Nikon Matrix metering idiosyncracies

Discussion in 'Digital Photography' started by eNo, Aug 6, 2009.

  1. eNo

    eNo Guest

    Good morning, fellow shutter bugs. Over the next 3 days (or as many as
    it takes) I will be posting a series of write-ups on Matrix metering
    as I am able to evaluate it with the Nikon D80 and D90. You can access
    part 1 at my blog (link below). I don't want to blow the suspense --
    and I also want to remain open to new findings as I take more sample
    images -- but let's just that at the outset, some of my early-on
    notions regarding the D90's metering are being challenged to the core.
    You will see more on this in part 2 of the series.

    I'm also putting out a call for anyone with other Nikon DSLRs (D40,
    D40x, D60, D200, D300, etc.) who can take some time to reproduce the 4
    basic test cases I provide in part 1 to contact me via a blog comment
    so we can arrange a way to incorporate their sample images into my
    blog. If you shoot with a Canon DSLR, I'd also be interested to see
    some comparable samples. Thanks.
    eNo, Aug 6, 2009
    #1
    1. Advertising

  2. eNo

    PDM Guest

    "eNo" <> wrote in message
    news:...
    > Good morning, fellow shutter bugs. Over the next 3 days (or as many as
    > it takes) I will be posting a series of write-ups on Matrix metering
    > as I am able to evaluate it with the Nikon D80 and D90. You can access
    > part 1 at my blog (link below). I don't want to blow the suspense --
    > and I also want to remain open to new findings as I take more sample
    > images -- but let's just that at the outset, some of my early-on
    > notions regarding the D90's metering are being challenged to the core.
    > You will see more on this in part 2 of the series.
    >
    > I'm also putting out a call for anyone with other Nikon DSLRs (D40,
    > D40x, D60, D200, D300, etc.) who can take some time to reproduce the 4
    > basic test cases I provide in part 1 to contact me via a blog comment
    > so we can arrange a way to incorporate their sample images into my
    > blog. If you shoot with a Canon DSLR, I'd also be interested to see
    > some comparable samples. Thanks.


    Where is the link?

    PDM
    PDM, Aug 6, 2009
    #2
    1. Advertising

  3. eNo

    Nicko Guest

    On Aug 6, 8:39 am, Glen Calvers <> wrote:

    > You will find that on *ANY* dSLR system that any metering light-path cannot
    > adequately compensate for varying light-cones from different focal-length
    > lenses.  A wide-angle lens will not read the same exposure at the same
    > f-ratio as a long zoom lens at the same f-ratio. This problem of the SLR
    > design has existed since the very first SLRs with in-camera metering. The
    > metering system is optimized for one focal-length only, and that's it.


    No, I shall not find that.

    What in gawd's name are you talking about? A meter reads what is
    presented to it. Period. Regardless of the length of the light path,
    if the meter is adequately adjusted, it will work the same with any
    lens. How in any sense does this relate to the focal length of the
    lens? What kind of camera have you been using for a bong?

    In
    > the past this used to be for the 50-55mm lens. The best you can hope for is
    > an approximation of what it should be when using the separate light-path in
    > the SLR design.


    Wow.

    This is why so many dSLR owners have to depend on RAW data
    > in order to regain what their metering system wrongly exposed in the first
    > place. Just one of the "lets not tell them" effects by promoters of the SLR
    > design.


    What are you smoking? Because I want to avoid it. I hate things that
    make me stupid.

    Unless you are talking about people using programmed-exposure-only
    dSLRs (is there such a thing?), the only issue here is the
    photographer's failure to interpret the meter reading and to make
    appropriate compensation when necessary. The metering system does not
    necessarily set the exposure; the photographer does. I'd wager that
    most SLR photographers who use RAW use it because of the post-
    processing versatility it offers, rather than to make up for poor
    exposure judgement (which, however, is a plus when you're drunk).

