Sensor resolution, any sites with actual measurements?

Discussion in 'Digital Photography' started by Rich, Mar 5, 2006.

  1. Rich

    Rich Guest

    I'm not talking about sites like dpreview.com that might be testing
    sensors with less
    than perfect lenses that happen to be used with some cameras. What I
    mean are sites that have done tests with lenses that resolve (clearly)
    beyond the sensor's capability.
    No Canon 5D shots with wide open WA lenses, or related horrors.
    I'm curious about how (for e.g.) the resolution of a FF camera compares
    to cameras with smaller sensors, to see what effect pixel size (as
    opposed to pixel count) has on absolute resolution, if any. I know its
    not possible to avoid in camera processing which varies from brand to
    brand,
    but you have to start somewhere.
    My preconception (correct by Mr. Roger Clark) was that pixel counts
    determine absolute
    resolution. The more pixels you can use to image something, the more
    detail you will see.
    But this seems to vary due to a lot of other factors.
    -Pixel count
    -Pixel size
    -lens quality
    -in-camera conversion and processing
    -in-camera sharpening
    -post camera processing
    Even the image (it's characteristics) will effect different sensors
    differently.
     
    Rich, Mar 5, 2006
    #1
    1. Advertising

  2. "Rich" <> wrote:
    > I'm not talking about sites like dpreview.com that might be testing
    > sensors with less
    > than perfect lenses that happen to be used with some cameras. What I
    > mean are sites that have done tests with lenses that resolve (clearly)
    > beyond the sensor's capability.


    The dpreview test chart images are all taken with decent lenses at sensible
    f stops.

    David J. Littleboy
    Tokyo, Japan
     
    David J. Littleboy, Mar 5, 2006
    #2
    1. Advertising

  3. Rich

    Alfred Molon Guest

    In article <>,
    says...

    > No Canon 5D shots with wide open WA lenses, or related horrors.


    Speaking about this, are there lenses for full-frame DSLRs with corner-
    to-corner sharpness even wide open (and also at wide angle, why not)?
    --

    Alfred Molon
    ------------------------------
    Olympus 50X0, 7070, 8080, E300, E330 and E500 forum at
    http://groups.yahoo.com/group/MyOlympus/
    Olympus E500 resource - http://myolympus.org/E500/
     
    Alfred Molon, Mar 5, 2006
    #3
  4. Rich

    Helen Guest

    "Rich" <> wrote in message
    news:...
    > Even the image (it's characteristics) will effect different sensors


    Still spelling its with an apostrophe I see.

    And it's 'affect', not 'effect'.
     
    Helen, Mar 5, 2006
    #4
  5. In article <>,
    Alfred Molon <> wrote:
    >In article <>,
    > says...
    >
    >> No Canon 5D shots with wide open WA lenses, or related horrors.

    >
    >Speaking about this, are there lenses for full-frame DSLRs with corner-
    >to-corner sharpness even wide open (and also at wide angle, why not)?


    'Wide open' is a rather useless requirement. It is quite possible that an
    f/1.4 lens set to f/4 is sharper than an f/4 lens wide open. But the f/4
    lens is likely to be sharper wide open than the f/1.4 lens.

    If you look at MTF graphs, wide angle lenses simply don't have great
    corner sharpness. You want high-end telephoto lenses for that (or stop
    down far enough).


    --
    That was it. Done. The faulty Monk was turned out into the desert where it
    could believe what it liked, including the idea that it had been hard done
    by. It was allowed to keep its horse, since horses were so cheap to make.
    -- Douglas Adams in Dirk Gently's Holistic Detective Agency
     
    Philip Homburg, Mar 5, 2006
    #5
  6. Rich

    Alfred Molon Guest

    In article <>,
    says...
    >
    > "Rich" <> wrote in message
    > news:...
    > > Even the image (it's characteristics) will effect different sensors

    >
    > Still spelling its with an apostrophe I see.
    >
    > And it's 'affect', not 'effect'.


