shooting at high ISO

Discussion in 'Digital Photography' started by Ockham's Razor, Nov 9, 2006.

  1. Other than noise, is there some problem with shooting at high ISO?

    I am using a Canon S60 for a P&S and it can shoot at ISO 400 (and does
    so without noise). Using high ISO seems to increase the versatility of
    the camera at low light levels without the flash or allowing me to shoot
    at higher speeds without using a low F stop.

    --
    There are two ways to spell Ockham/Occam. Britannica prefers the former.
     
    Ockham's Razor, Nov 9, 2006
    #1
    1. Advertising

  2. Ockham's Razor

    ASAAR Guest

    On Wed, 08 Nov 2006 16:33:24 -0800, Ockham's Razor wrote:

    > Other than noise, is there some problem with shooting at high ISO?
    >
    > I am using a Canon S60 for a P&S and it can shoot at ISO 400 (and does
    > so without noise). Using high ISO seems to increase the versatility of
    > the camera at low light levels without the flash or allowing me to shoot
    > at higher speeds without using a low F stop.


    With increased noise you also lose detail. You may think that the
    S60 is noise free at ISO 400, but it may be because you're either
    not sufficiently magnifying its images on your monitor or because
    the prints you make are relatively small. Check the two dpreview
    pages (links below), the first at ISO 50 and the latter ISO 400.
    You can clearly see the noise in the ISO 400 shots, which produce a
    grainy, mottled effect. The word "Wednesday" which appears on the
    watch face is barely legible at ISO 400 and the ruler markings in
    the crop below that is *much* cleaner in the ISO 50 shots.

    If you don't intend to make prints larger than 4" x 6", ISO 400
    may be acceptable, but it's better to shoot at high ISO only when
    necessary, since if you ever want to make prints of cropped portions
    of any the S60's images, the crisp detail won't be hopelessly lost.
    While a tripod won't help get you the versatility of being able to
    shoot at higher speeds with smaller apertures, it will in many cases
    allow you to shoot in even lower light levels at ISO 50 or ISO 100
    than you'd be able to do without a tripod at ISO 400, and the
    results should come out much sharper because with low light levels
    you'd still be using shutter speeds too slow to eliminate much of
    the camera's movement, even at ISO 400.

    http://www.dpreview.com/reviews/canons60/page6.asp

    http://www.dpreview.com/reviews/canons60/page7.asp
     
    ASAAR, Nov 9, 2006
    #2
    1. Advertising

  3. In article <>,
    ASAAR <> wrote:

    > On Wed, 08 Nov 2006 16:33:24 -0800, Ockham's Razor wrote:
    >
    > > Other than noise, is there some problem with shooting at high ISO?
    > >
    > > I am using a Canon S60 for a P&S and it can shoot at ISO 400 (and does
    > > so without noise). Using high ISO seems to increase the versatility of
    > > the camera at low light levels without the flash or allowing me to shoot
    > > at higher speeds without using a low F stop.

    >
    > With increased noise you also lose detail. You may think that the
    > S60 is noise free at ISO 400, but it may be because you're either
    > not sufficiently magnifying its images on your monitor or because
    > the prints you make are relatively small. Check the two dpreview
    > pages (links below), the first at ISO 50 and the latter ISO 400.
    > You can clearly see the noise in the ISO 400 shots, which produce a
    > grainy, mottled effect. The word "Wednesday" which appears on the
    > watch face is barely legible at ISO 400 and the ruler markings in
    > the crop below that is *much* cleaner in the ISO 50 shots.
    >
    > If you don't intend to make prints larger than 4" x 6", ISO 400
    > may be acceptable, but it's better to shoot at high ISO only when
    > necessary, since if you ever want to make prints of cropped portions
    > of any the S60's images, the crisp detail won't be hopelessly lost.
    > While a tripod won't help get you the versatility of being able to
    > shoot at higher speeds with smaller apertures, it will in many cases
    > allow you to shoot in even lower light levels at ISO 50 or ISO 100
    > than you'd be able to do without a tripod at ISO 400, and the
    > results should come out much sharper because with low light levels
    > you'd still be using shutter speeds too slow to eliminate much of
    > the camera's movement, even at ISO 400.
    >
    > http://www.dpreview.com/reviews/canons60/page6.asp
    >
    > http://www.dpreview.com/reviews/canons60/page7.asp


    Thank you. It is amazing how much there is to learn about these things.