    --
    YOP...
    Nicko, Aug 7, 2009
    #3
  4. Glen Calvers <> wrote:
    > On Fri, 07 Aug 2009 09:20:56 +0100, bugbear
    > <bugbear@trim_papermule.co.uk_trim> wrote:


    >>Glen Calvers wrote:
    >>> On Thu, 6 Aug 2009 06:00:37 -0700 (PDT), eNo <>

    >>
    >>>
    >>> You will find that on *ANY* dSLR system that any metering light-path cannot
    >>> adequately compensate for varying light-cones from different focal-length
    >>> lenses. A wide-angle lens will not read the same exposure at the same
    >>> f-ratio as a long zoom lens at the same f-ratio. This problem of the SLR
    >>> design has existed since the very first SLRs with in-camera metering. The
    >>> metering system is optimized for one focal-length only, and that's it. In
    >>> the past this used to be for the 50-55mm lens.

    >>
    >>That's interesting - can you point me at an explanation?
    >>
    >> BugBear


    > I've not bothered to look for one online, this problem has existed for ages
    > and has not changed. I know the problem is an inherent design flaw in all
    > SLR metering light-paths since the original in-camera metering designs were
    > invented. I would assume everyone knew this as basic common-knowledge about
    > the SLR design since it's existed for so long. Except for those in the past
    > that used OTF (or off-the-curtain) metering, bypassing the metering done in
    > the pentaprism's light path. But even then there were still some
    > discrepancies due to the angle of the light-bundles coming from the lens
    > and hitting the film or shutter curtain at a sharper or less sharp angle,
    > and the amount of reflectivity of the surface from which the measure was
    > taken (patterned shutter-curtain, or different reflectivities of film
    > surfaces). This is the same reason a movie-projection screen, projection TV
    > system, even LCD displays, are not always the same intensity depending on
    > your angle of view to it.


    > The problem is rather simple. If you have a wide-angle lens on the camera,
    > the light-cone coming from the back of that lens is very wide and shallow.
    > Light coming from the aperture will be striking the focusing screen, any
    > metering sensors, and even the CCD sensor itself at a sharper angle. (This
    > is why sensor blooming is more of a problem with wide-angle lenses and it
    > occurring more in the corners than the center of an image. This more
    > glancing light path spills over into adjoining pixels.) Conversely with a
    > telephoto lens the light-cone is narrow and long. Because the light-sensors
    > in the SLR design must be situated in a flat plane on the focusing screen
    > the sideways glancing light bundles from the wider and shallower cone from
    > a wide-angle lens won't all hit the exposure-metering sensors
    > perpendicularly. With a telephoto lens then they all hit those sensor spots
    > more perpendicularly.


    > The d/SLR's in-camera metering-system is calibrated/designed with this
    > inherent problem in mind. The amount of light hitting those sensors is
    > generally optimized for the angles of light coming from the back of a
    > "standard" 50-55mm lens. Above and below that focal-length of lens and the
    > light hitting those sensors won't be at the same angle. (Not unlike why you
    > have winter in the northern hemisphere when the earth is tilted away from
    > the sun. The light energy is spread over a wider area and is dimmer.) This
    > is true whether the light sensors are contained right within the focusing
    > screen itself, or if they are adjacent to the focusing screen somewhere
    > else in the housing. This problem is worse if the metering detection cells
    > are placed above the focusing screen and looking down from the sides into
    > the focusing screen. Then the apparent light hitting them will change even
    > more greatly depending on the focal-length of lens in use, the density of
    > "frost" on the focusing screen, and the focal-lengths of any fresnel lenses
    > incorporated into that focusing screen to ensure the brightest image
    > possible for your eye. This too is why more advanced camera systems offered
    > various focusing-screens depending on the focal-length of lens you were
    > going to use with them. Olympus SLRs addressing this problem the best in
    > the past. These focusing screens weren't just for providing an optimized
    > image for the eye (and the angle of any microprisms cut so they wouldn't
    > black-out), but also to make sure the in-camera metering system could also
    > perform more correctly with those focal-lengths for which those focusing
    > screens were designed.


    > The only true cure for this design flaw is to get your metering direct from
    > the sensor which is capturing the image.


