Nikon D2X review

Discussion in 'Nikon' started by Alan Browne, Jun 1, 2005.

  1. Alan Browne

    eawckyegcy Guest

    Of course you wouldn't, as that would be at odds with the Religion Of
    (Sound of forehead being smacked.) Holy shit! Fitting data to a
    model! I'll notify the authorities in Colorado at once: this Clark
    yahoo needs to be dealt with before he transmits these pernicious,
    odious, concepts into the impressionable youth!
    I guess you missed it, but Clark starts the sentence "Within the
    accuracy of the data". You can re-read it above if you don't believe

    Now can you figure out the rest, or does someone have to hold your
    You flip a coin 6 times and receive 4 heads. Is the coin fair?
    Quantify. Here, I'll help:
    Whatever you say, dude.
    Clark is an investigator at NASA. He designs/builds spacecraft
    instruments and the like. If I recall correctly, some of his gear is
    in orbit around Saturn right now.

    You, on the other hand, appear to be a guy with a Nikon and an
    uninformed opinion. Check it out:
    He didn't bother because it isn't relevant re: "photon noise limit".
    eawckyegcy, Jun 3, 2005
    1. Advertisements

  2. Yes, "It just looks like a curve that is fitted to the data" is one of
    the more amusing statements I've seen in a long time!

    Also look at Clark's treastise " Digital Cameras: Does Pixel Size
    Matter? Factors in Choosing a Digital Camera" at:
    "". It
    explains, in layman's terms, why pixel size is so critical in terms of
    noise (why the D2x isn't suitable for those who need to do high-ISO).

    I don't understand this desperate need to defend the D2x. It's a fine
    camera. As long as you don't need to shoot high-ISO, and don't care
    about the large crop factor, if you're already invested in Nikon lenses,
    then it's the best choice for a professional SLR

    Steven M. Scharf, Jun 3, 2005
    1. Advertisements

  3. Alan Browne

    Tom Scales Guest


    Come on. Cite some reasonable references, or it is OPINION.

    Regardless, we all know your unnatural predilection for Canon
    Tom Scales, Jun 3, 2005
  4. Sorry, I didn't know things were that bad where you live. I always assumed
    that when talking about photon shot noise, you have to know the number
    of photons present.

    But no, any curve that looks like a square root has to be photon shot noise.
    It is so obvious.
    And that is usually where people describe measurement errors, standard
    deviation, error distributions, etc.

    Fortunately, on the web you don't have to.
    Wow, we measure noise by flopping a coin 6 times. Great. If you want to
    measure noise to this level of accuracy, you have to average thousands of
    samples. Otherwise, you can just as well flip a coin.
    Hopefully, his designs for NASA are better thought out than his experiments.
    I know a bogus experiment when I see, I don't need a Nikon for that.
    Except that he never actually measured the number of photons.

    But worse, there is no way to perform a similar experiment with a different
    camera, because the illumination levels are unknown. All we know is the
    output of the A/D converter.

    Anyhow, what do I care about R N Clark's web-pages. If you want to take it
    as gospel, go ahead.
    Philip Homburg, Jun 3, 2005
  5. Alan Browne

    eawckyegcy Guest

    What does this have to do with the fact you are criticizing Clark for
    showing the noise curve of a 1DMkII sensor is completely consistent
    with a shot noise process?

    Oh, right, it has nothing to do with it! You are just raising a
    Propose a counter-explanation, dingbat.
    You are stooooopid beyond words. The fact that it was "above" the
    curve 4 out of 6 times is as meaningful as getting 4 heads out of 6
    coin flips.

    Have you ever been able to abstract?
    You really need to see it twice, so:
    Hey, his stuff is at Saturn and you are just a Nikon Nutcase. HA HA
    I know a total moron when I see one too. Is ignorance truly as
    blissful as they claim it is? Are you really this stooopid in real
    life, or is this an act just for me?
    You don't need to. You see, the curve has the same shape no matter
    what the sensitivity of the detector happens to be.

    "If Chewbacca lived on Endor, we need to know the illumination levels!"