    Maybe Rich is not a native English speaker.
    --

    Alfred Molon
    ------------------------------
    Olympus 50X0, 7070, 8080, E300, E330 and E500 forum at
    http://groups.yahoo.com/group/MyOlympus/
    Olympus E500 resource - http://myolympus.org/E500/
     
    Alfred Molon, Mar 5, 2006
    #6
  7. Rich

    Alfred Molon Guest

    In article <snq7dt1v8fhbggo9if3rmj9ks3@inews_id.stereo.hq.phicoh.net>,
    says...

    > 'Wide open' is a rather useless requirement.


    Well no, you might have to use the lens at the widest aperture because
    there is not sufficient light.

    > If you look at MTF graphs, wide angle lenses simply don't have great
    > corner sharpness. You want high-end telephoto lenses for that (or stop
    > down far enough).


    Why can't they design WA lenses with corner to corner sharpness (even
    wide open)?
    --

    Alfred Molon
    ------------------------------
    Olympus 50X0, 7070, 8080, E300, E330 and E500 forum at
    http://groups.yahoo.com/group/MyOlympus/
    Olympus E500 resource - http://myolympus.org/E500/
     
    Alfred Molon, Mar 5, 2006
    #7
  8. In article <>, Alfred Molon
    <> writes
    >In article <>,
    >says...
    >>
    >> "Rich" <> wrote in message
    >> news:...
    >> > Even the image (it's characteristics) will effect different sensors

    >>
    >> Still spelling its with an apostrophe I see.
    >>
    >> And it's 'affect', not 'effect'.

    >
    >Maybe Rich is not a native English speaker.


    Unlikely, his English isn't that good! ;-)
    --
    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, Mar 5, 2006
    #8
  9. In article <>,
    Alfred Molon <> wrote:
    >In article <snq7dt1v8fhbggo9if3rmj9ks3@inews_id.stereo.hq.phicoh.net>,
    > says...
    >
    >> 'Wide open' is a rather useless requirement.

    >
    >Well no, you might have to use the lens at the widest aperture because
    >there is not sufficient light.


    If you have enough light to use a f/2.8 lens wide open, then you have
    enough light to use an f/1.4 at f/2.8 as well. Or you have to compare
    the f/1.4 wide open with the f/2.8 at a two stops higher ISO or a two stops
    slower shutter speed.

    >> If you look at MTF graphs, wide angle lenses simply don't have great
    >> corner sharpness. You want high-end telephoto lenses for that (or stop
    >> down far enough).

    >
    >Why can't they design WA lenses with corner to corner sharpness (even
    >wide open)?


    It is probably physics that gets in the way. But you have to define 'corner
    to corner sharpness'. Does that refer to an MTF curve that approaches a
    straight horizontal line (ignoring the actual MTF value) or does it mean a
    specific minimum MTF in the corners for a given resolution?


    --
    That was it. Done. The faulty Monk was turned out into the desert where it
    could believe what it liked, including the idea that it had been hard done
    by. It was allowed to keep its horse, since horses were so cheap to make.
    -- Douglas Adams in Dirk Gently's Holistic Detective Agency
     
    Philip Homburg, Mar 5, 2006
    #9
  10. Rich

    Don Stauffer Guest

    Unless there is an uncommon amount of crosstalk, most CCD chips come
    close to their theoretical capability, which is about 0.7 of the line
    frequency of the detectors. That is, if there are 100 detectors per mm,
    then the chip will usually test out at about 70 lines (35 lp) per mm.
    While the mtf curve for a discrete sensor is shaped a bit different than
    for a scanning sensor (Kell factor), the curves are close near the
    cutoff frequency or the 2% contrast point.

    Rich wrote:
    > I'm not talking about sites like dpreview.com that might be testing
    > sensors with less
    > than perfect lenses that happen to be used with some cameras. What I
    > mean are sites that have done tests with lenses that resolve (clearly)
    > beyond the sensor's capability.
    > No Canon 5D shots with wide open WA lenses, or related horrors.
    > I'm curious about how (for e.g.) the resolution of a FF camera compares
    > to cameras with smaller sensors, to see what effect pixel size (as
    > opposed to pixel count) has on absolute resolution, if any. I know its
    > not possible to avoid in camera processing which varies from brand to
    > brand,
    > but you have to start somewhere.
    > My preconception (correct by Mr. Roger Clark) was that pixel counts
    > determine absolute
    > resolution. The more pixels you can use to image something, the more
    > detail you will see.
    > But this seems to vary due to a lot of other factors.
    > -Pixel count
    > -Pixel size
    > -lens quality
    > -in-camera conversion and processing
    > -in-camera sharpening
    > -post camera processing
    > Even the image (it's characteristics) will effect different sensors
    > differently.
    >
     
    Don Stauffer, Mar 5, 2006
    #10
  11. Rich

    Helen Guest

    "Alfred Molon" <> wrote in message
    news:...
    >> Still spelling its with an apostrophe I see.
    >>
    >> And it's 'affect', not 'effect'.

