High ISO noise

Discussion in 'Digital Photography' started by David Harper, Oct 17, 2008.

  1. David Harper

    David Harper Guest

    I take a picture with high ISO. Of, course it is a little noisy. My
    question is what kind of noise am I seeing? Is it random noise or something
    else, or a combination of things? Zooming in on a few pixels in a dark area
    I see individual pixels that are obviously wrong - too bright, incorrect
    color, etc. If I took the exact same picture again would these same pixels
    be off in the same way or would different pixels be incorrect this time? Are
    individual pixels in the camera's sensor non-linear in low-light situations?
    Is this non-linearity random or is it predicable for a given sensor?

    - David Harper
    David Harper, Oct 17, 2008
    #1
    1. Advertising

  2. David Harper

    Scott W Guest

    On Oct 17, 1:56 am, "David Harper" <> wrote:
    > I take a picture with high ISO.  Of, course it is a little noisy.  My
    > question is what kind of noise am I seeing?  Is it random noise or something
    > else, or a combination of things?  Zooming in on a few pixels in a dark area
    > I see individual pixels that are obviously wrong - too bright, incorrect
    > color, etc.  If I took the exact same picture again would these same pixels
    > be off in the same way or would different pixels be incorrect this time? Are
    > individual pixels in the camera's sensor non-linear in low-light situations?
    > Is this non-linearity random or is it predicable for a given sensor?


    The noise is a mix of a number of sources and the ratio of these
    sources depends on the camera you are using. Much of the noise is
    random, this is a combination of photon noise + electrical, if you
    take a whole lot of images and average them together you can reduce
    this noise. Then there is fixed pattern noise, which it more visible
    with long exposures, many camera have mode where they take two photos
    with long exposures one with the shutter closed, and then they
    subtract the frame where the shutter was closed from the one where it
    was open.

    Some of the fixed pattern noise is going to change depending on the
    temperature of the sensor.

    Scott
    Scott W, Oct 17, 2008
    #2
    1. Advertising

  3. David Harper wrote:
    > I take a picture with high ISO. Of, course it is a little noisy. My
    > question is what kind of noise am I seeing? Is it random noise or
    > something else, or a combination of things? Zooming in on a few pixels
    > in a dark area I see individual pixels that are obviously wrong - too
    > bright, incorrect color, etc. If I took the exact same picture again
    > would these same pixels be off in the same way or would different pixels
    > be incorrect this time? Are individual pixels in the camera's sensor
    > non-linear in low-light situations? Is this non-linearity random or is
    > it predicable for a given sensor?
    >
    > - David Harper
    >
    >



    Yes.

    Auntie
    Auntie Establishment, Oct 17, 2008
    #3
  4. Mxsmanic wrote:
    > Alan Browne writes:
    >
    >> (Film on the other hand has larger (x,y) dye 'blobs' with higher ISO
    >> as well as the dynamic variance. This is why, generally, one gets
    >> much better high ISO results with digital than film).

    >
    > My digital results at high ISO are no better than those of film.


    My ISO 1600 results with a DSLR are much better than those of ~ISO 400
    film.

    David
    David J Taylor, Oct 18, 2008
    #4
  5. Mxsmanic wrote:
    > David J Taylor writes:
    >
    >> My ISO 1600 results with a DSLR are much better than those of ~ISO
    >> 400 film.

    >
    > Sure they are.


    Yes, they are - here's my example, straight out of the camera:

    http://www.satsignal.eu/2008-10-04-1432-04.jpg

    David
    David J Taylor, Oct 18, 2008
    #5
  6. David Harper

    nospam Guest

    In article <>, Mxsmanic
    <> wrote:

    > > (Film on the other hand has larger (x,y) dye 'blobs' with higher ISO as
    > > well as the dynamic variance. This is why, generally, one gets much
    > > better high ISO results with digital than film).

    >
    > My digital results at high ISO are no better than those of film.


    then you're doing something very wrong.
    nospam, Oct 18, 2008
    #6
  7. Alan Browne wrote:
    > David J Taylor wrote:

    []
    >> http://www.satsignal.eu/2008-10-04-1432-04.jpg

    >
    > That has an undersaturated (overexposed) look to me.
    >
    > Did you use a tripod? If so, then why not ISO 100 or 200?
    >
    > Cheers,
    > Alan.


    No, purely hand held - a grab shot almost. One of several hundred shots
    taken during a visit to Bletchley Park in the UK.

    Cheers,
    David
    David J Taylor, Oct 18, 2008
    #7
  8. Mxsmanic wrote:
    > David J Taylor writes:
    >
    >> Yes, they are - here's my example, straight out of the camera:
    >>
    >> http://www.satsignal.eu/2008-10-04-1432-04.jpg

    >
    > Looks like a shot out of a drugstore disposable film camera.
    >
    > Where are your scanned film images for comparison?


