more green is better than less, you'll have to explain when that
    relationship stops and exactly why. As of now, all you say is that more
    green is always better since luminance is better than color, which means
    that 100% green has to be best.
     
    George Preddy, Dec 4, 2003
    #39
  20. Peter H

    Bill M Guest

    Uhm, possibly because you need red & blue (at least) to make other
    colors? Plus, I never said more green is better, YOU insist on putting
    words in peoples mouths. Your logic arriving at "100% green has to be
    best is flawed", much like your understanding of bayer demosaicing....
     
    Bill M, Dec 4, 2003
    #40
    1. Advertisements

Ask a Question

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

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.