Sensor cleaning mode maps out hot pixels

Discussion in 'Digital Photography' started by oparr@hotmail.com, Mar 31, 2006.

  1. Guest

    There seems to be a popular belief on http://www.dpreview.com/forums/
    that the Canon 5D and 30D will map out hot pixels when one goes into
    sensor cleaning mode. This sounds more like an April fool's joke to me.
    Does anyone here have any credible information on this?
     
    , Mar 31, 2006
    #1
    1. Advertising

  2. Jim Townsend Guest

    wrote:

    > There seems to be a popular belief on http://www.dpreview.com/forums/
    > that the Canon 5D and 30D will map out hot pixels when one goes into
    > sensor cleaning mode. This sounds more like an April fool's joke to me.
    > Does anyone here have any credible information on this?


    All Canon's sensor cleaning mode does is lift the mirror and open
    the shutter blades so the sensor is exposed. (It would be mighty
    hard to clean the sensor otherwise :)
     
    Jim Townsend, Mar 31, 2006
    #2
    1. Advertising

  3. Guest

    Jim Townsend wrote:
    > All Canon's sensor cleaning mode does is lift the mirror and open
    > the shutter blades so the sensor is exposed. (It would be mighty
    > hard to clean the sensor otherwise :)


    In other words as far as you know it does no more than the above on a
    5D or 30D? Do you have either of these cameras? I have two D60s and a
    20D and know what sensor cleaning mode does on them. I don't have a 5D
    or a 30D.
     
    , Mar 31, 2006
    #3
  4. "" <> writes:

    > Jim Townsend wrote:
    >> All Canon's sensor cleaning mode does is lift the mirror and open
    >> the shutter blades so the sensor is exposed. (It would be mighty
    >> hard to clean the sensor otherwise :)

    >
    > In other words as far as you know it does no more than the above on a
    > 5D or 30D? Do you have either of these cameras? I have two D60s and a
    > 20D and know what sensor cleaning mode does on them. I don't have a 5D
    > or a 30D.


    How could it possibly do anything about hot pixels?

    --
    Måns Rullgård
     
    =?iso-8859-1?Q?M=E5ns_Rullg=E5rd?=, Mar 31, 2006
    #4
  5. ASAAR Guest

    On Fri, 31 Mar 2006 19:37:28 +0100, Måns Rullgård wrote:

    >> In other words as far as you know it does no more than the above on a
    >> 5D or 30D? Do you have either of these cameras? I have two D60s and a
    >> 20D and know what sensor cleaning mode does on them. I don't have a 5D
    >> or a 30D.

    >
    > How could it possibly do anything about hot pixels?


    Some cameras need to be returned to the manufacturer to map out
    hot pixels. Others (such as some Olympus models) allow it to be
    done by the user - it's a menu option. If Canon wanted to allow
    their 5D or 30D models to map out hot pixels it could easily be
    added as a menu option, but one can safely assume that it wouldn't
    be added as an unannounced side-effect of a sensor cleaning mode. I
    would like to have been able to say that Canon maps out hot pixels
    whenever histograms are enabled or the 5D's hidden flash is
    extended, but it's not yet April 1 on this part of the globe.
     
    ASAAR, Mar 31, 2006
    #5
  6. Mxsmanic Guest

    ASAAR writes:

    > Some cameras need to be returned to the manufacturer to map out
    > hot pixels. Others (such as some Olympus models) allow it to be
    > done by the user - it's a menu option. If Canon wanted to allow
    > their 5D or 30D models to map out hot pixels it could easily be
    > added as a menu option, but one can safely assume that it wouldn't
    > be added as an unannounced side-effect of a sensor cleaning mode. I
    > would like to have been able to say that Canon maps out hot pixels
    > whenever histograms are enabled or the 5D's hidden flash is
    > extended, but it's not yet April 1 on this part of the globe.


    If the sensor is of decent quality, there are no hot pixels to "map
    out."

    --
    Transpose mxsmanic and gmail to reach me by e-mail.
     
