Re: Perfect lens

Discussion in 'Digital Photography' started by Martin Brown, Apr 2, 2013.

  1. Martin Brown

    Martin Brown Guest

    On 02/04/2013 09:08, Alfred Molon wrote:
    > Just wondering, if money played no role how good could lenses be?


    Lenses are typically made of glass and their chromatic aberrations can
    only be corrected exactly at two (achromat) or three (apochromat)
    wavelengths. If you allow the lens to be slow enough then it is possibel
    to make completely diffraction limited optics beyond about f5 these days
    for moderate to long focal lengths.
    >
    > No chromatic aberrations, no geometric distortions, huge sharpness from
    > corner to corner even wide open, or are there some physical constraints
    > which prevent from producing a perfect lens?
    >

    Mostly the fact that all glasses have a dispersive component. If you do
    away with the requirement of a flat focal plane then you can do much
    better with Schmidt camera designs of wide field extremely fast survey
    telescopes with something like a focal ratios of f3 although a few have
    been made with focal ratios less than 1.


    http://en.wikipedia.org/wiki/Schmidt_camera

    A derivative of this by Baker-Nunn was used to track and image
    satellites during the cold war. His obituary for super lenses:

    http://www.utsandiego.com/uniontrib/20050717/news_lz1j17baker.html

    A notable new all mirror design of telescope with a very fast ratio and
    essentially no aberrations is down to Willstrop of Cambridge using a
    three mirror design to get an f1.6 wide field diffraction limited scope.

    http://www.ast.cam.ac.uk/about/three-mirror.telescope

    That is about as close to a cost no object perfect lens as it is
    presently possible to manufacture. No use at all for an SLR though...

    A review (in French) of some of the fastest lenses ever made is online:

    http://www.dg77.net/photo/tech/fastex.htm

    Some of them are close to perfect diffraction limited performance when
    stopped down to normal working apertures.

    --
    Regards,
    Martin Brown
     
    Martin Brown, Apr 2, 2013
    #1
    1. Advertising

  2. Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    > On 02/04/2013 09:08, Alfred Molon wrote:


    >> Just wondering, if money played no role how good could lenses be?


    > Lenses are typically made of glass and their chromatic aberrations can
    > only be corrected exactly at two (achromat) or three (apochromat)
    > wavelengths.


    one, two, three wavelengths --- tell me why four, five, six
    aren't possible!


    >> No chromatic aberrations, no geometric distortions, huge sharpness from
    >> corner to corner even wide open, or are there some physical constraints
    >> which prevent from producing a perfect lens?


    > Mostly the fact that all glasses have a dispersive component.


    http://en.wikipedia.org/wiki/Negative_index_metamaterials
    does seem to open a few new ways.


    -Wolfgang
     
    Wolfgang Weisselberg, Apr 2, 2013
    #2
    1. Advertising

  3. Martin Brown

    RichA Guest

    On Apr 2, 2:12 am, Martin Brown <|||>
    wrote:
    > On 02/04/2013 09:08, Alfred Molon wrote:
    >
    > > Just wondering, if money played no role how good could lenses be?

    >
    > Lenses are typically made of glass and their chromatic aberrations can
    > only be corrected exactly at two (achromat) or three (apochromat)
    > wavelengths. If you allow the lens to be slow enough then it is possibel
    > to make completely diffraction limited optics beyond about f5 these days
    > for moderate to long focal lengths.
    >
    > > No chromatic aberrations, no geometric distortions, huge sharpness from
    > > corner to corner even wide open, or are there some physical constraints
    > > which prevent from producing a perfect lens?

    >
    > Mostly the fact that all glasses have a dispersive component. If you do
    > away with the requirement of a flat focal plane then you can do much
    > better with Schmidt camera designs of wide field extremely fast survey
    > telescopes with something like a focal ratios of f3 although a few have
    > been made with focal ratios less than 1.
    >
    > http://en.wikipedia.org/wiki/Schmidt_camera
    >
    > A derivative of this by Baker-Nunn was used to track and image
    > satellites during the cold war. His obituary for super lenses:
    >
    > http://www.utsandiego.com/uniontrib/20050717/news_lz1j17baker.html
    >
    > A notable new all mirror design of telescope with a very fast ratio and
    > essentially no aberrations is down to Willstrop of Cambridge using a
    > three mirror design to get an f1.6 wide field diffraction limited scope.
    >
    > http://www.ast.cam.ac.uk/about/three-mirror.telescope
    >
    > That is about as close to a cost no object perfect lens as it is
    > presently possible to manufacture. No use at all for an SLR though...
    >
    > A review (in French) of some of the fastest lenses ever made is online:
    >
    > http://www.dg77.net/photo/tech/fastex.htm
    >
    > Some of them are close to perfect diffraction limited performance when
    > stopped down to normal working apertures.
    >
    > --
    > Regards,
    > Martin Brown


