Sigma/Foveon change their tune (great technical article)

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

  1. From:


    Addendum of October, 2003

    Since the above "technote" was written in the fall of 2002, in reaction to
    customers wondering why Foveon and Sigma did not clearly state that the SD9
    was either a 3.4 megapixel or a 10.2 megapixel camera, we have had more
    experience with the proposed terminology in the marketplace. Many catalogs
    and reviews have not been able to accomodate the proposed terminology
    changes, and needed to put a single number into a megapixel slot;
    unfortunately they sometimes chose the 3.4 MP number. This number is very
    misleading, as it suggests that the SD9 is in some sense in the same
    category as 3 to 4 MP cameras, when in the fact the SD9 is delivering image
    resolution and sharpness that is outstanding in the DSLR category of 6 to 14

    In response to this misleading information in the marketplace, Sigma and
    Foveon now agree and insist that if only a single megapixel number can be
    used, then the 10.2 MP number, based on the number of photodetectors, is the
    only possibility. It is an objective count of the same kind of detector
    elements as are usually counted as megapixels. It is incomplete in that it
    does not fully represent the novel organization of pixel sensors into stacks
    of three, which allows image capture free of color artifacts and allows all
    the sharpness to fit naturally into a smaller output file. When more
    information can be used, a notation such as "10.2 MP (3.4 MP Red + 3.4 MP
    Green + 3.4 MP Blue)" is appropriate.
    George Preddy, Oct 27, 2003
    1. Advertisements

  2. George Preddy

    Todd Walker Guest

    Great! So glad this came from an UNBIASED source.

    Todd Walker
    Canon 10D
    Todd Walker, Oct 27, 2003
    1. Advertisements

  3. George Preddy

    Tony Spadaro Guest

    In other words they have decided to lie because no one is going to buy the
    camera unless they do. Very smart move which will really help them get good
    customer relations. Friends don't let friends buy Sigma.
    Tony Spadaro, Oct 27, 2003
  4. George - you'll find this of great interest I think ... ... the Foveon sensor
    now has micro-lenses and has lost its *false* aliasing sharpness
    characteristic as a result. Still, the cam' and its software look to be well
    improved IMHO

    Simon Stanmore, Oct 27, 2003
  5. George Preddy

    Alfred Molon Guest

    Alfred Molon, Oct 27, 2003
  6. George Preddy

    Alan Browne Guest

    This is not a "great technical article" it is manifestation of Sigma
    realizing that they are not selling the SD-9, the Foveon chip and Sigma
    lenses and accessories.
    Alan Browne, Oct 27, 2003
  7. George Preddy

    jriegle Guest

    They can say what they want. It all gets sorted out in the camera tests.

    The Foveon sensor had some potential. The 3.43 mp sensor was slightly better
    than the 6mp Canon D-SLR in the side-by-side test on dpreview except that
    the Foveon was free of the artifacts found in the bayer sensor cameras. This
    is considering the Foveon image was resampled to a 6mp image! It just shows
    all the processing that must be done to bayer sensors thanks to the spatial
    relationship of each color sensing site on the CCD/CMOS chip. It really made
    the bayer sensors look ugly.

    I'm still going to buy the Canon. I'm not putting my eggs in the Sigma
    basket. I hope Foveon finds its way out of Sigma.
    jriegle, Oct 27, 2003
  8. Sounds like Sigma marketing hype to me. When the output file is
    2268x1512, that ain't 10 megapixels.

    Of course, it might be tough to sell a 3.4 megapixel DSLR these days, so
    the marketing department has to switch into high gear. I sure hope they
    didn't threaten those who don't parrot their hype.

    Rational folks realize that an unbiased source should be used, not the
    camera maker. Let's look at what the reviewers are getting in terms of
    resolution. From Imaging Resource:

    Sigma SD10:

    "As before, the 3.4-megapixel 3-color-per-pixel Foveon sensor in the SD10
    produces overall resolution approaching that of a 6-megapixel sensor
    using a conventional color filter array pattern, with "strong detail"
    present in the laboratory resolution test image out to about 1,050 lines
    vertically, and 1,200 lines horizontally."