    Mxsmanic, Mar 31, 2006
    #6
  7. In article <>, Måns Rullgård
    <> writes
    >"" <> writes:
    >
    >> Jim Townsend wrote:
    >>> All Canon's sensor cleaning mode does is lift the mirror and open
    >>> the shutter blades so the sensor is exposed. (It would be mighty
    >>> hard to clean the sensor otherwise :)

    >>
    >> In other words as far as you know it does no more than the above on a
    >> 5D or 30D? Do you have either of these cameras? I have two D60s and a
    >> 20D and know what sensor cleaning mode does on them. I don't have a 5D
    >> or a 30D.

    >
    >How could it possibly do anything about hot pixels?
    >

    Relatively easily. Other cameras have a hot pixel mapping capability
    already.

    All it needs is a dark frame to be captured and then pixels above a
    certain threshold value to be mapped out as hot by the processor.
    Ideally the dark frame would be a reasonable length exposure.

    The Canon 5D has a 2-3second delay between pressing "OK" on the sensor
    clean dialog and the shutter actually opening that other Canon cameras
    don't seem to have, which suggests something is being done. This would
    certainly be a suitable time to implement the entire expose-mapping
    process.

    I haven't seen any hot pixels on my 5D to be able to check if this is
    true, but at the same time I wouldn't rule it out. Borderline pixels
    are likely to be more of an issue as pixel count increases, so the
    introduction of some in-camera process to deal with them on high
    resolution cameras shouldn't be too surprising.
    --
    Kennedy
    Yes, Socrates himself is particularly missed;
    A lovely little thinker, but a bugger when he's pissed.
    Python Philosophers (replace 'nospam' with 'kennedym' when replying)
     
    Kennedy McEwen, Apr 1, 2006
    #7
  8. In article <>, Mxsmanic
    <> writes
    >
    >If the sensor is of decent quality, there are no hot pixels to "map
    >out."
    >

    Really? Do the arithmetic. The current industrial maxim for
    "perfection" is six-sigma. That corresponds to 3.4 defects in every
    million units produced. To put it in perspective that means a service
    outage of less than 20 minutes every ten years for telecoms, power or
    broadcast systems. Few manufacturers or delivery systems actually
    achieve it, but it is a valid aspiration.

    In terms of sensor in digital cameras, six sigma is 3.4 defective pixels
    per megapixel. So in 12.8million pixels, best practice standards would
    result in an *average* of 44 hot or cold pixels in each sensor produced.

    You can certainly buy sensors with no "hot" pixels to map out, but I am
    fairly confident that you are unlikely to earn enough in you entire
    working life to afford one! So your body will be in a rather indecent
    state before you'll ever own what you consider to be a "decent quality"
    digital camera!
    --
    Kennedy
    Yes, Socrates himself is particularly missed;
    A lovely little thinker, but a bugger when he's pissed.
    Python Philosophers (replace 'nospam' with 'kennedym' when replying)
     
    Kennedy McEwen, Apr 1, 2006
    #8
  9. Guest

    , Apr 1, 2006
    #9
  10. Guest

    >All it needs is a dark frame to be captured and then pixels above a
    >certain threshold value to be mapped out as hot by the processor.


    Yes it's feasible that sensor cleaning mode could be used to map out
    pixels that become hot after a second or two. It would then be left up
    to the user to turn on NR (dark frame noise reduction) to handle any
    after a second or more if necessary. But why would Canon not document
    this feature? One reason is given in the link below but it makes no
    sense IMO;

    http://forums.dpreview.com/forums/read.asp?forum=1032&message=17838784

    Kennedy McEwen wrote:
    >
     
    , Apr 1, 2006
    #10
  11. Paul Allen Guest

    Måns Rullgård wrote:
    > "" <> writes:
    >
    >> Jim Townsend wrote:
    >>> All Canon's sensor cleaning mode does is lift the mirror and open
    >>> the shutter blades so the sensor is exposed. (It would be mighty
    >>> hard to clean the sensor otherwise :)

    >> In other words as far as you know it does no more than the above on a
    >> 5D or 30D? Do you have either of these cameras? I have two D60s and a
    >> 20D and know what sensor cleaning mode does on them. I don't have a 5D
    >> or a 30D.

    >
    > How could it possibly do anything about hot pixels?