    I've seen one of the Baker Super Schmidts in the flesh. 12"
    correctors, like two inverted punch bowls and a 24" primary mirror.
     
    RichA, Apr 2, 2013
    #3
  4. Martin Brown

    Martin Brown Guest

    On 02/04/2013 15:00, Wolfgang Weisselberg wrote:
    > Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    >> On 02/04/2013 09:08, Alfred Molon wrote:

    >
    >>> Just wondering, if money played no role how good could lenses be?

    >
    >> Lenses are typically made of glass and their chromatic aberrations can
    >> only be corrected exactly at two (achromat) or three (apochromat)
    >> wavelengths.

    >
    > one, two, three wavelengths --- tell me why four, five, six
    > aren't possible!


    Four is possible but it is rarely done in practice. You are into the law
    of diminishing returns and every air to glass surface loses some
    contrast. When it is they are called superachromats and typically extend
    to near infrared eg.

    http://en.wikipedia.org/wiki/Superachromat
    (delta f ~ 0.001

    For realistic choices of lens material there is little or nothing to be
    gained by using more than three wavelengths. You are stuck with the
    dispersive properties that the clear glasses happen to have.

    The last big thing was using calcium fluorite and now similarly behaved
    exotic glasses without the brittle handling characteristics of fluorite.

    Apochromats are usually as good as you need to get for photography

    http://en.wikipedia.org/wiki/Apochromat
    (delta f ~ 0.002 on one corrected for near IR as well

    Achromat is the basic first order correction right in red and blue.

    http://en.wikipedia.org/wiki/Achromat
    (delta f ~0.005

    They do tend to show purple fringing on high contrast highlights.

    >>> No chromatic aberrations, no geometric distortions, huge sharpness from
    >>> corner to corner even wide open, or are there some physical constraints
    >>> which prevent from producing a perfect lens?

    >
    >> Mostly the fact that all glasses have a dispersive component.

    >
    > http://en.wikipedia.org/wiki/Negative_index_metamaterials
    > does seem to open a few new ways.
    >
    >
    > -Wolfgang


    They may but they are unlikely to ever be used for optical components in
    the visible band.

    --
    Regards,
    Martin Brown
     
    Martin Brown, Apr 3, 2013
    #4
  5. Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    > On 02/04/2013 15:00, Wolfgang Weisselberg wrote:
    >> Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    >>> On 02/04/2013 09:08, Alfred Molon wrote:


    >>>> Just wondering, if money played no role how good could lenses be?


    >>> Lenses are typically made of glass and their chromatic aberrations can
    >>> only be corrected exactly at two (achromat) or three (apochromat)
    >>> wavelengths.


    >> one, two, three wavelengths --- tell me why four, five, six
    >> aren't possible!


    > Four is possible but it is rarely done in practice. You are into the law
    > of diminishing returns


    Money plays no role. (Nor does weight or carryability.)

    > and every air to glass surface loses some
    > contrast.


    Check out how many surfaces one of the 70-200 IS zoom lenses
    have ...

    > When it is they are called superachromats and typically extend
    > to near infrared eg.


    > http://en.wikipedia.org/wiki/Superachromat
    > (delta f ~ 0.001


    > For realistic choices of lens material there is little or nothing to be
    > gained by using more than three wavelengths. You are stuck with the
    > dispersive properties that the clear glasses happen to have.


    That's maybe a common sense and certainly a cost limitation;
    the OP threw out that concern out of the window.

    And the wikipedia link you gave says it's "highly beneficial"
    for certain uses.


    > The last big thing was using calcium fluorite and now similarly behaved
    > exotic glasses without the brittle handling characteristics of fluorite.


    > Apochromats are usually as good as you need to get for photography


    .... currently.

    But we're not concerned here with what you need, but what if
    you had infinite money.


