8MP vs. 5MP

Discussion in 'Digital Photography' started by Alfred Molon, Feb 14, 2004.

  1. Alfred Molon

    Alfred Molon Guest

    Since the new 2/3" 8MP CCD has pixels of the same size as a 1/1.8" 5MP
    CCD, noise should be more or less the same, shouldn't it ?

    In other words an 8MP camera with a 2/3" CCD should have the same noise
    level as the current generation of 5MP digital cameras with a 1/1.8"
    CCD.

    Or could it be the case that at 8MP the lenses cannot keep up with the
    increased resolution and as a consequence the camera has to sharpen a
    lot internally, thereby amplifying noise ?
    --

    Alfred Molon
    ------------------------------
    http://groups.yahoo.com/group/Olympus_405060/
    Olympus 5050 resource - http://www.molon.de/5050.html
    Olympus 5060 resource - http://www.molon.de/5060.html
     
    Alfred Molon, Feb 14, 2004
    #1
    1. Advertising

  2. In article <>,
    Alfred Molon <> wrote:

    > Since the new 2/3" 8MP CCD has pixels of the same size as a 1/1.8" 5MP
    > CCD, noise should be more or less the same, shouldn't it ?
    >
    > In other words an 8MP camera with a 2/3" CCD should have the same noise
    > level as the current generation of 5MP digital cameras with a 1/1.8"
    > CCD.
    >
    > Or could it be the case that at 8MP the lenses cannot keep up with the
    > increased resolution and as a consequence the camera has to sharpen a
    > lot internally, thereby amplifying noise ?


    Any of the above can be true, or not. You'll just have to check the
    reviews and sample photos.
     
    Kevin McMurtrie, Feb 15, 2004
    #2
    1. Advertising

  3. Alfred Molon

    Mark Herring Guest

    On Sat, 14 Feb 2004 23:39:22 GMT, Alfred Molon
    <> wrote:

    >Since the new 2/3" 8MP CCD has pixels of the same size as a 1/1.8" 5MP
    >CCD, noise should be more or less the same, shouldn't it ?

    Yes, assuming the same optics and other factors
    >
    >In other words an 8MP camera with a 2/3" CCD should have the same noise
    >level as the current generation of 5MP digital cameras with a 1/1.8"
    >CCD.

    but worse than a 2/3" 5Mp sensor
    >
    >Or could it be the case that at 8MP the lenses cannot keep up with the
    >increased resolution and as a consequence the camera has to sharpen a
    >lot internally, thereby amplifying noise ?

    No---noise is tied to the system gain and the size of the
    pixel----processing does not change the intrinsic signal to noise
    ratio.

    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Feb 15, 2004
    #3
  4. Alfred Molon

    Alfred Molon Guest

    In article <>, says...

    > >Or could it be the case that at 8MP the lenses cannot keep up with the
    > >increased resolution and as a consequence the camera has to sharpen a
    > >lot internally, thereby amplifying noise ?

    > No---noise is tied to the system gain and the size of the
    > pixel----processing does not change the intrinsic signal to noise
    > ratio.


    I would have thought that a not so sharp lens could be compensated by
    additional sharpening (via software) in the camera.
    --

    Alfred Molon
    ------------------------------
    http://groups.yahoo.com/group/Olympus_405060/
    Olympus 5050 resource - http://www.molon.de/5050.html
    Olympus 5060 resource - http://www.molon.de/5060.html
     
    Alfred Molon, Feb 15, 2004
    #4
  5. > I would have thought that a not so sharp lens could be compensated by
    > additional sharpening (via software) in the camera.
    > --
    >
    > Alfred Molon


    I would have thought it better to keep the lens-related sharpening the
    same.

    Then with 8MP you get more samples across the image, making for:

    - a higher spatial frequency noise, and therefore less visible noise
    (measured at the same viewing angle)

    - fewer JPEG artefacts for a given compression ratio, given that there is
    a greater pixel-to-pixel correlation.

    Cheers,
    David
     
    David J Taylor, Feb 15, 2004
    #5
  6. Alfred Molon

    Guest

    In message <>,
    Alfred Molon <> wrote:

    >I would have thought that a not so sharp lens could be compensated by
    >additional sharpening (via software) in the camera.