    Or to use metering sensors which have the same response to angled
    light as the image sensor. A perfect solution which would be easy to
    implement in a digital camera :)

    --
    Chris Malcolm
    Chris Malcolm, Aug 7, 2009
    #4
  5. eNo

    eNo Guest

    On Aug 6, 10:55 am, "PDM" <pdcm99[deletethisbit]@tiscali.co.uk> wrote:
    > "eNo" <> wrote in message
    >
    > news:...
    >
    > > Good morning, fellow shutter bugs. Over the next 3 days (or as many as
    > > it takes) I will be posting a series of write-ups on Matrix metering
    > > as I am able to evaluate it with the Nikon D80 and D90. You can access
    > > part 1 at my blog (link below). I don't want to blow the suspense --
    > > and I also want to remain open to new findings as I take more sample
    > > images -- but let's just that at the outset, some of my early-on
    > > notions regarding the D90's metering are being challenged to the core.
    > > You will see more on this in part 2 of the series.

    >
    > > I'm also putting out a call for anyone with other Nikon DSLRs (D40,
    > > D40x, D60, D200, D300, etc.) who can take some time to reproduce the 4
    > > basic test cases I provide in part 1 to contact me via a blog comment
    > > so we can arrange a way to incorporate their sample images into my
    > > blog. If you shoot with a Canon DSLR, I'd also be interested to see
    > > some comparable samples. Thanks.

    >
    > Where is the link?
    >
    > PDM


    http://esfotoclix.com/blog1
    eNo, Aug 7, 2009
    #5
  6. eNo

    eNo Guest

    On Aug 7, 2:29 am, Glen Calvers <> wrote:
    > On Fri, 07 Aug 2009 09:20:56 +0100, bugbear
    >
    >
    >
    >
    >
    > <bugbear@trim_papermule.co.uk_trim> wrote:
    > >Glen Calvers wrote:
    > >> On Thu, 6 Aug 2009 06:00:37 -0700 (PDT), eNo <>

    >
    > >> You will find that on *ANY* dSLR system that any metering light-path cannot
    > >> adequately compensate for varying light-cones from different focal-length
    > >> lenses.  A wide-angle lens will not read the same exposure at the same
    > >> f-ratio as a long zoom lens at the same f-ratio. This problem of the SLR
    > >> design has existed since the very first SLRs with in-camera metering. The
    > >> metering system is optimized for one focal-length only, and that's it. In
    > >> the past this used to be for the 50-55mm lens.

    >
    > >That's interesting - can you point me at an explanation?

    >
    > >   BugBear

    >
    > I've not bothered to look for one online, this problem has existed for ages
    > and has not changed. I know the problem is an inherent design flaw in all
    > SLR metering light-paths since the original in-camera metering designs were
    > invented. I would assume everyone knew this as basic common-knowledge about
    > the SLR design since it's existed for so long. Except for those in the past
    > that used OTF (or off-the-curtain) metering, bypassing the metering done in
    > the pentaprism's light path. But even then there were still some
    > discrepancies due to the angle of the light-bundles coming from the lens
    > and hitting the film or shutter curtain at a sharper or less sharp angle,
    > and the amount of reflectivity of the surface from which the measure was
    > taken (patterned shutter-curtain, or different reflectivities of film
    > surfaces). This is the same reason a movie-projection screen, projection TV
    > system, even LCD displays, are not always the same intensity depending on
    > your angle of view to it.
    >
    > The problem is rather simple. If you have a wide-angle lens on the camera,
    > the light-cone coming from the back of that lens is very wide and shallow..
    > Light coming from the aperture will be striking the focusing screen, any
    > metering sensors, and even the CCD sensor itself at a sharper angle. (This
    > is why sensor blooming is more of a problem with wide-angle lenses and it
    > occurring more in the corners than the center of an image. This more
    > glancing light path spills over into adjoining pixels.) Conversely with a
    > telephoto lens the light-cone is narrow and long. Because the light-sensors
    > in the SLR design must be situated in a flat plane on the focusing screen
    > the sideways glancing light bundles from the wider and shallower cone from
    > a wide-angle lens won't all hit the exposure-metering sensors
    > perpendicularly. With a telephoto lens then they all hit those sensor spots
    > more perpendicularly.
    >
    > The d/SLR's in-camera metering-system is calibrated/designed with this
    > inherent problem in mind. The amount of light hitting those sensors is
    > generally optimized for the angles of light coming from the back of a
    > "standard" 50-55mm lens. Above and below that focal-length of lens and the
    > light hitting those sensors won't be at the same angle. (Not unlike why you
    > have winter in the northern hemisphere when the earth is tilted away from
    > the sun. The light energy is spread over a wider area and is dimmer.) This
    > is true whether the light sensors are contained right within the focusing
    > screen itself, or if they are adjacent to the focusing screen somewhere
    > else in the housing. This problem is worse if the metering detection cells
    > are placed above the focusing screen and looking down from the sides into
    > the focusing screen. Then the apparent light hitting them will change even
    > more greatly depending on the focal-length of lens in use, the density of
    > "frost" on the focusing screen, and the focal-lengths of any fresnel lenses
    > incorporated into that focusing screen to ensure the brightest image
    > possible for your eye. This too is why more advanced camera systems offered
    > various focusing-screens depending on the focal-length of lens you were
    > going to use with them. Olympus SLRs addressing this problem the best in
    > the past. These focusing screens weren't just for providing an optimized
    > image for the eye (and the angle of any microprisms cut so they wouldn't
    > black-out), but also to make sure the in-camera metering system could also
    > perform more correctly with those focal-lengths for which those focusing
    > screens were designed.
    >
    > The only true cure for this design flaw is to get your metering direct from
    > the sensor which is capturing the image.


