Putting the SD9 "yellow myth" to bed, a follow along RAW demo

Discussion in 'Digital Photography' started by George Preddy, Dec 24, 2003.

  1. George Preddy

    Mark M Guest

    Good thing you never do that regarding cameras, aye George?
     
    Mark M, Jan 3, 2004
    1. Advertisements

  2. George, I didn't "start out with" that information. I was answering a
    direct question from Mark. I've been posting to this thread for a week,
    and to this newsgroup for many years, and almost never refer to my
    educational background.

    You were not even invited to the conversation, but felt the need to
    butt in anyway. Why did you? Are you feeling insecure about your own
    level of knowledge?

    Dave
     
    Dave Martindale, Jan 3, 2004
    1. Advertisements

  3. I did. The first search result was a page that pointed out that Sigma's
    advertising claims about being a 10 MP camera were just advertising
    hype, and the camera is really just 3.4 MP.

    I looked at a couple of other pages, but they just had comparisons of
    photos - no actual resolution targets. So, at a first glance at least,
    that search showed *no* actual resolution test supporting the idea that
    the Sigma cameras exceed the resolution of a 6 MP Canon (or other
    similar camera).

    If you think there is a web page anywhere that does have a proper test
    that comes to this conclusion, why are you so shy about pointing it
    out?

    George, almost nothing you write is "fact".

    Bayer sensors do *not* measure 3 colours in one spot, which is what is
    being discussed. They do measure 3 colours in different locations, in
    order to produce RGB images, but that neither increases nor decreases
    luminance resolution. Measuring 3 colours in one location, as the
    Foveon sensors do, does not help luminance resolution at all. You claim
    it makes the 3.4 MP Foveon sensor behave like a 10 or 14 MP sensor, but
    that's all in your head. There's nothing in signal processing theory or
    in test results that supports your multiplication by 3 or 4.

    No they are not - the pixel counts are wrong unless you use your strange
    non-standard definition of "pixel". Everyone else counts one spatial
    location as one pixel, whether it is a 1-colour or 3-colour sample.
    Further, there's no justification for redefining it "your way".

    That's not what "spatial resolution to match" means. You're twisting
    what other people write again. More precisely, so there is no possible
    ambiguity:

    The 6M 1-colour sensels in the Bayer sensor give the same spatial luminance
    resolution as a 6 MP B&W sensor. Thus we call it a 6 MP sensor.

    The 3.4M 3-colour sensels in the Foveon sensor give the same spatial
    luminance resolution as a 3.4 MP B&W sensor. Thus we call it a 3.4 MP
    sensor.

    You can say that the Foveon chip has 10.3 M sensors, but they're
    stacked together in groups of 3 at the same image plane location, so
    the image is being measured at only 3.4M locations, and the chip has no
    better luminance resolution than a 3.4 MP B&W sensor. Thus calling it a
    "10.3 MP chip" is pointless. It's a 3.4 MP sensor with a useful
    resolution of at most 1500 lines per picture height (since there are
    only 1500 rows of sensels in total). (And that's a theoretical maximum;
    artifacts start appearing around 1300 l/ph.) In addition, it has aliasing
    artifacts above 1500 l/ph.

    In comparison, the Canon D60/10D/300D sensor cannot possibly be 1.5 MP
    or 2 MP. If it was, the resolution could not exceed 1000 or 1300 l/ph.
    Actual tests show a real resolution of about 1600 l/ph. And it renders
    grey instead of aliasing artifacts for detail too fine to resolve.

    The 11MP 1Ds clearly resolves in excess of 2000 l/ph (the limit of the
    test chart that dpreview uses). With resolution like that, it couldn't
    possibly be equivalent to a 3.7 MP sensor as your math claims. The
    Foveon sensors are clearly not even close in resolution.
    B&W, because that's how resolution tests are normally done, because the
    human eye can see 10 times finer detail in luminance than in colour.

    But you could do the same tests with two colours that differed in
    luminance, and you'd get the same results - the Canon cameras would
    still beat the Foveon. This is representative of real-world colour
    photographic subjects.

    The *only* way to get the Canon Bayer resolution to measure less than
    the Foveon is to use a specially-constructed colour target that
    minimizes luminance difference and so measures colour resolution only.
    The outbackphoto people are open about the fact that this is what they
    did, to demonstrate the effect.

