Steve's Digicam's classifies the SD-9 and SD-10 as 10.2MP

Discussion in 'Digital Cameras' started by George Preddy, Oct 27, 2003.

  1. George Preddy

    Mark Herring Guest

    WRONG---See maybe 200 or more posts in recent weeks explaining EXACTLY
    why George is wrong.

    Detailed explanation by private e-mail (but using only material
    already posted here)
    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Nov 12, 2003
    #81
    1. Advertisements

  2. George Preddy

    Guest Guest

    Look George, I can do it too....

    The Sigma cameras only sample 3.4 million spatial sampes. The competitors
    sample 6 million spatial samples.
     
    Guest, Nov 12, 2003
    #82
    1. Advertisements

  3.  
    George Preddy, Nov 12, 2003
    #83
  4. It is purely factual. If you think Bayers capture 6M full color pixels,
    that's exactly where you got duped.
     
    George Preddy, Nov 12, 2003
    #84
  5. George Preddy

    Guest Guest

    The Sigma cameras only sample 3.4 million spatial samples. The competitors
    sample 6 million.
     
    Guest, Nov 12, 2003
    #85
  6. George Preddy

    Mark Herring Guest

    George, old buddy;

    If you wil read the record, you will see that I never claimed 6M
    full-color pixels. Bayer sensors sacrifice color resoltuion in favor
    of luminance resolution---we all know that.

    The point is---in case anyone has forgotten, is that most vendors use
    pixel to mean something that contributes to the LUMINANCE spatial
    resolution**. The Megapixels rating is thus a firly consisten measure
    of spatial resolving power that one would see in a monochrome image.
    The reason Bayer is effective is that we dont NEED the same resoltuion
    in color.

    What is your reason for not responding directly to people's arguments?


    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Nov 12, 2003
    #86
  7. In the interests of accuracy, the Foveon's sensors are *not* stacked
    in the sense that they are all the same size and arranged one directly
    over the other, despite the huge number of illustrations from Foveon
    to the contrary. Take a look at this page at DPReview which shows
    both the technical reality and the marketing "reality":

    http://www.dpreview.com/news/0202/02021102foveonx3tech.asp

    The only "accurate" illustration is the first one, the rest are over
    simplified at best and downright misleading at worst, perhaps even
    deliberately so. Even then, the scale is out of proportion in the
    first illustration and although not shown it is horizontally
    compressesed as well (based on size/resolution), but does provide a
    more accurate idea of the physical sensor's appearance. Specifically,
    it is a series of (I assume square) bull's eyes with red on the
    outside and blue in the middle. I suspect that some might take that
    as evidence for a spatial resolution of 10.2M pixels, in the final
    image, but there you go...

    Andy
     
    Andy Blanchard, Nov 12, 2003
    #87
  8. I've posted on this at least 10X... Foveon doesn't need to "guess luminance
    better" it is known for each and every pixel.
     
    George Preddy, Nov 12, 2003
    #88
  9. Great link. Just to clarify, you don't need to "unstack" to have an RGB
    spatial resolution of 10.3MP, since the color channels sample from mutually
    exclusive populations. IOWs, the sensors don't "block" one another by
    stacking them.
     
    George Preddy, Nov 12, 2003
    #89
  10. Yes, I would. And by the identical standard...

    6.0MP (1.5 R + 3.0 G + 1.5 B) = 1.5 MP

    I've never been into counting 1/3rd pixels as full pixels, I think it's a
    dumb standard for both types of sensors.
     
    George Preddy, Nov 14, 2003
    #90
  11. George Preddy

    Nils Rostedt Guest

    Mike wrote :
    My advice is to consider all aspects of the camera and not only the
    resolution. The consensus wisdom seems to be that the 3.43mp Foveon more or
    less equals a 5 mp Bayer in resolution. But, its noise and color
    reproduction characteristics (especially at higher ISO settings) are
    apparently not as good as the best contemporary 6 mp Bayer cameras. It seems
    that there are tricks such as ISO push that can be used to improve the
    situation somewhat, but they require a fair bit of expertise with the
    camera. Some reviews also report undesirable halo effects in particular
    lighting situations with the Foveon sensor. The Sigma SD9 also has a fairly
    unconventional "sports" viewfinder and only one focusing point. Also, it
    doesn't allow JPEG shooting (only RAW) so getting the final results is a bit
    more complicated, which may be an important issue for the more casual
    photographer.

    If I were in your situation I'd let the Foveon technology mature for one ore
    two new-product cycles (probably 1-2 years) at which point the snags are
    likely to be ironed out. The Nikon D100 seems to be regarded fairly
    favorably, so that may be a wise choice especially if your existing lenses
    are of the higher quality and you want to keep them and stay with Nikon.
    Canon is the obvious market leader with a decent price (the 300D), so it's a
    safe choice. I saw some reports on the new Pentax *ist that seemed
    promising, especially the small size and light weight .

    Just my 2 cents - YMMV.
    /N
     
    Nils Rostedt, Nov 14, 2003
    #91
  12. SNIP
    Larger than a 5x7 gets visibly blurry.

    Bart
     
    Bart van der Wolf, Nov 14, 2003
    #92
  13. http://www.pbase.com/pennychallenge

    The Sony P-72 shown is a 3.2 MP Bayer.
     
    George Preddy, Nov 15, 2003
    #93
  14. This is good advice. The big thing is to separate the terms "resolution" as
    it applies ot output pixels and "resolution" as it applies to optical
    rsolution. The two are actually unrelated, any optical resolution can be
    presented using any number of output pixels (as long as it doesn't get
    truncated by using too few).
    I'd say more like 6MP equivalent, but that's using black and white targets
    to meause. Using a color target (which is all that really matters) the
    Foveon shows its full 10.3MP advantage:
    http://www.outbackphoto.com/artofraw/raw_05/essay.html
    A lot actually, but Mike said he is considering the SD-10, which does that
    automatically for you up to ISO 1600.
    That used to be the conventional view, but recently released SPP2 eliminated
    the problem, so it turned out not to be Foveon charge leaking as many had
    incorrectly assumed, including me.
    This is true, though for high quality prints, which Mike said he wants, RAW
    mode is non-negotiable.
    SPP2 ironed out all of the "typical" SD-9 criticisms of emphsized yellows
    and highlight blooming in areas of overexposure. The SD-10 adds explicit
    high ISOs so you don't have to do the RAW push (and you get a bright LCD
    preivew that way).

    "X3 Fill Light" is something you didn't mention, and it gives SPP yet
    another big advantage, as it effectively boosts dynamic range to previously
    unheard of levels for DSLRs.
     
    George Preddy, Nov 15, 2003
    #94
  15. George Preddy, Nov 15, 2003
    #95
    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.