    Haven't done film for over ten years, and I no longer have an interface
    for my scanner. You can provide those.

    David
    David J Taylor, Oct 19, 2008
    #8
  9. David Harper

    J. Clarke Guest

    David J Taylor wrote:
    > Mxsmanic wrote:
    >> David J Taylor writes:
    >>
    >>> My ISO 1600 results with a DSLR are much better than those of ~ISO
    >>> 400 film.

    >>
    >> Sure they are.

    >
    > Yes, they are - here's my example, straight out of the camera:
    >
    > http://www.satsignal.eu/2008-10-04-1432-04.jpg
    >
    > David


    I'm not seeing mxsmanic's posts, but I would invite him to show us a
    film sequence comparable to the one at
    http://www.flickr.com/photos/39383723@N00/sets/72157607585960643/.
    Yeah, at 400,000 the image is mostly noise but you can still tell what
    time it is by the wall clock.

    I shot that one while dinking around with the camera and Elements one
    night--just got curious as to how far I could go and still see
    anything in the image.

    For those who would critique focus or composition, it's not intended
    as art.

    --
    --
    --John
    to email, dial "usenet" and validate
    (was jclarke at eye bee em dot net)
    J. Clarke, Oct 19, 2008
    #9
  10. Mxsmanic <> wrote:
    > nospam writes:


    >> then you're doing something very wrong.


    > No, camera manufacturers are concentrating on pixel counts to the detriment of
    > noise reduction.


    I'm old enough to have struggled a lot in a past life with the awkward
    compromises between grain and resolution in fast colour film. I also
    have a Sony A350, often cited as a notorious example of Sony's lunacy
    in pushing for high pixel counts at the cost of dreadful high ISO
    noise. So I was very surprised to find that even at ISO 3200 the
    careful use of good third party noise reduction software and
    downsizing to a 7MP image produces much cleaner sharper results than I
    ever got from ISO 400 colour film.

    If someone offered me a camera with half the high ISO noise of my A350
    but also half the low ISO resolution I wouldn't want it. In fact if
    someone offered me one with twice the high ISO noise and twice the low
    ISO resolution of my A350 (plus lenses to match the resolution :) I'd
    jump at it!

    So from my own personal point of view I don't agree that even the
    worst offenders in the poor noise but high pixel count camera makers,
    such as Sony, are going too far. For my taste they haven't yet gone
    far enough.

    But of course that may be because decades ago I was psychologically
    traumatised by the poor performance of high ISO colour film :)

    --
    Chris Malcolm
    Chris Malcolm, Oct 19, 2008
    #10
  11. David Harper

    DavidM Guest

    David Harper wrote:
    > I take a picture with high ISO. Of, course it is a little noisy. My
    > question is what kind of noise am I seeing? Is it random noise or
    > something else, or a combination of things? Zooming in on a few pixels
    > in a dark area I see individual pixels that are obviously wrong - too
    > bright, incorrect color, etc. If I took the exact same picture again
    > would these same pixels be off in the same way or would different pixels
    > be incorrect this time? Are individual pixels in the camera's sensor
    > non-linear in low-light situations? Is this non-linearity random or is
    > it predicable for a given sensor?


    The noise, while random, is characteristic of your camera or scanner.
    The CCD sensor will have all kinds of thresholds, sensitivities and
    software parameters that are individual to that model.
    I've used a package called Neat Image with very good results
    (http://www.neatimage.com/). They have a free version that you can try.

    Once an image is loaded the software will box off an area of low detail,
    then use that to build a profile for the CCD. You can also change it's
    position in order to achieve a better match. My advice is to aim for a
    noise characterisation level of 90%+. Well worth experimenting with, the
    cleaned up images can look great.

    DavidM
    DavidM, Oct 19, 2008
    #11
  12. J. Clarke wrote:
    []
    > I'm not seeing mxsmanic's posts, but I would invite him to show us a
    > film sequence comparable to the one at
    > http://www.flickr.com/photos/39383723@N00/sets/72157607585960643/.
    > Yeah, at 400,000 the image is mostly noise but you can still tell what
    > time it is by the wall clock.
    >
    > I shot that one while dinking around with the camera and Elements one
    > night--just got curious as to how far I could go and still see
    > anything in the image.
    >
    > For those who would critique focus or composition, it's not intended
    > as art.


    Thanks for posting that sequence, John. It shows me that ISO 1600 is just
    the start of the "low-noise, high-ISO" zone.

    David
    David J Taylor, Oct 19, 2008
    #12
  13. David Harper

    Viperdoc Guest

    For those who don't know, Anthony Atkielski is a well known troll who posts
    thousands of messages. He writes like an expert (see some rec.aviation posts
    for example), but the fact is that he has little or no actual experience.