    But the result doesn't apply to general photography, unless you often
    shoot saturated red/blue test targets in the wild. The Sigma cameras
    are good for that particular task.

    Dave
     
    Dave Martindale, Jan 3, 2004
  4. (Dave Martindale) wrote in
    Hey Dave don't sweat it, you are a valued contributer to this NG while
    George is, George is a what, a clown, perhaps that is even too
    generous.
     
    Mike Latondresse, Jan 3, 2004
  5. George Preddy

    Mark Herring Guest

    <<<SCREAM>>>>

    I've kill-filed George, I've marked all the Sigma threads to ignore,
    but it still comes.

    Let's ALL join the movement to not reply to GP posts
    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Jan 3, 2004
  6. It goes without saying that everything said which is "unlinked" to some
    factual reference in this forum is opinion.
     
    George Preddy, Jan 3, 2004
  7. Unfortunately, telling me doesn't increase the 1Ds's relatively low RGGB
    sensor count.
     
    George Preddy, Jan 3, 2004
  8. George Preddy

    imbsysop Guest

    then Georgie boy *you* are just a vague opinion .. LOL !
     
    imbsysop, Jan 3, 2004
  9. George Preddy

    Larry Lynch Guest

    Clowns have a powerfull Union, you could get in trouble
    insulting them that way.
     
    Larry Lynch, Jan 3, 2004
  10. George Preddy

    Bill M Guest

    No such thing as a "RGGB sensor" count either....quit makin' stuff up.
    Each location on a Bayer captures either (R)ed, (G)reen or (B)lue
    light and corresponding luminance...the 1Ds camera has 11MP worth of
    pixels in the image plane and all your twisting and turning won't ever
    change that....
     
    Bill M, Jan 4, 2004
  11. Way too low a full color RGGB sensor count.
     
    George Preddy, Jan 4, 2004
  12. The 2.76MP 1Ds, its advertised as 11MegaOneThirdPixels.
     
    George Preddy, Jan 4, 2004
  13. Stay away from the light! Walk away from the light!!
     
    George Preddy, Jan 4, 2004
  14. George Preddy

    Mark Herring Guest

    Never try to teach a pig to sing---it wastes your time and annoys the
    pig.
    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Jan 4, 2004
  15. George Preddy

    Bill M Guest

    Stop babbling....there is no such thing. Bayers do not have full color
    sensors.
     
    Bill M, Jan 4, 2004
  16. George Preddy

    Jeff Shoaf Guest

    And it evidently needs saying that linking to "factual" information that is
    marketing hype or articles written by someone with a vested interest in one
    technology comparing that technology to another are in no means objective
    and should not be considered to be really factual unless validated by
    several truely objective, impartial, independent, and knowledgeble sources.
     
    Jeff Shoaf, Jan 4, 2004
  17. George Preddy

    Jeff Shoaf Guest

    Per George's own comment, this is his non-factual opinion: "It goes without
    saying that everything said which is "unlinked" to some factual reference
    in this forum is opinion."
     
    Jeff Shoaf, Jan 4, 2004
  18. George Preddy

    Jeff Shoaf Guest

    Per this quote from George himself in another thread, this is George's
    opinion and a non-factual post.George said, "It goes without saying that
    everything said which is "unlinked" to some factual reference in this forum
    is opinion."
     
    Jeff Shoaf, Jan 4, 2004
  19. George Preddy

    Mark M Guest

    Not all expressed opinion requires reference links to be effective and
    convincing.
    There are many posts here which create their own credibility through clear
    common sense and experience which is verified through convincing
    delineations of processes and factors involved.
    George's major problem is that he lacks in too many basics of terminology,
    logic, argumentation, and even the most simple facts concerning even his own
    gear. What undermines his arguments perhaps more than anything else is,
    IMO, his inability to recognize this lack, and his consitent presentation of
    the same flawed arguments which remain un-revised in the face of clear
    correction by others.
     
    Mark M, Jan 4, 2004
  20. While I agree with you, I would take a step back and suggest that
    George's biggest problem is motivation. He needs to "win", have the
    last word, feel he has the best equipment, had made the best choice,
    etc. To feed this obsession he will sacrifice everything, even
    self-respect. Sad, but all too common, behaviour.
     
    The Black Sheep, Jan 5, 2004
    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.