    >
    > Maybe Rich is not a native English speaker.
    > --
    >

    No excuse (but he is anyway).
    He's been told how to spell 'its' before.
     
    Helen, Mar 5, 2006
    #11
  12. Rich

    Rich Guest

    I think some of the higher-end Canon L-telephotos are nearly as sharp
    wide open
    as stopped down because of precise engineering and the use of the best
    lens materials.
     
    Rich, Mar 5, 2006
    #12
  13. Rich

    Rich Guest

    If this is the case, then why don't digicams with small sensors and
    tiny pixels
    resolve more than DSLRs?
     
    Rich, Mar 5, 2006
    #13
  14. Rich wrote:
    > If this is the case, then why don't digicams with small sensors and
    > tiny pixels
    > resolve more than DSLRs?


    In lines per mm terms, they do.

    David
     
    David J Taylor, Mar 5, 2006
    #14
  15. Rich

    Neil Ellwood Guest

    On Sun, 05 Mar 2006 19:30:28 +0000, David J Taylor wrote:

    > Rich wrote:
    >> If this is the case, then why don't digicams with small sensors and
    >> tiny pixels
    >> resolve more than DSLRs?

    >
    > In lines per mm terms, they do.
    >
    > David

    But only on the sensor.
    --
    Neil
    Delete 'l' to reply by email
     
    Neil Ellwood, Mar 5, 2006
    #15
  16. Rich

    Neil Ellwood Guest

    On Sun, 05 Mar 2006 14:41:31 +0100, Alfred Molon wrote:

    > In article <snq7dt1v8fhbggo9if3rmj9ks3@inews_id.stereo.hq.phicoh.net>,
    > says...
    >
    >> 'Wide open' is a rather useless requirement.

    >
    > Well no, you might have to use the lens at the widest aperture because
    > there is not sufficient light.
    >
    >> If you look at MTF graphs, wide angle lenses simply don't have great
    >> corner sharpness. You want high-end telephoto lenses for that (or stop
    >> down far enough).

    >
    > Why can't they design WA lenses with corner to corner sharpness (even
    > wide open)?

    Ask a lens designer - they will probably tell you that it is a combination
    of physics, engineering tolerances and available glass specifications.

    --
    Neil
    Delete 'l' to reply by email
     
    Neil Ellwood, Mar 5, 2006
    #16
  17. Rich

    Alfred Molon Guest

    In article <h8EOf.11802$>,
    says...
    > Unless there is an uncommon amount of crosstalk, most CCD chips come
    > close to their theoretical capability, which is about 0.7 of the line
    > frequency of the detectors. That is, if there are 100 detectors per mm,
    > then the chip will usually test out at about 70 lines (35 lp) per mm.
    > While the mtf curve for a discrete sensor is shaped a bit different than
    > for a scanning sensor (Kell factor), the curves are close near the
    > cutoff frequency or the 2% contrast point.


    Where do you get this 70%? If a pattern with n white lines and n black
    lines falls exactly on the lines of the CCD, so that each CCD line sees
    either a black line or a white line, the CCD will have a resolution of
    100% of the nominal pixel count. Maximum MTF. On the other hand if each
    CCD line sees half of a black line and half of a white line, the CCD
    will see a uniform gray (zero MTF).
    --

    Alfred Molon
    ------------------------------
    Olympus 50X0, 7070, 8080, E300, E330 and E500 forum at
    http://groups.yahoo.com/group/MyOlympus/
    Olympus E500 resource - http://myolympus.org/E500/
     
    Alfred Molon, Mar 5, 2006
    #17
  18. Rich

    Don Stauffer Guest

    Neil Ellwood wrote:
    > On Sun, 05 Mar 2006 19:30:28 +0000, David J Taylor wrote:
    >
    >
    >>Rich wrote:
    >>
    >>>If this is the case, then why don't digicams with small sensors and
    >>>tiny pixels
    >>>resolve more than DSLRs?