    It replaces their data with data from spare pixels around the edge of
    the sensor. Have you ever noticed how cameras are spec'd as having
    a certain number of actual pixels and a slightly smaller number of
    "effective" pixels. The difference between the actual and effective
    pixel count is made up by the spare pixels around the edge. You'd
    think that it would be easier to remap photons to a spare pixel that
    was near the pixel it replaces. Researchers have discovered, much
    to their surprise, that distributing the spare pixels thoughout the
    sensor disrupts the final image. So instead they employ some fancy
    optical footwork to redirect the hot pixel's photons to a spare
    that's out of the way at the edge of the sensor. Neat, huh?

    Paul Allen
     
    Paul Allen, Apr 1, 2006
    #11
  12. In article <>,
    "" <> writes

    >But why would Canon not document
    >this feature? One reason is given in the link below but it makes no
    >sense IMO;
    >

    Canon choose not to document a huge number of "features". For example,
    there isn't one single official Canon documentation of how their flash
    system works - only unofficial, mainly composed of rumour, experiment,
    trial and error and reverse engineering descriptions. In terms of
    documentation they are almost as bad as Sony, if that were possible! ;-)

    I wouldn't rely on lack of documentation as a reason to discount a
    feature.
    --
    Kennedy
    Yes, Socrates himself is particularly missed;
    A lovely little thinker, but a bugger when he's pissed.
    Python Philosophers (replace 'nospam' with 'kennedym' when replying)
     
    Kennedy McEwen, Apr 1, 2006
    #12
  13. SamSez Guest

    "" <> wrote in
    news::

    > There seems to be a popular belief on http://www.dpreview.com/forums/
    > that the Canon 5D and 30D will map out hot pixels when one goes into
    > sensor cleaning mode. This sounds more like an April fool's joke to me.
    > Does anyone here have any credible information on this?
    >


    Sorry, I don't believe it. While it is certainly possible to provide hot
    pixel mapping as an in-camera function [my oly has it], it makes absolutely
    no sense to 'combine' it with the current mirror-up, [lens off,] time to
    blow off the sensor functionality of the canon series. Having the mirror
    up and possibly the lens off is contradictory with finding hot pixels. If
    they added another menu function, maybe -- but not combined with the very
    function that the sensor cleaning menu choice provides.
     
    SamSez, Apr 1, 2006
    #13
  14. Rich Guest

    On Sat, 1 Apr 2006 01:05:21 +0100, Kennedy McEwen
    <> wrote:

    >In article <>, Mxsmanic
    ><> writes
    >>
    >>If the sensor is of decent quality, there are no hot pixels to "map
    >>out."
    >>

    >Really? Do the arithmetic. The current industrial maxim for
    >"perfection" is six-sigma. That corresponds to 3.4 defects in every
    >million units produced. To put it in perspective that means a service
    >outage of less than 20 minutes every ten years for telecoms, power or
    >broadcast systems. Few manufacturers or delivery systems actually
    >achieve it, but it is a valid aspiration.


    A worthless aspiration right now as a medical grade CCD (I don't even
    know if they make CMOS at that level) costs about 8-10x what the
    consumer stuff costs. Would you pay $40,000-$60,000 for a 1DsMkII?
    -Rich
     
    Rich, Apr 1, 2006
    #14
  15. In article <Xns9797D0CE48F36samthemanverizonnet@199.45.49.11>, SamSez
    <> writes
    >"" <> wrote in
    >news::
    >
    >> There seems to be a popular belief on http://www.dpreview.com/forums/
    >> that the Canon 5D and 30D will map out hot pixels when one goes into
    >> sensor cleaning mode. This sounds more like an April fool's joke to me.
    >> Does anyone here have any credible information on this?
    >>

    >
    >Sorry, I don't believe it. While it is certainly possible to provide hot
    >pixel mapping as an in-camera function [my oly has it], it makes absolutely
    >no sense to 'combine' it with the current mirror-up, [lens off,] time to
    >blow off the sensor functionality of the canon series.


    Why do you think that? There is an inordinately long time between
    selecting OK on the sensor clean dialog and anything physical happening.
    This delay doesn't appear to be there with the 20D or other Canon
    cameras, so it certainly appears to be a new function that has been
    introduced on this model.

    > Having the mirror
    >up and possibly the lens off is contradictory with finding hot pixels.