    --
    There are two ways to spell Ockham/Occam. Britannica prefers the former.
     
    Ockham's Razor, Nov 9, 2006
    #3
  4. Ockham's Razor

    bugbear Guest

    Ockham's Razor wrote:
    > Other than noise, is there some problem with shooting at high ISO?


    Yeah. In bright light (e.g. Egyptian desert) you might not
    have a fast enough shutter speed at wide apertures.

    BugBear
     
    bugbear, Nov 9, 2006
    #4
  5. Ockham's Razor

    Dave Cohen Guest

    Ockham's Razor wrote:
    > Other than noise, is there some problem with shooting at high ISO?
    >
    > I am using a Canon S60 for a P&S and it can shoot at ISO 400 (and does
    > so without noise). Using high ISO seems to increase the versatility of
    > the camera at low light levels without the flash or allowing me to shoot
    > at higher speeds without using a low F stop.
    >


    The iso question with digital isn't very different from film. Ideally
    one shoots at lower iso, but if it's a choice between that and no
    picture one uses higher iso. The great advantage of digital over film
    was unless one carried two bodies or had a camera with replaceable back,
    you were stuck with whatever you had loaded.
    With digital, not only can you change for each shot, depending on the
    shooting situation you may be able to shoot the same scene using
    different iso settings.
    Dave Cohen
     
    Dave Cohen, Nov 9, 2006
    #5
  6. Ockham's Razor

    Alan Meyer Guest

    ASAAR wrote:
    > On Wed, 08 Nov 2006 16:33:24 -0800, Ockham's Razor wrote:
    > ...
    > > Other than noise, is there some problem with shooting at high ISO?

    >
    > With increased noise you also lose detail.

    ....

    I believe that this is right. Here's why:

    A sensor can respond to a certain dynamic range of light,
    i.e., a certain range of values from very dark to very bright.

    Increasing the ISO in a photograph doesn't change the
    dynamic range of the sensor. It just takes the signal that
    is there and amplifies it.

    In digital terms you can think of it like this. If a sensor
    produces 12 bits of dynamic resolution, increasing the
    ISO by a factor of 4 (going from 100 to 400), multiplies
    the values by 4. 3/4ths of the resulting possible values will
    be off the scale that the processor can accept. The
    remaining 1/4th will now have gaps of at least 4 values
    between them instead of 1. The picture now has only
    1/4th as much information in it as a properly exposed
    ISO 100 image.

    Or to put it in binary terms, the 12 bit data is shifted
    2 bits to the left (where "left" means high order here)
    and the two low order bits are zero filled.

    Computer programmers will know what this means. I
    hope it is understandable to others

    However, boosting the ISO is still often the best way
    to get a usable shot..

    Alan
     
    Alan Meyer, Nov 9, 2006
    #6
  7. Alan Meyer wrote:
    > ASAAR wrote:
    >
    >>On Wed, 08 Nov 2006 16:33:24 -0800, Ockham's Razor wrote:
    >>...
    >>
    >>>Other than noise, is there some problem with shooting at high ISO?

    >>
    >> With increased noise you also lose detail.