    I'll have to check this effect out. I'm assuming it would affect spot
    and center-weighed metering as well...

    In the meantime, here's the next installment in the Matrix metering
    series:
    http://esfotoclix.com/blog1/?p=280
    eNo, Aug 7, 2009
    #6
  7. eNo

    nospam Guest

    In article <>, Glen Calvers
    <> wrote:

    > I've not bothered to look for one online, this problem has existed for ages
    > and has not changed. I know the problem is an inherent design flaw in all
    > SLR metering light-paths since the original in-camera metering designs were
    > invented. I would assume everyone knew this as basic common-knowledge about
    > the SLR design since it's existed for so long. Except for those in the past
    > that used OTF (or off-the-curtain) metering, bypassing the metering done in
    > the pentaprism's light path. But even then there were still some
    > discrepancies due to the angle of the light-bundles coming from the lens
    > and hitting the film or shutter curtain at a sharper or less sharp angle,
    > and the amount of reflectivity of the surface from which the measure was
    > taken (patterned shutter-curtain, or different reflectivities of film
    > surfaces). This is the same reason a movie-projection screen, projection TV
    > system, even LCD displays, are not always the same intensity depending on
    > your angle of view to it.


    what a load of rubbish.

    > The problem is rather simple. If you have a wide-angle lens on the camera,
    > the light-cone coming from the back of that lens is very wide and shallow.
    > Light coming from the aperture will be striking the focusing screen, any
    > metering sensors, and even the CCD sensor itself at a sharper angle. (This
    > is why sensor blooming is more of a problem with wide-angle lenses and it
    > occurring more in the corners than the center of an image. This more
    > glancing light path spills over into adjoining pixels.)


    not true. what matters is the exit pupil, not the focal length, and
    even wide angle lenses can have their exit pupil fairly forward.
    olympus tried to make a big deal about it with their 'digital lenses.'

    > Conversely with a
    > telephoto lens the light-cone is narrow and long. Because the light-sensors
    > in the SLR design must be situated in a flat plane on the focusing screen
    > the sideways glancing light bundles from the wider and shallower cone from
    > a wide-angle lens won't all hit the exposure-metering sensors
    > perpendicularly. With a telephoto lens then they all hit those sensor spots
    > more perpendicularly.


    also not true.

    > The d/SLR's in-camera metering-system is calibrated/designed with this
    > inherent problem in mind. The amount of light hitting those sensors is
    > generally optimized for the angles of light coming from the back of a
    > "standard" 50-55mm lens. Above and below that focal-length of lens and the
    > light hitting those sensors won't be at the same angle.


    nonsense.

    > The only true cure for this design flaw is to get your metering direct from
    > the sensor which is capturing the image.


    nonsense.
    nospam, Aug 7, 2009
    #7
  8. eNo

    nospam Guest

    In article <>, Glen Calvers
    <> wrote:

    > >I'm assuming it would affect spot and center-weighed metering as well...