    There is no contradiction if the operation is completed *before* the
    mirror goes up and the shutter opens - everything occurs when the sensor
    is in the dark.

    >If
    >they added another menu function, maybe -- but not combined with the very
    >function that the sensor cleaning menu choice provides.


    Again, it makes perfect sense to do it then. It isn't something that
    needs to be done on a regular basis such as power-up, since pixel maps
    are relatively stable over time.

    I am not supporting the claims that this feature is present in the
    sensor clean function, since it isn't something I have been able to
    check for myself, but the anecdotal evidence which supports it from
    those who have experience makes more sense than your arguments that it
    doesn't or can't be done.
    --
    Kennedy
    Yes, Socrates himself is particularly missed;
    A lovely little thinker, but a bugger when he's pissed.
    Python Philosophers (replace 'nospam' with 'kennedym' when replying)
     
    Kennedy McEwen, Apr 1, 2006
    #15
  16. SamSez Guest

    Kennedy McEwen <> wrote in
    news::

    > In article <Xns9797D0CE48F36samthemanverizonnet@199.45.49.11>, SamSez
    > <> writes
    >>"" <> wrote in
    >>news::
    >>
    >>> There seems to be a popular belief on
    >>> http://www.dpreview.com/forums/ that the Canon 5D and 30D will map
    >>> out hot pixels when one goes into sensor cleaning mode. This sounds
    >>> more like an April fool's joke to me. Does anyone here have any
    >>> credible information on this?
    >>>

    >>
    >>Sorry, I don't believe it. While it is certainly possible to provide
    >>hot pixel mapping as an in-camera function [my oly has it], it makes
    >>absolutely no sense to 'combine' it with the current mirror-up, [lens
    >>off,] time to blow off the sensor functionality of the canon series.

    >
    > Why do you think that? There is an inordinately long time between
    > selecting OK on the sensor clean dialog and anything physical
    > happening. This delay doesn't appear to be there with the 20D or other
    > Canon cameras, so it certainly appears to be a new function that has
    > been introduced on this model.
    >
    >> Having the mirror
    >>up and possibly the lens off is contradictory with finding hot pixels.

    >
    > There is no contradiction if the operation is completed *before* the
    > mirror goes up and the shutter opens - everything occurs when the
    > sensor is in the dark.
    >
    >>If
    >>they added another menu function, maybe -- but not combined with the
    >>very function that the sensor cleaning menu choice provides.

    >
    > Again, it makes perfect sense to do it then. It isn't something that
    > needs to be done on a regular basis such as power-up, since pixel maps
    > are relatively stable over time.
    >
    > I am not supporting the claims that this feature is present in the
    > sensor clean function, since it isn't something I have been able to
    > check for myself, but the anecdotal evidence which supports it from
    > those who have experience makes more sense than your arguments that it
    > doesn't or can't be done.


    [in your little mind perhaps...]
     
    SamSez, Apr 1, 2006
    #16
  17. Guest

    >Canon choose not to document a huge number of "features".

    Assuming the feature existed in the 5D, I'm wondering if Canon's
    silence has something to do with the fact that their more costly pro 1
    series DLSRs don't have this feature. There is also an eerie silence in
    the DPReview forum that deals with both 1 and 5 series bodies over this
    issue. The most I'm able to get out of them is that it was discussed
    long ago and reference to the mentioned single thread. Very few
    opinions are given. It's as though one is supposed to either disbelieve
    or take a leap of faith and believe.

    Kennedy McEwen wrote:
    >
     
    , Apr 1, 2006
    #17
  18. Mxsmanic Guest

    Kennedy McEwen writes:

    > Really? Do the arithmetic.


    Skip the arithmetic. Do the economics.

    > The current industrial maxim for
    > "perfection" is six-sigma. That corresponds to 3.4 defects in every
    > million units produced. To put it in perspective that means a service
    > outage of less than 20 minutes every ten years for telecoms, power or
    > broadcast systems. Few manufacturers or delivery systems actually
    > achieve it, but it is a valid aspiration.


    The telephone companies achieve it. Disk drives achieve it.
    Electronic memory modules achieve it. Good flat-panel monitors
    achieve it.

    > In terms of sensor in digital cameras, six sigma is 3.4 defective pixels
    > per megapixel. So in 12.8million pixels, best practice standards would
    > result in an *average* of 44 hot or cold pixels in each sensor produced.