    >
    > ...
    >
    > I believe that this is right. Here's why:
    >
    > A sensor can respond to a certain dynamic range of light,
    > i.e., a certain range of values from very dark to very bright.
    >
    > Increasing the ISO in a photograph doesn't change the
    > dynamic range of the sensor. It just takes the signal that
    > is there and amplifies it.
    >
    > In digital terms you can think of it like this. If a sensor
    > produces 12 bits of dynamic resolution, increasing the
    > ISO by a factor of 4 (going from 100 to 400), multiplies
    > the values by 4. 3/4ths of the resulting possible values will
    > be off the scale that the processor can accept. The
    > remaining 1/4th will now have gaps of at least 4 values
    > between them instead of 1. The picture now has only
    > 1/4th as much information in it as a properly exposed
    > ISO 100 image.
    >
    > Or to put it in binary terms, the 12 bit data is shifted
    > 2 bits to the left (where "left" means high order here)
    > and the two low order bits are zero filled.
    >
    > Computer programmers will know what this means. I
    > hope it is understandable to others
    >
    > However, boosting the ISO is still often the best way
    > to get a usable shot..
    >
    > Alan
    >

    This is not correct. Boosting ISO changes an analog gain, not a bit
    shift. Electronic sensors have one sensitivity (the quantum efficiency),
    but the analog signal from the sensor is amplified and sent to an
    A/D converter. At low ISO, many cameras (e.g. most DSLRs) are A/D
    limited, and as one increases ISO, the limit becomes sensor
    read noise limited for the low end, and as ISO increases, the high
    end (how many photons hit the 12-bit limit) decreases. Overall,
    dynamic range decreases.

    It is best illustrated in the sensor analysis:
    http://www.clarkvision.com/imagedetail/index.html#sensor_analysis

    For example, for a Nikon 200, table 1 at:
    http://www.clarkvision.com/imagedetail/evaluation-nikon-d200
    shows the dynamic range as a function of ISO.

    Roger
     
    Roger N. Clark (change username to rnclark), Nov 9, 2006
    #7
  8. Ockham's Razor

    ASAAR Guest

    On 9 Nov 2006 12:42:22 -0800, Alan Meyer wrote:

    >>> Other than noise, is there some problem with shooting at high ISO?

    >>
    >> With increased noise you also lose detail.

    > ...
    >
    > I believe that this is right. Here's why:


    Whatever the reason, I think that there are two factors
    responsible for loss of detail. I think that what you are
    describing is actually less related to reduced detail than it is to
    increasing the posterization or banding that often accompanies
    higher ISO and increased noise. Of the two factors, one might be
    the dynamic range of the sensor/amplifier combination that you and
    Roger appear to be describing. Perhaps more significant is the loss
    of detail attributed to overly aggressive in-camera noise reduction
    processing as well as using a high in-camera "sharpness" setting.
    By keeping both of these as low as possible, the results may appear
    inferior to what many P&S cameras produce using higher default
    settings. But by using computer software to reduce noise and
    enhance sharpness, the initial somewhat disappointing results can be
    improved a great deal, and without losing as much detail as is
    typically lost when allowing the camera to do the image processing.
     
    ASAAR, Nov 9, 2006
    #8
  9. Ockham's Razor

    Alan Meyer Guest

    Roger N. Clark (change username to rnclark) wrote:
    > > ...

    > This is not correct. Boosting ISO changes an analog gain, not a bit
    > shift.


    I didn't (stating the obvious) know that. I thought all processing was
    done after analog to digital conversion.

    > Electronic sensors have one sensitivity (the quantum efficiency),
    > but the analog signal from the sensor is amplified and sent to an
    > A/D converter. At low ISO, many cameras (e.g. most DSLRs) are A/D
    > limited, and as one increases ISO, the limit becomes sensor
    > read noise limited for the low end,


    Would it be correct to say then that the division of the signal
    into, say, 12 bits of digital data will produce 12 bits of
    resolution no matter how much the signal is amplified? If I
    understand your web pages correctly, that will happen, but the
    differences between adjacent values will become too small to
    know whether they are due to real differences or to thermal
    noise. In the limiting case, I guess the difference might amount to
    less than one electron - though the noise issue will dominate
    the picture quality problem long before we get to that point.

