Excellent tests Sony A350

Discussion in 'Sony' started by Focus, Jun 8, 2008.

  1. Focus

    Focus Guest

    Highly Recommended 88/100 by Camera Labs:
    http://www.cameralabs.com/reviews/Sony_Alpha_DSLR_A350/verdict.shtml

    Photoreview rates it a solid 9 out of 10:
    http://www.photoreview.com.au/reviews/digitalslr/sony-dslra350.aspx

    9 out of 10 from Pocket-Lint:
    http://www.pocket-lint.co.uk/reviews/review.phtml/3132/4156/Sony-Alpha-350-digital-camera.phtml

    What Digital Camera rates 90%
    http://www.whatdigitalcamera.com/equipment/review/Sony_DSLR_A350_11717.php?offset=&offset=5

    Popphoto rates IQ (100-800 ISO, 1600-3200: extremely high), color accuracy
    and resolution all excellent:
    http://www.popphoto.com/cameras/5234/camera-test-sony-alpha-350-vital-statistics-page3.html


    Seems DPReview stands alone with it's only "recommended" and 8 points for
    IQ.
     
    Focus, Jun 8, 2008
    #1
    1. Advertisements

  2. Focus

    OldBoy Guest

    What's your problem?
    You allready managed to get the all-time best camera thrice.
    You luky [email protected]@rd :)
     
    OldBoy, Jun 8, 2008
    #2
    1. Advertisements

  3. Focus

    dwight Guest

    Y'know...

    Your Sony may be the world's greatest camera, I have no idea. I don't own
    one, haven't considered buying one. The only cameras I do know about are the
    ones I have. I couldn't even tell you how they compare to your Sony, since
    I've never tried a Sony.

    But you spend an inordinate amount of time defending your decision.

    I have found positive and negative reviews for every purchase I've ever
    made, but the only opinion that counts is mine. When I plunk down my money,
    I've done my research, I'm making a somewhat informed purchase. The product
    will either disappoint, live up to expectations, or excel. I use dpreview as
    a guide, among several other sources, and I don't take any one source as
    gospel.

    I will never offer an opinion on equipment I do not own and have never used.

    But I'll say this - once I've bought something, I never feel the need to
    justify that choice to anyone.

    Just keep showing your pictures. In the end, that's all anyone really cares
    about.

    dwight
    www.tfrog.com
     
    dwight, Jun 8, 2008
    #3
  4. Focus

    Alan Browne Guest

    Focus wrote: <some irrelevant stuff>

    Go shoot some photos and stop worrying about what websites say.

    I'm still waiting for that raw file (or even the full sized JPG will
    do). Pls send via yousendit.com to my e-mail minus the "freelunch".

    However, I'm compelled to say that with a pixel density of 195 px/mm,
    the -350 is bound to be noisy compared to lower density cameras.

    Even the FF from Sony due out later this year at 24.5 Mpix will have a
    very high pix density of about 168 px/mm. It will be hard to keep the
    noise down and corner performance may be compromised. Hopefully they
    add a microlens layer to correct for the angle of light arrival.
     
    Alan Browne, Jun 8, 2008
    #4
  5. Focus

    ransley Guest

    If it is so "Excellent" please explain why dpreview [ who does the
    most comprehensive of all testing ] gives " Image quality" of the Sony
    only an 8, And the Canon 450d 9. Im sure its a very good camera, but
    "Image Quality" is what I want most, don`t you?
     
    ransley, Jun 8, 2008
    #5
  6. Focus

    Dica Photo Guest

    _____________________
    /| /| | |
    ||__|| | Do not feed the |
    / O O\__ | trolls. Thank you. |
    / \ | --Mgt. |
    / \ \|_____________________|
    / _ \ \ ||
    / |\____\ \ ||
    / | | | |\____/ ||
    / \|_|_|/ | _||
    / / \ |____| ||
    / | | | --|
    | | | |____ --|
    * _ | |_|_|_| | \-/
    *-- _--\ _ \ | ||
    / _ \\ | / `
    * / \_ /- | | |
    * ___ c_c_c_C/ \C_c_c_c____________
     
    Dica Photo, Jun 8, 2008
    #6
  7. Focus

    frederick Guest

    Most review sites are "advertorial".
    They don't want to give honest reviews because it would cost them
    advertising revenue.
    If you think the a350 got a rough ride from DPReview - wait for the
    Pentax k20d review. The sample photo crops in the a350 review are shockers.
     
    frederick, Jun 8, 2008
    #7
  8. Focus

    ransley Guest

    Ive seen dpreview put many sonys at the tops, thats why I own 3 of
    them. Consumer reports most always puts sony P&S tops. Sony does make
    good stuff, but you have to wonder about money and ratings. dpreview
    has put other sony dslr at the top, but an 8-9 on image rating on the
    350 vs 450d seems harsch.
     
    ransley, Jun 8, 2008
    #8
  9. Focus

    frederick Guest

    Why?
    It looks to me like the 450d has at least a full stop advantage over the
    a350 in noise, and the comments that they make about soft jpegs with
    poor low contrast detail and awful in-camera NR seem pretty much on the
    mark.
    The Canon 450d has better image quality. "One point" better? I dunno -
    but that seems quite fair.
     
    frederick, Jun 8, 2008
    #9
  10. Focus

    Mr.T Guest

    Not necessarily so. For the same final image pixel count, you can do noise
    averaging.
    So you can have EITHER, similar noise and pixels, OR more pixels and noise.
    Choice is yours.
    (all else being equal of course, which it never is :)

    MrT.
     