    His modus is to post inflammatory or contradictory messages that contain
    half truths, in order to get some pathetic attention for himself.

    He single handedly has ruined several fine newsgroups.

    If you look back, less than a month ago he was asking about batteries for
    his camera, and today he writes like the world's expert on digital cameras.
    Viperdoc, Oct 19, 2008
    #13
  14. David Harper

    Steve Guest

    On Sun, 19 Oct 2008 15:16:07 +0200, Mxsmanic <>
    wrote:

    >Floyd L. Davidson writes:
    >
    >> Indeed, but not one with ISO 1600 film!

    >
    >The disposables are usually loaded with fast, wide-latitude film that can
    >produce a usable image in all sorts of situations. Automatic compensation at
    >the lab allows just about every frame to come out.
    >
    >> Shouldn't that be where are *your* film examples?

    >
    >http://www.mxsmanic.com/VelviaScan.jpg
    >
    >Shot on Velvia at ISO 50, handheld, ambient light on an overcast day, 90 mm
    >lens. About 1/9 of original size.
    >
    >http://www.mxsmanic.com/artists.jpg
    >
    >Shot on Portra BW if I remember correctly, handheld, nighttime, 50 mm lens,
    >again, about 1/9 of original size.
    >
    >There are many more.


    Where's the film color shot pushed to ISO 1600?

    Steve
    Steve, Oct 19, 2008
    #14
  15. David Harper

    Viperdoc Guest

    Anthony, still a master at stating the obvious while trying to pass off as
    an "expert"?
    Viperdoc, Oct 19, 2008
    #15
  16. David Harper

    Viperdoc Guest

    Anthony will never admit that he's wrong, or that there is any alternative
    to his narrow, yet inexperienced, opinion.
    Viperdoc, Oct 19, 2008
    #16
  17. David Harper

    Viperdoc Guest


    >
    > So you're comparing what you see today to ten-year-old memories.


    Anthony, at least he has some experience. How long have you been shooting
    digital? What kind of gear and software?
    Viperdoc, Oct 19, 2008
    #17
  18. David Harper

    Viperdoc Guest

    Anthony will never respond to your challenge, and instead will deflect the
    issue, since he ultimately knows he is wrong and unable to respond.
    Viperdoc, Oct 19, 2008
    #18
  19. Mxsmanic wrote:
    > David J Taylor writes:
    >
    >> Haven't done film for over ten years, and I no longer have an
    >> interface for my scanner.

    >
    > So you're comparing what you see today to ten-year-old memories.


    I probably haven't /taken/ prints for 40 years. Doesn't mean I haven't
    seen other people's results. Still waiting for your ISO 400 negative scan
    to compare with what I posted...

    David
    David J Taylor, Oct 19, 2008
    #19
  20. Alan Browne wrote:
    > David J Taylor wrote:
    >
    >> No, purely hand held - a grab shot almost. One of several hundred
    >> shots taken during a visit to Bletchley Park in the UK.

    >
    > With some effort, I get away with tripods when I visit places (though
    > most museums won't allow it...). I somehow often manage to forget the
    > monopod at home...
    >
    > http://photo.net/photodb/photo?photo_id=8044331&size=lg ISO 25


    Sometimes, Alan, I wish I had that discipline as well but, to be honest,
    today I carry the minimum. I'm not a professional photographer - it's
    just a hobby for me. My tripods are exclusively for home use.

    My wife has a walking stick which can double as a monopod, but she said it
    was more bother than it was worth. "At the times when it would have
    helped my photography, it was actually more use to keep me upright. Well,
    it /was/ in windy Antarctica!".

    Cheers,
    David
    David J Taylor, Oct 19, 2008
    #20
    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. just44

    kodak 14n noise at high iso

    just44, Mar 5, 2004, in forum: Digital Photography
    Replies:
    2
    Views:
    283
  2. Winston

    Canon 20D noise reduction at high ISO's

    Winston, Feb 17, 2005, in forum: Digital Photography
    Replies:
    12
    Views:
    518
    Kevin McMurtrie
    Feb 22, 2005
  3. Larry R Harrison Jr

    D70 ISO 200 Noise vs EOS 300D ISO 100

    Larry R Harrison Jr, Jul 17, 2005, in forum: Digital Photography
    Replies:
    9
    Views:
    395
  4. Cynicor

    ISO noise vs. long exposure noise

    Cynicor, Aug 27, 2005, in forum: Digital Photography
    Replies:
    5
    Views:
    484
    Dirty Harry
    Aug 27, 2005
  5. Freedom55

    How imporrtant is high ISO noise?

    Freedom55, Apr 12, 2006, in forum: Digital Photography
    Replies:
    27
    Views:
    713
Loading...

Share This Page