    > ... and as ISO increases, the high
    > end (how many photons hit the 12-bit limit) decreases. Overall,
    > dynamic range decreases.


    Is amplification applied uniformly over the whole range of the
    signal? Or do the cameras try to compress the differences
    at the high end while expanding the low end in order to try
    to avoid "clipping" bright parts?

    > It is best illustrated in the sensor analysis:
    > http://www.clarkvision.com/imagedetail/index.html#sensor_analysis


    Your website provides the best explanations of this stuff that
    I've seen anywhere on the web. Thank you very much for
    your efforts to educate us.

    Since I've got you on the figurative line here, I'd also like to
    ask you a question that's bugged me ever since I got into
    digital cameras - namely, How do electronic shutters work?

    Is there a good explanation somewhere on the web I can
    look at?

    Thanks.

    Alan
     
    Alan Meyer, Nov 9, 2006
    #9
  10. Alan Meyer wrote:
    > Roger N. Clark (change username to rnclark) wrote:


    >> Electronic sensors have one sensitivity (the quantum efficiency),
    >>but the analog signal from the sensor is amplified and sent to an
    >>A/D converter. At low ISO, many cameras (e.g. most DSLRs) are A/D
    >>limited, and as one increases ISO, the limit becomes sensor
    >>read noise limited for the low end,

    >
    > Would it be correct to say then that the division of the signal
    > into, say, 12 bits of digital data will produce 12 bits of
    > resolution no matter how much the signal is amplified? If I
    > understand your web pages correctly, that will happen, but the
    > differences between adjacent values will become too small to
    > know whether they are due to real differences or to thermal
    > noise. In the limiting case, I guess the difference might amount to
    > less than one electron - though the noise issue will dominate
    > the picture quality problem long before we get to that point.


    Yes, that is correct. For example, the 20D has a full well
    of about 50,000 electrons. At ISO 100, 12-bits (level 4095 is set
    to 50,000. At ISO 200, the 12 bits are spread over 25,000
    electrons; at ISO 400, about 12,500 electrons, up to ISO 3200,
    which is spread over 50,000/32 ~ 1560 electrons.
    12-bits, or 4095 levels spread over 1560 electrons means 0.38
    electrons per bit, but your really can't measure a fraction
    of an electron.

    Then consider a small pixel camera, like the Canon S70 with
    a full well of 8200 electrons at ISO 50. If the S70 had gains
    to go up to ISO 3200 (it does not), one would get a maximum
    of only 8200 /(3200/50) = 128 electrons. 12 bits spread over
    128 electrons does not make much sense. This explains why
    small pixel size cameras (like many P&S) have such poor low
    light performance, and why the manufacturers do not put such
    high ISOs on those cameras.

    >>... and as ISO increases, the high
    >>end (how many photons hit the 12-bit limit) decreases. Overall,
    >>dynamic range decreases.

    >
    > Is amplification applied uniformly over the whole range of the
    > signal? Or do the cameras try to compress the differences
    > at the high end while expanding the low end in order to try
    > to avoid "clipping" bright parts?


    The sensors respond linearly to light, and the analog to digital
    conversion is done on that linear signal. After conversion, a tone
    curve is applied, digitally.

    >>It is best illustrated in the sensor analysis:
    >>http://www.clarkvision.com/imagedetail/index.html#sensor_analysis

    >
    > Your website provides the best explanations of this stuff that
    > I've seen anywhere on the web. Thank you very much for
    > your efforts to educate us.


    Thanks.

    > Since I've got you on the figurative line here, I'd also like to
    > ask you a question that's bugged me ever since I got into
    > digital cameras - namely, How do electronic shutters work?


    It is done with voltages applied to the sensor.
    Here is a good explanation:
    http://www.olympusmicro.com/primer/java/photomicrography/ccd/shutter/index.html
    (or do a google search for electronic shutter).

