Dpreview's 10.2MP SD10 image vs. a new 8MP Bayer image

Discussion in 'Digital Photography' started by George Preddy, May 3, 2004.

  1. George Preddy, May 3, 2004
    1. Advertisements

  2. George Preddy

    Alfred Molon Guest

    Alfred Molon, May 3, 2004
    1. Advertisements

  3. George Preddy

    Skip M Guest

    George, what are you trying to say here? That the teeny sensor of the 8mp
    Nikon CoolPix is better than the subAPS sized sensor of the Sigma? Lord,
    look at the zigzag hair in the Sigma image.
    Skip M, May 3, 2004
  4. http://img2.dpreview.com/gallery/nikoncp8700_samples/originals/040331-1313-42.jpg

    How typical for Steve/George to compare a mediocre underexposed JPEG image
    to a better exposed Raw conversion. Try this for an 8MP image:
    After proper sharpening (yet another concept alien to 'Preddy') that image
    really shows its potential.

    http://forums.dpreview.com/forums/read.asp?forum=1032&message=8538085 has an
    interesting thread about the 1D Mark 2.

    Bart van der Wolf, May 3, 2004
  5. George Preddy

    ~ Darrell ~ Guest

    ~ Darrell ~, May 3, 2004
  6. George Preddy

    Larry Guest

    I really WOULD like to know what his point is??????

    He has compared two different subjects taken at different ISO settings with
    two different focal lengnths and two different lighting conditions....

    Thats not apple and oranges.... its more like chainsaws and violins.

    (just in case George is reading this I will explain chainsaws and violins:
    They both make noise, and thats all they have in common in this context)
    Larry, May 3, 2004
  7. George Preddy

    paleryder Guest

    Except with the Sigma slr's, which also make noise.
    paleryder, May 3, 2004
  8. Blurry with a blown out pale white sky, must be from a Canon.
    "Potential" is an interesting word choice. I agree, the image isn't
    ready for prime time. Who knows if Canon will ever perfect a full
    color sensor, if not, they'll remain out of the running for good.
    George Preddy, May 4, 2004
  9. (George Preddy) wrote in
    Yep. Sigma's sure showing 'em a thing or two, and obviously the customers
    are abandoning Canon in droves for Sigma. <snort>

    George, you're so dumb it's amazing.
    Albert Nurick, May 4, 2004
  10. Thanks for proving my point. You are clueless, also about sharpening and

    Bart van der Wolf, May 4, 2004
  11. George Preddy

    Chris Cox Guest

    Chris Cox, May 4, 2004

  12. Preddy doesn't half spark off some verbal fart-lighting. For anyone who
    is vaguely bothered about facts, I was shooting yesterday with a Minolta
    A2 and Sigma SD10 side by side, because I need to send files resized up
    to 48 megabytes minimum to the picture library I use. I'm travelling to
    Turkey next week and concerned to take the right kit.

    The A2 was used with noise reduction off, at 64 ISO, at full res, with
    no sharpening, reading RAW files using the new Photoshop CS plugin which
    offers vignetting and chromatic aberration correction. The SD10 images
    were processed through PhotoPro2 but for comparison, some were also
    processed using the PS plugin with chromatic aberration from the rather
    poor Sigma 18-50mm removed. In all cases sharpness, luminance and colour
    smoothing were very carefully adjusted to get the best possible final
    image from each camera.

    After editing and resizing all the duplicated shots - it was a superb
    day and they were good, sharp sunny views - I showed some at 100 per
    cent view on screen to my wife, who was not aware which was which.

    The Sigma wins. It shows more detail. Shirley felt the colour in the
    Minolta example we looked at most closely was better, at first, but then
    we studied an area of trees, and the Sigma showed much greater
    differentiation between many different leaf green shades, while the
    Minolta made them all look rather similar. We studied twigs on bare
    branches at pixel level, and no matter what is said here, the 'snap to
    pixel' effect was no great on the Sigma than on the Minolta. The Sigma
    fell down badly with some narrow white window-frames which showed an odd
    colour aliasing (not expected at all from this camera) but the CS
    plug-in colour noise reduction function totally eliminated that leaving
    only a slight evidence of phasing of tone.