    >
    > Center-weighted yes, to a lesser degree. Spot metering no. Unless you
    > choose some metering spot in your FOV other than the exact center. Then
    > spot-metering from the edges will be just as prone to metering error caused
    > by the varying light-levels at the periphery from various focal-length
    > lenses as when using average and center-weighted. The d/SLR's metering
    > light-path has always been a huge compromise over accuracy and has not
    > changed. From vague memory, of the many articles published on this problem
    > during the 1970's when they were trying to find work-arounds in the SLR
    > camera design, I recall the differences in metering from a 24mm lens and a
    > 400mm lens could be as much as 2 (or more?) stops.


    nonsense.
    >
    > Much depending on the orientation of the metering sensors to the
    > image-plane as well as the focusing-screen's quality and fresnel lens
    > properties. Many, but not all, of today's cameras incorporate the metering
    > sensors right in the focusing-screen assembly itself, instead of using the
    > more error-prone design of reading the incident light off of the screen
    > using sensors at the sides of the pentaprism housing.


    no, they use the 'error prone' method which is not error prone at all,
    as can be seen by zillions of well exposed photos. if it's really 2 or
    more stops off, why do the photos require little to no manipulation?

    > In-screen sensors
    > lessening the problem but the problem still has not disappeared. It's only
    > become slightly less of an issue. This is the very reason they tried to
    > compensate for it by using in-screen sensor methods.
    >
    > One of those things that nobody wants to talk about for fear of losing
    > sales and converts. The end-user just chalks-up the metering errors in
    > their SLR design cameras to scene discrepancies or user error, then try to
    > compensate for it by forever needing RAW data to recover what their camera
    > failed to properly meter in the first place. This is why camera makers come
    > up with elaborate metering schemes like "evaluative" or "multi-program"
    > metering (automated "Scene" modes).


    nonsense. that's for real world conditions, such as backlighting or
    lots of sky or snow.

    > Trying to compare the light and dark
    > elements in the scene to a database of known common lighting situations and
    > adjusting accordingly. Basically trying to adjust your metering compared to
    > a database of average snapshots. The camera owner never realizing these
    > metering errors are caused by zooming their lens from 28mm to 200mm. Not
    > apparent to the user because the scene itself changes during that zooming
    > and usually requires a new exposure setting anyway, or due to the change in
    > available f/stops due to the longer focal-lengths being limited to smaller
    > apertures.


    and the camera compensates for the different aperture.

    > Few put two and two together today unless told to do so by the
    > manufacturer. Then they'll be assured by the manufacturer that two and two
    > is equal to three or five and it will be believed. It must be believed. Or
    > else you'll have to question, "you get what you pay for." Nobody wants to
    > go there after shelling out $5000 or more for body and lenses. They'd
    > rather believe that two and two equals three or five. It's easier to do
    > that. The comforting and pleasant bliss of ignorance is a mighty attractor
    > and panacea.


    except most people with dslrs don't spent anywhere near $5000.
    nospam, Aug 7, 2009
    #8
  9. The P&S troll "Glen Calvers" <> wrote:

    > You will find that


    "Glen Calvers" is the P&S troll.

    -Wolfgang
    Wolfgang Weisselberg, Aug 7, 2009
    #9
  10. On Sat, 8 Aug 2009 00:15:31 +0200, Wolfgang Weisselberg
    <> wrote:

    >The P&S troll "Glen Calvers" <> wrote:
    >
    >> You will find that

    >
    >"Glen Calvers" is the P&S troll.
    >
    >-Wolfgang


    Wolfgang is the resident virtual-photographer troll who can't refute one
    thing stated. This is all he has. Just like all trolls like him.
    Oh Look! Wolfgang The Moron!, Aug 8, 2009
    #10
  11. Glen Calvers <> wrote:
    > On 7 Aug 2009 10:10:58 GMT, Chris Malcolm <> wrote:


    >>Or to use metering sensors which have the same response to angled
    >>light as the image sensor. A perfect solution which would be easy to
    >>implement in a digital camera :)