    Canon 10D:

    "I found "strong detail" out to at least 1,400 lines horizontally and
    1200 vertically, although there was still meaningful detail beyond that
    point. "Extinction" of the target patterns didn't occur until about 1,550
    lines. "

    Sounds like Sigma can demand that folks call it a 10MP, but the
    resolution falls short of their 6MP competition.

    If it's a 10MP camera, it's a really, really bad one.
    Albert Nurick, Oct 27, 2003
  9. George Preddy

    Darrell Guest

    Too bad the micro-lenses are Sigma's ;)
    Darrell, Oct 27, 2003

  10. Lionel

    1) I posted a link to a review and that's where to find the research.
    2) On Usenet one man's tripe is another man's truffle: Even asking kindly
    will never change that

    Simon Stanmore, Oct 27, 2003
  11. George Preddy

    Chris Brown Guest

    He would seem to have a point. This chart:

    Shows quite clearly that the SD10 has significantly less pixel-level
    aliasing than the SD9 (although it hasn't been eliminated). This is, of
    course, at the expense of apparent sharpness, simply because the apparent
    sharpness of SD9 images was, in part, due to aliasing. I guess the large
    light gathering area for each pixel is responsible for eliminating some of
    this high-frequency artifacting.
    Chris Brown, Oct 28, 2003
  12. I have much less animosity toward the Foveon sensor and the SD9 than many
    people here, but that could change fast with BS like this.

    A pixel is one dot in a two-dimensional array of dots. You can't have
    stacked pixels. I don't believe the Foveon sensor is competitive with 10-14
    megapixel Bayer sensors, and, if Sigma and Foveon try to make this BS
    claim, I think it will backfire on them much like Fuji's attempt to call
    their 3 megapixel "SuperCCDs" "6 Megapixel Sensors" backfired.

    Here's a suggestion for a more truthful claim - "The Foveon sensor has 3.4
    real megapixels which are equal to 5 Bayer megapixels". This claim would be
    in-line with the reviews I've read from reputable sources like
    Tony Whitaker, Oct 28, 2003
  13. What a joke.

    George ... if you've got the time: care to tell
    how you came to be a member of this cargo cult ?

    Stanley Krute, Oct 28, 2003
  14. That is exactly right, they are finally going to start lying and start
    listing 1/3rd pixels as full pixels. In a way, it is a sad day for the
    truth, now there are no honest manufactures.
    George Preddy, Oct 28, 2003
  15. It looks like a nice update, the price will be interesting. As will be SD-9

    It looks like they've fixed the high ISO workaround, the SD-9 could always
    take nice ISO 1600 shots, but you have to underexpose and push in RAW, the
    explicit higher ISO setting don't work nearly well in low light for some
    reason. The SPP Fill Light looks like a very convenient way to increase
    dynamic range using automated exposure blending (using RAW adjusted exposure
    baselines), again it is doable with the SD-9 too given SPPs great exposure
    control, but its a good deal of work per image replaced with a quick slider.
    Great new feature. But SPP was always, and still is at v2.0, a great Sigma
    strength for those who are interested in shooting RAW.
    George Preddy, Oct 28, 2003
  16. But at least the camera is capable image their sharpness. ;) ;)
    George Preddy, Oct 28, 2003
  17. It'll be interesting to see how it's received. I suspect most Sigma users
    would rather use the SD-9/10's normal 13.7MP output when more smoothing is
    desired, rather than have it done by default, like for poster size and up
    George Preddy, Oct 28, 2003
  18. Didn't you just finish saying the lack of an SD-9 successor proved that?
    George Preddy, Oct 28, 2003
  19. Right, just as the article points out, it is about a 6MP Bayer for B&W
    targets, and about a 14MP Bayer equivent for a full spectrum color targets.
    So their new 10.2MP rating seems sufficiently conservative.
    George Preddy, Oct 28, 2003
  20. Correct, the output file is 13.7MP.
    George Preddy, Oct 28, 2003
    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.