    That would also mean seven bad pixels in a 1600x1200 monitor, which is
    far above what the good monitors actually show (many have no bad
    pixels at all).

    > You can certainly buy sensors with no "hot" pixels to map out, but I am
    > fairly confident that you are unlikely to earn enough in you entire
    > working life to afford one!


    I'm not. They aren't that expensive, although they are more expensive
    than defective sensors. If you're paying $8000 for a camera body,
    it's entirely reasonable to expect zero defective pixels. If you're
    buying consumer junk, obviously all bets are off.

    --
    Transpose mxsmanic and gmail to reach me by e-mail.
     
    Mxsmanic, Apr 1, 2006
    #18
  19. Paul Allen <"paul dot l dot allen at comcast dot net"> writes:

    > Måns Rullgård wrote:
    >> "" <> writes:
    >>
    >>> Jim Townsend wrote:
    >>>> All Canon's sensor cleaning mode does is lift the mirror and open
    >>>> the shutter blades so the sensor is exposed. (It would be mighty
    >>>> hard to clean the sensor otherwise :)
    >>> In other words as far as you know it does no more than the above on a
    >>> 5D or 30D? Do you have either of these cameras? I have two D60s and a
    >>> 20D and know what sensor cleaning mode does on them. I don't have a 5D
    >>> or a 30D.

    >> How could it possibly do anything about hot pixels?

    >
    > It replaces their data with data from spare pixels around the edge of
    > the sensor. Have you ever noticed how cameras are spec'd as having
    > a certain number of actual pixels and a slightly smaller number of
    > "effective" pixels. The difference between the actual and effective
    > pixel count is made up by the spare pixels around the edge. You'd
    > think that it would be easier to remap photons to a spare pixel that
    > was near the pixel it replaces. Researchers have discovered, much
    > to their surprise, that distributing the spare pixels thoughout the
    > sensor disrupts the final image. So instead they employ some fancy
    > optical footwork to redirect the hot pixel's photons to a spare
    > that's out of the way at the edge of the sensor. Neat, huh?


    Well, it is April 1...

    I know how hot pixels can be dealt with. Lifting the mirror and
    opening the shutter is certainly not going make finding them any
    easier.

    --
    Måns Rullgård
     
    =?iso-8859-1?Q?M=E5ns_Rullg=E5rd?=, Apr 1, 2006
    #19
  20. In article <>, Rich
    <> writes
    >On Sat, 1 Apr 2006 01:05:21 +0100, Kennedy McEwen
    ><> wrote:
    >
    >>In article <>, Mxsmanic
    >><> writes
    >>>
    >>>If the sensor is of decent quality, there are no hot pixels to "map
    >>>out."
    >>>

    >>Really? Do the arithmetic. The current industrial maxim for
    >>"perfection" is six-sigma. That corresponds to 3.4 defects in every
    >>million units produced. To put it in perspective that means a service
    >>outage of less than 20 minutes every ten years for telecoms, power or
    >>broadcast systems. Few manufacturers or delivery systems actually
    >>achieve it, but it is a valid aspiration.

    >
    >A worthless aspiration


    It is clear to see why a non-entity would consider the aspiration of
    perfection to be worthless!
    --
    Kennedy
    Yes, Socrates himself is particularly missed;
    A lovely little thinker, but a bugger when he's pissed.
    Python Philosophers (replace 'nospam' with 'kennedym' when replying)
     
    Kennedy McEwen, Apr 1, 2006
    #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. Mark Grady
    Replies:
    10
    Views:
    881
  2. notreallyme
    Replies:
    14
    Views:
    983
  3. kl_tom

    Can hot pixels become dead pixels?

    kl_tom, Oct 4, 2006, in forum: Digital Photography
    Replies:
    4
    Views:
    541
    Bill Funk
    Oct 5, 2006
  4. Ralph Fox

    MSN maps vs. Google maps

    Ralph Fox, Jul 30, 2007, in forum: NZ Computing
    Replies:
    0
    Views:
    595
    Ralph Fox
    Jul 30, 2007
  5. RichA
    Replies:
    9
    Views:
    759
Loading...

Share This Page