    If you use sharpening to bring back some of the detail rolled off by the
    AA filter, or the limits of the lens if it is the limiting factor, but
    sharpening an out-of focus or motion-blurred image usually looks
    artificial to me. The worst looking USM you see is usually stuff in the
    1.5 to 10 pixel range; that is the range that I have never found any use
    for. I generally use 0.3 to 0.7, or something big like 25.
    --

    <>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
    John P Sheehy <>
    ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
     
    , Feb 15, 2004
    #6
  7. Alfred Molon

    Mark Herring Guest

    On Sun, 15 Feb 2004 08:58:50 GMT, Alfred Molon
    <> wrote:

    >In article <>, says...
    >
    >> >Or could it be the case that at 8MP the lenses cannot keep up with the
    >> >increased resolution and as a consequence the camera has to sharpen a
    >> >lot internally, thereby amplifying noise ?

    >> No---noise is tied to the system gain and the size of the
    >> pixel----processing does not change the intrinsic signal to noise
    >> ratio.

    >
    >I would have thought that a not so sharp lens could be compensated by
    >additional sharpening (via software) in the camera.

    Correct---but this does not increase noise
    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Feb 15, 2004
    #7
  8. "Alfred Molon" <> wrote in message
    news:...

    > Since the new 2/3" 8MP CCD has pixels of the same size as a 1/1.8" 5MP
    > CCD, noise should be more or less the same, shouldn't it ?


    > In other words an 8MP camera with a 2/3" CCD should have the same noise
    > level as the current generation of 5MP digital cameras with a 1/1.8"
    > CCD.


    Not in the way one normally thinks of it. The total amount of thermal noise
    would be the same, but the signal is divided up among more pixels. The
    noise is therefore averaged over a smaller area for each pixel and has a
    relatively greater effect.

    All other things being equal (which they rarely are), increasing the number
    of pixels will increase both resolution and noise. A similar phenomenon
    occurs in development of black-and-white film: Increasing developer
    dilution often increases sharpness and also increases grain.
     
    Andrew Koenig, Feb 15, 2004
    #8
  9. Alfred Molon

    Alfred Molon Guest

    In article <>,
    says...
    > >> >Or could it be the case that at 8MP the lenses cannot keep up with the
    > >> >increased resolution and as a consequence the camera has to sharpen a
    > >> >lot internally, thereby amplifying noise ?
    > >> No---noise is tied to the system gain and the size of the
    > >> pixel----processing does not change the intrinsic signal to noise
    > >> ratio.

    > >
    > >I would have thought that a not so sharp lens could be compensated by
    > >additional sharpening (via software) in the camera.

    > Correct---but this does not increase noise


    It does - any sharpening increases noise.
    --

    Alfred Molon
    ------------------------------
    http://groups.yahoo.com/group/Olympus_405060/
    Olympus 5050 resource - http://www.molon.de/5050.html
    Olympus 5060 resource - http://www.molon.de/5060.html
     
    Alfred Molon, Feb 15, 2004
    #9
  10. <> wrote in message
    news:...
    > In message <>,
    > Alfred Molon <> wrote:
    >
    > >I would have thought that a not so sharp lens could be compensated by
    > >additional sharpening (via software) in the camera.

    >
    > If you use sharpening to bring back some of the detail rolled off by the
    > AA filter, or the limits of the lens if it is the limiting factor, but
    > sharpening an out-of focus or motion-blurred image usually looks
    > artificial to me.


    Yes, although there are techniques that work better than a simple USM.
    Deconvolution (in Fourier space) will recover some of the original
    sharpness, although there are limits to what can be done.

    > The worst looking USM you see is usually stuff in the
    > 1.5 to 10 pixel range; that is the range that I have never found any use
    > for. I generally use 0.3 to 0.7, or something big like 25.


    Yes those are the ranges I also use most on the final image size. Sharpening
    for print can be a bit stronger to compensate for additional losses.

    Bart
     
    Bart van der Wolf, Feb 15, 2004
    #10
  11. Alfred Molon

    Nimous Guest

    can u please tellm e more about d convolution and Fourier space??
    or links for more information?