    >>>> No chromatic aberrations, no geometric distortions, huge sharpness from
    >>>> corner to corner even wide open, or are there some physical constraints
    >>>> which prevent from producing a perfect lens?


    >>> Mostly the fact that all glasses have a dispersive component.


    >> http://en.wikipedia.org/wiki/Negative_index_metamaterials
    >> does seem to open a few new ways.


    > They may but they are unlikely to ever be used for optical components in
    > the visible band.


    Just as unlikely as people using transistors for photography.

    -Wolfgang
     
    Wolfgang Weisselberg, Apr 3, 2013
    #5
  6. Martin Brown

    Martin Brown Guest

    On 04/04/2013 10:29, Neil Ellwood wrote:
    > On Tue, 02 Apr 2013 16:00:55 +0200, Wolfgang Weisselberg wrote:
    >
    >> Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    >>> On 02/04/2013 09:08, Alfred Molon wrote:

    >>
    >>>> Just wondering, if money played no role how good could lenses be?

    >>
    >>> Lenses are typically made of glass and their chromatic aberrations can
    >>> only be corrected exactly at two (achromat) or three (apochromat)
    >>> wavelengths.

    >>
    >> one, two, three wavelengths --- tell me why four, five, six aren't
    >> possible!


    > We can normally see only three wavelengths (red, green and blue) and to
    > correct for more wavelengths would increase the cost and complexity by a
    > great deal.


    Actually the cones in our eyes are sensitive to roughly blue, green and
    yellow. Our perception of red is computed in the brain as yellow-green
    which means that you can play tricks on its perception of colour by
    using a strong minus yellow bandpass filter. See for example:

    http://openwetware.org/wiki/BIO254:Phototransduction

    A few people do have abnormal extended red sensitivity that allows them
    to see a difference between live and dead plant foliage. If you wrap
    some unexposed colour slide film against your eyes and wait for dark
    adaption to bring the scene back into colour you can boost the relative
    amount of IR detection (black slide film is transparent in the near IR
    and dense for visible). The result is you see live foliage as lighter.

    A good camera lens has to bring all light to a decent focus as otherwise
    you get the characteristic purple haloes around highlights. The latter
    being a signature of basic achromat designs at best focus.

    --
    Regards,
    Martin Brown
     
    Martin Brown, Apr 4, 2013
    #6
  7. Neil Ellwood <> wrote:
    > On Tue, 02 Apr 2013 16:00:55 +0200, Wolfgang Weisselberg wrote:
    >> Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:


    >>> Lenses are typically made of glass and their chromatic aberrations can
    >>> only be corrected exactly at two (achromat) or three (apochromat)
    >>> wavelengths.


    >> one, two, three wavelengths --- tell me why four, five, six aren't
    >> possible!


    > We can normally see only three wavelengths (red, green and blue) and to
    > correct for more wavelengths would increase the cost and complexity by a
    > great deal.


    That's may be an argument that it's not economic, not that it
    can't be done.

    In addition, while you may have only 3 colour sensor types in
    your eye, each of them react to a fairly wide frequency range.
    The same is true for your digital consumer camera sensors ---
    in fact, red pixels also (need to) react to blueish frequencies.

    So since we're not seeing and recording 3 single (or very
    narrow) wavelengths, both perfectly corrected and ... less
    perfectly corrected wavelengths overlap in any image formed
    even by apochromatic lenses ...

    -Wolfgang
     
    Wolfgang Weisselberg, Apr 5, 2013
    #7
  8. Martin Brown

    Paul Ciszek Guest

    In article <>,
    Alfred Molon <> wrote:
    >Perhaps another question: would they make a mirror lens for a camera and
    >the shape of the mirrors were perfect (assume zero error), no
    >displacement errors etc., would such a lens be perfect? Unlimited corner
    >to corner sharpness wide open, no geometric or chromatic aberrations?


    Undergraduate optics was a long time ago, but I seem to remember the textbook
    saying something along these lines: Even a mathematically perfect lens or mirror
    could only do one, not-very-useful thing perfectly: Bring a point source of light,
    on-axis, to a single point focus, also on-axis. If you use small-angle approximations
    and ignore the higher order terms, it turns out you can also sort of focus other
    points a little bit a way from the optical axis and thereby form an image. But
    imaging is by its very nature deviating from the mathematical idea into the realm
    of "how good is good enough?". I think our teacher said something to the effect that
    a non-spherical lens designed to perfectly focus an on-axis point source was actually
    *worse* for imaging purposes than a spherical lens; I don't recall what he said about
    mirrors specifically, but it was the same theme: Imaging is an apporoximation.