    >>
    >>In lines per mm terms, they do.
    >>
    >>David

    >
    > But only on the sensor.

    But I thought that was what the question was about, resolution of the
    SENSOR.
     
    Don Stauffer, Mar 6, 2006
    #18
  19. Rich

    Don Stauffer Guest

    Neil Ellwood wrote:

    >>
    >>Why can't they design WA lenses with corner to corner sharpness (even
    >>wide open)?

    >
    > Ask a lens designer - they will probably tell you that it is a combination
    > of physics, engineering tolerances and available glass specifications.
    >

    As a one time lens designer, I'll take a poke at this. It is the physics
    and glasses, not tolerances.

    In many cases lens design is a situation of solving a set of equations
    with more variables than the number of equations- not enough degrees of
    freedom. Each and every field angle is a new design parameter. What
    provides proper correction at one field angle doesn't work at another
    field angle. Each field angle differing from the previous case by the
    desired resolution of the system is really a new case. The wider the
    total field of view, the greater the number of variables (field angle in
    this case), so the harder the job.

    One of the nasty things that happens is that near the edge of the field,
    the incidence angle on some surfaces gets very high. Surface reflection
    goes up as incidence angle goes up and one starts getting a lot of flare
    and ghost images, yet the greater the field angle, the harder it is to
    design baffling.

    Lens coatings only work well over a narrow range of angles.
     
    Don Stauffer, Mar 6, 2006
    #19
  20. Rich

    Don Stauffer Guest

    Alfred Molon wrote:
    > In article <h8EOf.11802$>,
    > says...
    >
    >>Unless there is an uncommon amount of crosstalk, most CCD chips come
    >>close to their theoretical capability, which is about 0.7 of the line
    >>frequency of the detectors. That is, if there are 100 detectors per mm,
    >>then the chip will usually test out at about 70 lines (35 lp) per mm.
    >>While the mtf curve for a discrete sensor is shaped a bit different than
    >>for a scanning sensor (Kell factor), the curves are close near the
    >>cutoff frequency or the 2% contrast point.

    >
    >
    > Where do you get this 70%? If a pattern with n white lines and n black
    > lines falls exactly on the lines of the CCD, so that each CCD line sees
    > either a black line or a white line, the CCD will have a resolution of
    > 100% of the nominal pixel count. Maximum MTF. On the other hand if each
    > CCD line sees half of a black line and half of a white line, the CCD
    > will see a uniform gray (zero MTF).


    Ah, but in a real resolution test there is NO guarantee that the lines
    in the object chart will fall exactly on the line! That is exactly the
    point. The kell factor and the 70% represent the PROBABILITY that the
    lines and detectors line up exactly.

    In real world imaging there is no way to ensure that exact line up in
    phase, or that the bar chart will be aligned exactly in angle to be
    parallel with rows or columns.
     
    Don Stauffer, Mar 6, 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. Mike

    Epson printers and lightfastness measurements

    Mike, Dec 15, 2003, in forum: Digital Photography
    Replies:
    13
    Views:
    554
    John Horner
    Dec 22, 2003
  2. Steven M. Scharf

    D-SLR Sensor Resolution and Sensor Size Comparison Size Matters!

    Steven M. Scharf, May 14, 2004, in forum: Digital Photography
    Replies:
    32
    Views:
    5,489
    Georgette Preddy
    May 16, 2004
  3. Chuck McDevitt

    Nikon D70 -- What are the actual color points for the sensor?

    Chuck McDevitt, Aug 8, 2004, in forum: Digital Photography
    Replies:
    6
    Views:
    394
    Mitch Alsup
    Aug 14, 2004
  4. RichA
    Replies:
    4
    Views:
    185
    RichA
    Aug 27, 2013
  5. Dale

    measurements to verify color workflow

    Dale, Mar 18, 2014, in forum: Digital Photography
    Replies:
    0
    Views:
    106
Loading...

Share This Page