    > You would need them to interact with angles of light coming from the lens
    > in the same contiguous area as the full sensor does. These sensors would
    > also have to be placed before the fresnel lens in the focusing-screen
    > because that lens' own light-cone is one of the main problems to cause this
    > metering problem in all d/SLR designs. It needs to be metered before
    > interacting with any other optics in the system. Otherwise you're back at
    > square one. Averaging and center-weighted is just that. It's a complete
    > average or calculated emphasis on one area but taking all the light hitting
    > that full sensor area into account. If you have tiny CCD chips with the
    > same micro-lenses and Bayer array on it to duplicate the exact same sensing
    > response elsewhere in the light path, then you'll still end up with lower
    > light levels in the corners and brighter in the center when using a
    > wide-angle lens, exacerbated even more if not done before the fresnel lens.
    > Taking an average or center-weighted metering from just a few sampled areas
    > so as not to impair the viewfinder's usefulness just won't cut it. Metering
    > off of the whole sensor without any other optics interfering is the only
    > viable solution for predictable accuracy no matter the focal-length of the
    > lens in use.


    The Sony A350 DSLR (and probably some other related models) has a
    separate image sensor for live view which is only switched into
    operation when live view is selected, so that you can use the LCD live
    view without losing the fast phase detection autofocus. When live view
    and matrix metering are in use it switches from the usual exposure
    sensors to using the live view sensor as an exposure meter. A number
    of people have commented on how extraordinarily accurately it seems to
    be able to nail exposure when using matrix metering and live view
    together. Having seen how well it works myself I now sometimes switch
    to live view before taking the shot just to get a more predictably
    accurate and consistent exposure in difficult conditions.

    Does that approach avoid the DSLR exposure metering problems you have
    been raising?

    --
    Chris Malcolm
    Chris Malcolm, Aug 8, 2009
    #11
  12. Glen Calvers <> wrote:
    > On 8 Aug 2009 09:51:43 GMT, Chris Malcolm <> wrote:


    >>
    >>The Sony A350 DSLR (and probably some other related models) has a
    >>separate image sensor for live view which is only switched into
    >>operation when live view is selected, so that you can use the LCD live
    >>view without losing the fast phase detection autofocus. When live view
    >>and matrix metering are in use it switches from the usual exposure
    >>sensors to using the live view sensor as an exposure meter. A number
    >>of people have commented on how extraordinarily accurately it seems to
    >>be able to nail exposure when using matrix metering and live view
    >>together. Having seen how well it works myself I now sometimes switch
    >>to live view before taking the shot just to get a more predictably
    >>accurate and consistent exposure in difficult conditions.
    >>
    >>Does that approach avoid the DSLR exposure metering problems you have
    >>been raising?


    > Can you direct me to a link showing a diagram of how those two systems are
    > implemented? I don't understand how phase focusing can remain engaged, it
    > being dependent on a separate light-path than the one leading to the sensor
    > and dependent on the mirror being down.


    When live view is engaged in the A350 the mirror stays down, and the
    light path from the mirror that leads to the viewfinder is diverted to
    a second lower resolution sensor the purpose of which is to drive
    the live view LCD. So phase detection autofocus remains engaged.

    > "Live view" metering, far as I
    > know, *is* the metering taken direct from the full sensor, just as in all
    > non-reflex digital camera designs. Unless Sony has implemented some other
    > way of metering when using the sensor for live-view.


    It uses a separate second sensor for the live view. It uses that
    sensor, when live view is engaged and it has been switched into the
    light path from the mirror, for matrix metering, in effect giving it a
    very much larger exposure sensing matrix. When live view is not
    engaged it uses the normal DSLR matrix exposure metering.

    > The accuracy that you are perceiving from the "live-view" metering is
    > precisely why I choose only cameras with that feature today. It perfectly
    > circumvents this problem of all reflex-camera designs. All non-reflex
    > digital cameras meter this way, this is why RAW from most of these cameras
    > becomes superfluous, they get the exposure right the first time. No need to
    > manipulate the data later to get a useful image from them.


    It seems Sony found a way round that problem.