    Mr.T, Jun 9, 2008
    #10
  11. You don't really have to average the pixels at all. Just display the
    images at the same size, which of course, is impossible with the coarse
    resolution monitors we use. Reducing pixel count in an image only
    reduces noise when the result is viewed at the smaller size. For any
    given size output, downsampling or binning does not really lose noise; it
    just re-distributes it in the the image frequency spectrum, and is more
    likely to leave it concentrated at the same frequency as important image
    detail. Downsampling and binning, at the image level, generally just
    ruin the image, resolution-wise.
     
    John P Sheehy, Jun 9, 2008
    #11
  12. Focus

    Alan Browne Guest

    As the pixel "pit" aperture goes down (geometrically) as resolution goes
    up, the signal also goes down geometrically v. the pretty much constant
    noise.
    ..
    "Smoothing" the image to reduce the appearance of noise is the same as
    the loss of resolution because of noise presence in the first place.

    The advantages of the higher sampling line f. are lost to filtering.
    Damned physics.

    As these resolutions are also at or beyond the performance of the lenses
    used, moreso.

    It would be interesting to try the -350 with my highest res lenses (100
    f/2.8, 135 f/1.8) but I think I can resist purchasing the body for now
    and wait for the FF body at 24.5 Mpix ... and fatter pixel pitch.
     
    Alan Browne, Jun 9, 2008
    #12
  13. Focus

    Mr.T Guest

    You MUST compare at the same size (screen or print) for both images
    regardless of number of pixels, or else the comparison is meaningless.
    That depends totally on how you process the file. Since noise is more
    apparent in "plain" areas, one can simply process those areas.
    And since you already started with higher resolution anyway, reducing it to
    the same level as the camera you are comparing to, should not " ruin the
    image, resolution-wise", unless you have done something wrong, or as I
    already said "all else is NOT equal".

    MrT.
     
    Mr.T, Jun 10, 2008
    #13
  14. Focus

    Mr.T Guest

    Yes, all else being equal.
     
    Mr.T, Jun 10, 2008
    #14
  15. 40D: 90/100
    450D: 88/100
    40D: not reviewed
    450D: 9/10 (9.5 for image quality, where the A350 only gets 9.0)
    40D: 8/10
    450D: 8/10
    Note: the A350 has NO ENTRY for image quality (I wonder why?),
    and only a half-page statement, not the 6+ pages of real review
    other cameras have.

    40D: 90%
    450D: 89%
    A350: IQ ISO 100-800: barely reaching 6/6
    IQ ISO 1600-3200: lower third of 5/6
    Resolution ISO 100-800: barely reaching 5/6
    Color Accuracy: 5/6

    40D: IQ -ISO 800: as A350
    IQ -ISO HI: as A350
    Resolution: complete 6/6 (touching the right border!)
    Color Accuracy: as A350
    The 40D also has better noise, except for ISO 1600+
    --- where the 350 blurs everything.

    450D: IQ -ISO 1600: barely reaching 6/6
    Resolution: a good 5/6 (near 6/6)
    Color Accuracy: as A350
    Seems that the 450D is as good as teh A350, and the 40D is
    even better.


    Of course, DPReview might care about
    - tiny viewfinder
    - excessive noise and excessive noise reduction at ISO >400
    - very stupid (broken) dark frame substraction
    - bad ergonomics
    - slow to respond
    - being designed for "live view" and pessimized for
    eye-on-viewfinder use --- like it was an oversized compact
    camera with exchangeable lenses instead of an DSLR
    and of course, a rather uninspired JPEG output and a sub-standard
    kit lens --- yes, the detail's in the RAW, it's just not used.

    -Wolfgang
     
    Wolfgang Weisselberg, Jun 10, 2008
    #15
  16. Uhm ... what *is* the point of higher resolution then?

    -Wolfgang
     
    Wolfgang Weisselberg, Jun 10, 2008
    #16
  17. Wolfgang Weisselberg wrote:
    []
    To see the lens defects in greater detail?

    David
     
    David J Taylor, Jun 10, 2008
    #17
  18. Monitor resolution doesn't come into it at all, just declare
    the size to be 800x600 or whatever.
    Smaller than what? Pixels have no inherent size.

    What if I used 6,000,000m x 4,000,000 meters as image size?
    Would that be smaller than 3x2 meters at arms' lenth?

    What if that large image was at ~ 400,000,000m distance?
    (Think: painting the moon)
    Do you claim that resolution available in the pixels can be
    used at any output size?

    If not, downsampling or cropping will happen as you exceed the
    resolution of the output device.
    The signal adds linearly, but random noise, being random, only
    adds at the square root. So a 2x2 binning (on chip) increases
    the signal-to-noise-ratio by a factor of 2 (and only adds read
    noise time once, instead of SQRT(4) times, to the signal).

    Downsampling does the same, except for the read noise reduction.

    So how do you figure increasing the signal-to-noise-ratio
    doesn't reduce noise or shift frequencies?
    That depends a lot on what you need, resolution-wise and what
    you have, noise-wise and bandwidth-wise.

    -Wolfgang
     
    Wolfgang Weisselberg, Jun 10, 2008
    #18
  19. Let it. Its images will not improve, never fear.
    Which is why it defends his camera choice, having been burned for its pics.

    -Wolfgang
     
    Wolfgang Weisselberg, Jun 10, 2008
    #19
  20. Wolfgang Weisselberg wrote:
    []
    Why is downsampling off-chip any different to binning on-chip? Isn't the
    read noise equally significant?

    David
     
    David J Taylor, Jun 10, 2008
    #20
    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.