    Roger
     
    Roger N. Clark (change username to rnclark), Nov 10, 2006
    #10
  11. ASAAR wrote:

    > On 9 Nov 2006 12:42:22 -0800, Alan Meyer wrote:
    >
    >>>>Other than noise, is there some problem with shooting at high ISO?
    >>> With increased noise you also lose detail.

    >>I believe that this is right. Here's why:

    >
    > Whatever the reason, I think that there are two factors
    > responsible for loss of detail. I think that what you are
    > describing is actually less related to reduced detail than it is to
    > increasing the posterization or banding that often accompanies
    > higher ISO and increased noise. Of the two factors, one might be
    > the dynamic range of the sensor/amplifier combination that you and
    > Roger appear to be describing. Perhaps more significant is the loss
    > of detail attributed to overly aggressive in-camera noise reduction
    > processing as well as using a high in-camera "sharpness" setting.
    > By keeping both of these as low as possible, the results may appear
    > inferior to what many P&S cameras produce using higher default
    > settings. But by using computer software to reduce noise and
    > enhance sharpness, the initial somewhat disappointing results can be
    > improved a great deal, and without losing as much detail as is
    > typically lost when allowing the camera to do the image processing.


    I agree that probably in recent cameras, especially high megapixel
    small sensor size, that the camera manufacturers are averaging
    pixel in an attempt to reduce noise, and it comes at the expense of
    detail. If you shot raw, you have a better chance of controlling
    that loss, but it does depend on the raw converter.

    But there is a fundamental loss in detail as noise increases,
    even if no noise reduction/pixel averaging is done. Consider the
    MTF of a system. Let's say you achieve an MTF of 60 line pairs
    per mm (lpm) at extinction MTF (0% MTF) at a signal-to-noise
    ratio (S/N) of 100, and at 30 lpm the MTF is 50%.
    Now lets add extreme noise so the S/N drops to 2. The 50% MTF
    at 30 lpm would be just detectable, and the high frequencies would
    be lost in the noise. The result of adding that noise is about
    a factor of 2 loss in detail. So while noise does decrease detail,
    it takes a fair amount of noise to do it (after all we can
    detect pretty good detail is high speed film that it really
    very noisy).

    Roger
     
    Roger N. Clark (change username to rnclark), Nov 10, 2006
    #11
  12. Ockham's Razor

    Alan Meyer Guest

    Great information Roger.

    Thank you.

    Alan
     
    Alan Meyer, Nov 10, 2006
    #12
  13. Ockham's Razor

    Shawn Hirn Guest

    In article <>,
    Ockham's Razor <> wrote:

    > Other than noise, is there some problem with shooting at high ISO?


    Not in dimly lighted areas. In brightly lighted scenes, a high ISO might
    overexpose the subject.
     
    Shawn Hirn, Nov 12, 2006
    #13
    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. JD

    How do you convert a CD ISO to DVD ISO?

    JD, Aug 20, 2004, in forum: Computer Support
    Replies:
    9
    Views:
    29,398
  2. Bill Smith
    Replies:
    1
    Views:
    906
    DaveJ
    Aug 1, 2003
  3. Salem Derisavi

    Figuring out ISO settings in Auto ISO mode

    Salem Derisavi, Sep 28, 2003, in forum: Digital Photography
    Replies:
    3
    Views:
    1,575
    Andrew McDonald
    Sep 29, 2003
  4. Pumpkin King

    Tips for Shooting High School Wrestling?

    Pumpkin King, Dec 1, 2003, in forum: Digital Photography
    Replies:
    0
    Views:
    614
    Pumpkin King
    Dec 1, 2003
  5. EarGuy

    Shooting High School Football + Using flash

    EarGuy, Sep 8, 2006, in forum: Digital Photography
    Replies:
    5
    Views:
    485
    John McWilliams
    Sep 8, 2006
Loading...

Share This Page