    --
    Chris Malcolm
    Chris Malcolm, Aug 8, 2009
    #12
  13. eNo

    DEvON Guest

    On 8 Aug 2009 20:04:32 GMT, Chris Malcolm <> wrote:

    >Glen Calvers <> wrote:
    >> On 8 Aug 2009 09:51:43 GMT, Chris Malcolm <> wrote:

    >
    >>>
    >>>The Sony A350 DSLR (and probably some other related models) has a
    >>>separate image sensor for live view which is only switched into
    >>>operation when live view is selected, so that you can use the LCD live
    >>>view without losing the fast phase detection autofocus. When live view
    >>>and matrix metering are in use it switches from the usual exposure
    >>>sensors to using the live view sensor as an exposure meter. A number
    >>>of people have commented on how extraordinarily accurately it seems to
    >>>be able to nail exposure when using matrix metering and live view
    >>>together. Having seen how well it works myself I now sometimes switch
    >>>to live view before taking the shot just to get a more predictably
    >>>accurate and consistent exposure in difficult conditions.
    >>>
    >>>Does that approach avoid the DSLR exposure metering problems you have
    >>>been raising?

    >
    >> Can you direct me to a link showing a diagram of how those two systems are
    >> implemented? I don't understand how phase focusing can remain engaged, it
    >> being dependent on a separate light-path than the one leading to the sensor
    >> and dependent on the mirror being down.

    >
    >When live view is engaged in the A350 the mirror stays down, and the
    >light path from the mirror that leads to the viewfinder is diverted to
    >a second lower resolution sensor the purpose of which is to drive
    >the live view LCD. So phase detection autofocus remains engaged.
    >
    >> "Live view" metering, far as I
    >> know, *is* the metering taken direct from the full sensor, just as in all
    >> non-reflex digital camera designs. Unless Sony has implemented some other
    >> way of metering when using the sensor for live-view.

    >
    >It uses a separate second sensor for the live view. It uses that
    >sensor, when live view is engaged and it has been switched into the
    >light path from the mirror, for matrix metering, in effect giving it a
    >very much larger exposure sensing matrix. When live view is not
    >engaged it uses the normal DSLR matrix exposure metering.
    >
    >> The accuracy that you are perceiving from the "live-view" metering is
    >> precisely why I choose only cameras with that feature today. It perfectly
    >> circumvents this problem of all reflex-camera designs. All non-reflex
    >> digital cameras meter this way, this is why RAW from most of these cameras
    >> becomes superfluous, they get the exposure right the first time. No need to
    >> manipulate the data later to get a useful image from them.

    >
    >It seems Sony found a way round that problem.


    Now if they can just find a way around the slow and loud camera-shaking
    mirror and the slow flash-sync and loud slapping moving-subject-distorting
    shutter. But wait, they already have found a way around all of these
    drawbacks, they're available in all their non-dSLR designs. Granted, their
    dSLRs are some of the best on the market. Some of them addressing many of
    the drawbacks of the SLR design. Like inaccurate viewfinder framing and
    metering inaccuracies of all dSLRs. But they still come complete with the
    main drawbacks for which I will not go back to the SLR design ever again no
    matter who makes it.
    DEvON, Aug 8, 2009
    #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. Cliff Spicer

    S2Pro and Matrix Metering with Flash

    Cliff Spicer, Jul 22, 2003, in forum: Digital Photography
    Replies:
    0
    Views:
    352
    Cliff Spicer
    Jul 22, 2003
  2. Replies:
    3
    Views:
    459
    [BnH]
    Oct 26, 2005
  3. Doug Jewell

    Re: Nikon D90 defective Matrix metering

    Doug Jewell, Mar 28, 2009, in forum: Digital Photography
    Replies:
    17
    Views:
    2,710
  4. Rob Morley

    Re: Nikon D90 defective Matrix metering

    Rob Morley, Apr 7, 2009, in forum: Digital Photography
    Replies:
    18
    Views:
    608
    Paul Furman
    Apr 28, 2009
  5. Doug Jewell

    Re: Nikon D90 defective Matrix metering

    Doug Jewell, Apr 8, 2009, in forum: Digital Photography
    Replies:
    11
    Views:
    3,020
    Bruce
    Apr 12, 2009
Loading...

Share This Page