    However, neither camera really cuts it when blown up to 48 megabyte
    files, in my view. The picture library seems to find the enlarged files
    not only acceptable but more desirable than nice clean correctly sized
    ones, and since it's a condition of their terms that we resize all
    digital camera images this way, we'll take the money and ignore common

    It's hard to guess whether a Nikon or Sony 8 megapixel consumer cam
    would clearly better the Sigma's over-enlarged 3.4 megapixel file, where
    the A2 fails, but these A2 images are in their own right not bad - far
    better than my first efforts with the camera, and better than any
    examples I've seen on net test sites. That is down to rigid adherence to
    RAW format, 64 ISO, no noise reduction - etc.

    Photoshop's CS plugin does allow Sigma XF3 files to be processed to
    approximately 8 megapixels (a little larger, but not the 2X scale given
    by the PhotoPro app). When these are compared with unrescaled A2 images,
    the situation is a little better for the A2, but you would still not
    really be able to say it was any better than the Sigma and on many
    subjects the SD10 image is far better. A good example was a dramtically
    yellow field of oilseed rape overlooking the North Sea - the Sigma shot
    shows a full range of tones in every petal of the flower-heads, the
    Minolta shot is one even saturated yellow mass without any tone or
    texture in the yellow. However, the Minolta yellow looks more like the
    real thing, while the Sigma yellow is a bit beefy - slightly orange
    rather than lemon.

    David Kilpatrick, May 4, 2004
  13. George Preddy

    imbsysop Guest

    wow you must have a hell of a screen to show an A2 picture (3264x2448
    full pix) at 100% on screen ...
    interesting statement comparing a 3,34 Mpx cam to a 8 Mpx ..
    plse elaborate what is won by "blowing up" a pict-file to 48 Mb ? &
    how does one "blow up" a file ? with dynamite ?

    hardly convincing argumentation ..
    imbsysop, May 4, 2004
  14. That is true, Foveon's Pro 10M only natively outputs 13.72MP images
    when interpolated to the same degree as all Bayers.
    George Preddy, May 4, 2004
  15. I never, ever "sharpen" well focused images, so you are right there, I
    know nothing about it. Current Bayer-interpolated cripples all need
    mega artificial sharpening, which is why spending more than $100 on a
    lens for an interpolated Bayer camera is a ludicrous thought.
    George Preddy, May 4, 2004
  16. George Preddy

    Crownfield Guest

    gee, george,
    it was 3.xx
    then it was 10,
    now its 13. its been 13 for months now.
    i am surprised that you have not boosted it to 20 or 30 mp by now.

    its growth rate has slowed a lot lately.
    Crownfield, May 4, 2004
  17. He's confused between sensors and pixels. The Foveon sensor layers three
    sensors per pixel location, but it's still 3.4 Mpixel resolution, and
    resolution is the most important thing. If he read the Foveon web site then
    he'd understand the difference.

    If Foveon could come out with even a 6Mp sensor then they might stand a
    chance of getting into some higher end SLRs. Their not going anywhere with
    Steven M. Scharf, May 4, 2004
  18. If you had a camera to begin with, this statement might make sense.
    And it sure ain't the only thing...

    Jukka-Pekka Suominen, May 4, 2004
  19. I guess that Moore's law doesn't apply for Georgie-boy. Might have
    something to do with the fact that neither does Shannon's sampling
    theorems or many other basic math fundamentals?

    Jukka-Pekka Suominen, May 4, 2004
  20. SNIP
    So much is clear, you are clueless.
    Each pixel is the single average of all detail within it (assuming an area
    sample and not a point sample). Averaging means reducing local contrast. The
    loss of contrast can be restored by enhancing edge (and/or pixel) contrast.
    Sharpening procedures do just that.

    Your SD-9 images get sharpened with Raw conversion. The Raw data itself is
    very low contrast and gets boosted quite a bit by default.

    Bart van der Wolf, May 4, 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.