    "Bart van der Wolf" <> skrev i melding
    news:402fbd90$0$560$4all.nl...
    ..........
    > Yes, although there are techniques that work better than a simple USM.
    > Deconvolution (in Fourier space) will recover some of the original
    > sharpness, although there are limits to what can be done.

    ..........
     
    Nimous, Feb 15, 2004
    #11
  12. that's not what happens with film. most developed contain a solvent to
    decrease grain by dissolving the grain making it appear finer this can cause
    a reduction in E.I.. Diluting the developer reduces the solvent effect.

    Sheldon Strauss
    www.shel.focalfix.com
     
    Sheldon Strauss, Feb 16, 2004
    #12
  13. "Nimous" <> wrote in message
    news:04TXb.33161$...
    > can u please tellm e more about d convolution and Fourier space??
    > or links for more information?


    In very simple (too simple) words:
    Convolution is a mathematical way of applying a uniform distortion (such as
    a blur) to, in our case, image data.
    Fourier space is a different way of expressing, in our case, image data
    after a reversible decomposition of an image into its sine and cosine
    components. The benefit of doing such a conversion is that some calculations
    are much easier to perform, like the addition/removal of repetitive
    structures, blur or motion. For that to be successful, a lot of precautions
    are necessary.
    Deconvolution is, again simplified, taking out a given amount of distortion,
    and is often used in astronomy for reduction of lens defects (and it is very
    effective there because of the relative simplicity of the
    images -pointsources plus distortion- make it easier to describe the
    distortion) .

    Some software tries to take a lot of experimentation out of the process of
    determining/describing the distortion by automating the removal of that.
    Examples are: http://www.unshake.co.uk/ (with examples at
    http://www.unshake.co.uk/examples/index.html), which is mostly geared
    towards Camera Shake and Motion artifacts, or
    http://www.reindeergraphics.com/tutorial/chap4/fourier12.html and
    http://www.reindeergraphics.com/foveapro/deconvolution.shtml.
    http://meesoft.logicnet.dk/Analyzer/ also offers a very interesting (mostly
    8-bit/channel) possibility to deconvolve.

    And as a general resource for mathematical image processing:
    http://homepages.inf.ed.ac.uk/rbf/HIPR2/index.htm .

    Bart
     
    Bart van der Wolf, Feb 16, 2004
    #13
  14. Alfred Molon

    Mark Herring Guest

    On Sun, 15 Feb 2004 17:38:09 GMT, Alfred Molon
    <> wrote:

    >In article <>,
    > says...
    >> >
    >> >I would have thought that a not so sharp lens could be compensated by
    >> >additional sharpening (via software) in the camera.

    >> Correct---but this does not increase noise

    >
    >It does - any sharpening increases noise.


    I hear the faint sounds of hairs being split.

    Sharpening of an image involves altering the contrast selectively
    with spatial frequency. It does not change the intrinsic noise
    content. What it DOES do is make noise more visible. From the
    viewpoint of the end user, the distinction is maybe academic.
    **************************
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
     
    Mark Herring, Feb 16, 2004
    #14
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. ArtKramr

    5mp Kodak vs 5mp Canon

    ArtKramr, Oct 2, 2003, in forum: Digital Photography
    Replies:
    11
    Views:
    549
    ArtKramr
    Oct 3, 2003
  2. Jason Smith

    Sony 8MP Camera (model DSC-F828)

    Jason Smith, Nov 23, 2003, in forum: Digital Photography
    Replies:
    4
    Views:
    480
    David J. Littleboy
    Nov 24, 2003
  3. VT

    8Mp Digicams?

    VT, Jan 10, 2004, in forum: Digital Photography
    Replies:
    5
    Views:
    433
  4. Turbohat

    8Mp Vs 5Mp

    Turbohat, Jul 8, 2004, in forum: Digital Photography
    Replies:
    76
    Views:
    4,312
    Ray Fischer
    Jul 15, 2004
  5. spasmous2

    Using 8MP camera at 5MP resolution

    spasmous2, Nov 16, 2008, in forum: Digital Photography
    Replies:
    5
    Views:
    2,240
    Paul Furman
    Nov 17, 2008
Loading...

Share This Page