    AAAAAHAHAHAHAHA. What's next, Homburg, will you pull a monkey out of
    your pocket?
    If you didn't care to know that the sensors in modern cameras are
    dominated (to the exclusion of almost all other effects) by poisson
    noise, then why did you ask for as much?
    That you reject physically observed reality is in full accordance to
    all religious doctrine.
    eawckyegcy, Jun 3, 2005
  6. SNIP
    That's correct, however we only know (if we're lucky) the potential
    well depth of a sensor. Given that physical maximum, we can calculate
    the theoretical photon shot noise, which is Sqrt(Mean), for that
    sensor. We do need to understand that not all sensels are created
    equal, so some sensels won't reach the maximum. However, since we're
    sampling (very significant number of samples), the response is
    probably close to the response from the total population.
    That would require a more scientific approach to prove with some
    credibility, which seems to be your main objection against Roger
    Clark's observation, and I would agree. However, it is kind of
    difficult for an individual to meet that criterion with only a single
    camera, but then Roger wouldn't object to getting more test result
    feedback from others. The problem with that is that the exact same
    procedure needs to be followed, which is kind of hard to achieve and
    I wouldn't qualify it as "bogus", especially given Roger's background.
    The "bogus" qualification kind of puts the burdon of (dis)proof on
    your shoulders, it seems.

    It seems that for each photon that gets converted, an electron can be
    counted (+/- readout noise). It's the number of converted photons that
    count for photon shot noise statistics.
    I'm not sure what you're getting at, 'all' we need to know is the
    number of electrons stored in the well, and the saturation level.
    There might be an issue in that we don't know upfront if the ADC makes
    a linear conversion from electrons to a DN (quantified digital
    number), and I would be weary as one approaches the well saturation
    level (beware of anti-blooming measures gradually kicking in), but
    that can be estimated from the average DN in response to exposure.

    Bart van der Wolf, Jun 4, 2005
  7. Alan Browne

    JPS Guest

    In message <3lWne.93073$>,
    Well, the 12-24 is not going to be anything like having an 18-36 on a
    full-frame sensor, because wide angle lenses just aren't all that sharp,
    especially zooms, and with small apertures, and you need a larger
    capture area to really get all the detail. Why do you think that large
    formats are favored by film landscape photographers?
    JPS, Jun 4, 2005
  8. Alan Browne

    JPS Guest

    In message <>,
    They do, but it is a difficult basis for comparing cameras, because
    there are too many variables.
    JPS, Jun 4, 2005
  9. Alan Browne

    JPS Guest

    In message <>,
    This is a newsgroup for discussion of *TECHNOLOGY*. What is common to
    all photography is not what this group is intended for.
    JPS, Jun 4, 2005
  10. Alan Browne

    Roxy d'Urban Guest

    No, they are not inaccurate, just the truth. I found this out by doing
    some work for an international sports picture agency who supplied the
    400mm f/2.8 lens. Do you really believe that a beleaguered sports
    photographer earning the going rate for his images is going to be able to
    afford a couple of 1D's, 400mm, 300mm, 70-200mm, 24-70mm and 16-35mm
    lenses? Sure, some do, but in most cases the newspapers and picture
    agencies provide the gear. They get it from Canon. On lease. Canon gets
    ongoing revenues from its assets in the field and the agencies get to
    write off the whole cost as a tax deduction, whereas if they had bought it
    outright they would have had to enter it onto their asset register and
    could only claim depreciation over 5 years. At the end of the 5 years
    there would be a newer, better lens from Canon and they would be stuck
    with having spent a lot of money on the older technology.
    Nonsense. There is no perceptible difference in image quality between a
    Nikon or Canon pro telephoto lens.
    You are talking through your ass. You seem to be forgetting that the D1
    pre-dates the 1D by a good couple of years. You also seem to be forgetting
    that the D2x is considerably more advanced than the 1DMkII.

    As far as competing with the 20D goes...look no further than the D70.
    Roxy d'Urban, Jun 4, 2005
  11. The problem I have with his noise measurements is that he doesn't just
    say using the sqrt of the DN is good way to model the noise in a
    sensor (which is rather obvious from his data, certainly the low ISO ones).

    Instead he makes claims about the performance of future CCD designs based
    on the assumption that most of the noise is due to photon noise, but he
    doesn't prove that.

    My guess is that if you assume twice or four time as many photons, you
    subtract the noise predicted by that number of photons and you insert
    other sources of noise, then the resulting graph can be made to match
    his experiments. Certainly for high ISO, his model doesn't match the
    maesured data all that well.

    Then there is the lack of additional statistical data. What are the
    confidence intervals for this signal-to-noise figures. He doesn't
    verify that the noise behaves like the independent random variable you

    In the end we have the rather un-interesting result that for high intensity
    (low ISO, close to saturation) the sensor is strongly influenced by photon

    In interesting cases (high ISO) we don't know, and can't predict what the
    performance of future sensors is going to be because all interesting variables
    have been illimenated from the model.

    The sentence 'The measured values (points and lines in color) show that
    within measurement error, the 1D Mark II is performing at the theoretical
    maximum.' suggests that he is talking about physics theory. But he isn't,
    he is talking about his own model where he assumes that there is only
    photon noise.
    Philip Homburg, Jun 4, 2005
  12. Alan Browne

    Tom Scales Guest

    Theory, theory and more theory.

    Too bad the actual facts (read: photographs) are simply stunning. Even
    using it on my 6mp D70 results in 24 x 36 prints that are simply stunning.

    Oh, but wait, that print on my wall must not really exist, because it
    doesn't match the theory!

    Tom Scales, Jun 4, 2005
  13. Alan Browne

    JPS Guest

    In message <A2goe.94331$>,
    The fact that you are pleased with your prints is only an actual fact
    about your feelings. It tells us nothin about how much better a 12MP
    full-frame with an 18mm lens would be compared to a 6 or 12MP 1/1.5x
    crop would be.
    No, it is completely subjective, with nothing to compare to.
    JPS, Jun 4, 2005
  14. Alan Browne

    Tom Scales Guest

    Thank you for making my point. This whole thread is about subjective
    theory, not actual prints.

    Tom Scales, Jun 4, 2005
  15. Alan Browne

    JPS Guest

    In message <Dehoe.94972$>,
    Exactly what have I said that is stupid?

    You're acting like an irrational product-defender.
    JPS, Jun 4, 2005
  16. Alan Browne

    Tom Scales Guest

    No, no, no. Not you -- the whole silly thread. The D2X is a fine camera.
    So is the 1DMkII, the 1DsMkII and the Rebel XT. Heck, for some people, the
    SD9 is a fine camera, just not for me. Heck, I'd love a 1DsMkII, but the
    extra money couldn't be justified.

    This is a typical "my **** is bigger than your thread".
    Tom Scales, Jun 4, 2005
  17. Alan Browne

    Alan Browne Guest

    Everything has limitations. There is a print size where a photo breaks
    down into crud. Better combinations of subject, lighting, lens and
    sensor (not to mention technique) will go furthest.

    That should be clear to most.

    Alan Browne, Jun 4, 2005
  18. This is true. I'll never understand people that purchase a product, and
    then feel compelled to prove to the world that whatever they purchased
    is the absolute best product in the universe for everyone.

    We've seen this same behavior from this individual in the past, so it
    shouldn't be surprising. In 2004, he purchased a Pentax Optio S4i, and
    immediately began massive numbers of posts defending his choice, "darn
    good quality," and "I love mine," and got very upset whenever anyone
    pointed out some of the limitations of this product.
    Steven M. Scharf, Jun 4, 2005
  19. Alan Browne

    JPS Guest

    In message <5Uhoe.66004$>,
    I'd like one; but I don't have any Nikon lenses.
    Well, that one I'd skip. Too much aliasing; too little discrimination
    between color channels. That's not to say that someone couldn't take
    interesting images with it.
    My monthly real estate expenses just went up by $750/month in December,
    and my income has dropped about $25K/yr due to overtime becoming
    history, and my hobbies are hurting for it.
    But isn't all like that. There is a lot of assumption going on, and if
    you're not careful, you start believing A's assumptions about B. There
    are outright bigots involved here, but there are also people challenging
    certain points, and they are assumed to be bigots by defensive people.

    For example, if someone links to an image to demonstrate a lack of
    noise, and it is downsized to 20%, and you bring up the fact that you
    can't properly demonstrate noise in a 20% resize, you often
    automatically become anti-that-camera, in defendee's eyes. You really
    have to sift through this stuff to find any rational though.
    JPS, Jun 4, 2005
  20. Alan Browne

    Tom Scales Guest

    Oh Steven, you're so full of shit.

    You're the one that defends Canon over every objection. I'm not defending
    the D2X, just pointing out that theoretical discussions of what should be
    its limitations are meaningless. It's a fine camera with limitations, just
    as the 1dsMkII has limitations. Just different ones. They both excel in
    different areas.

    Heck, I've challenged you a thousand times and you're too wimpy to even

    As for the s4i, yes, I challenged you assertations of what you, in theory,
    thought of a camera you didn't own.

    Is it perfect, no.
    Is it a little soft, sometimes.
    Is it so darn small its always in my pocket, yes.
    Do I catch photos you miss because you don't have a camera with you. Bet I

    It's amazing how you've changed. Let's face it, anyone that feels the need
    to build a website to tear down the Sigma cameras -- again, cameras you
    don't own, has issues. Serious issues.
    Tom Scales, Jun 4, 2005
    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.