    The one thing that a mirror *will* do for you is kill chromatic aberration. That is
    why Newton was able to get such good results with his first attempt at a reflector
    telescope.

    --
    "Remember when teachers, public employees, Planned Parenthood, NPR and PBS
    crashed the stock market, wiped out half of our 401Ks, took trillions in
    TARP money, spilled oil in the Gulf of Mexico, gave themselves billions in
    bonuses, and paid no taxes? Yeah, me neither."
     
    Paul Ciszek, Apr 7, 2013
    #8
  9. Martin Brown

    Martin Brown Guest

    On 07/04/2013 08:38, Alfred Molon wrote:

    > Perhaps another question: would they make a mirror lens for a camera and
    > the shape of the mirrors were perfect (assume zero error), no
    > displacement errors etc., would such a lens be perfect? Unlimited corner
    > to corner sharpness wide open, no geometric or chromatic aberrations?


    There are always some small aberrations off axis but they can be kept
    very well controlled iff you are prepared to sacrifice having a flat
    focal plane. The best wide field diffraction limited optics all have one
    thing in common - a curved focal plane (usually spherical).

    Here is an example optimisation for a Schmidt telescope design:
    http://sitemaker.umich.edu/anupam.garge/files/me555_07_09_finalreport_final_final.pdf

    (you may not like the maths)

    Sharpness is always diffraction limited by the finite aperture of the
    lens - you can't get around that without an infinite lens. It is for
    this reason that astronomers are always building bigger telescopes.

    The atmosphere also limits resolution in long telephoto lenses.

    --
    Regards,
    Martin Brown
     
    Martin Brown, Apr 8, 2013
    #9
  10. Alfred Molon <> wrote:
    > Perhaps another question: would they make a mirror lens for a camera and
    > the shape of the mirrors were perfect (assume zero error), no
    > displacement errors etc., would such a lens be perfect?


    Take a look at the bokeh of a mirror lens to be cured of that
    idea. :)

    Of course, if your light rays are coming in perfectly parallel
    (i.e. infinite distance) it means there's no closer or further
    and thus no bokeh problems.

    -Wolfgang
     
    Wolfgang Weisselberg, Apr 10, 2013
    #10
  11. Martin Brown

    Martin Brown Guest

    On 10/04/2013 22:44, Wolfgang Weisselberg wrote:
    > Alfred Molon <> wrote:


    >> Perhaps another question: would they make a mirror lens for a camera and
    >> the shape of the mirrors were perfect (assume zero error), no
    >> displacement errors etc., would such a lens be perfect?

    >
    > Take a look at the bokeh of a mirror lens to be cured of that
    > idea. :)


    Only a problem where the image formed is out of focus.
    Donuts can look sometimes OK on the right subject.

    > Of course, if your light rays are coming in perfectly parallel
    > (i.e. infinite distance) it means there's no closer or further
    > and thus no bokeh problems.
    >
    > -Wolfgang


    If the target is all in focus then bokeh is not a problem.

    The diffraction pattern of a centre obstructed circular aperture is
    slightly different to that of an unobstructed one. Low to mid spatial
    frequencies are suppressed slightly which leads to a slightly higher
    effective resolution in the point spread coupled with a worse series of
    diffraction rings around it. A reasonable description of the behaviour
    and limitations of obstructed apertures in telescopes is online at

    http://www.telescope-optics.net/obstruction.htm

    It is worth pointing out here that almost all professionally used
    astronomical telescopes these days are obstructed aperture reflectors.

    The largest lens is in the Yerkes observatory made from a pair of 42"
    blanks the history of which includes land price bubbles and intrigue.

    http://astro.uchicago.edu/yerkes/history/1892.html

    You can't realistically make lenses any bigger - they sag under their
    own weight and cannot be mounted without stressing the component. Some
    of the earlier once worlds largest telescope lenses have suffered a form
    of surface recrystalisation hazing due to the experimental formulation
    of flint glass used in their doublet to correct chromatic abberation.
    Airy's designed 12" OG Northumberlan Telescope in Cambridge has
    undergone replacement with a modern lens after the flint component
    eventually became too thin to repolish out the damage.

    http://www.ast.cam.ac.uk/about/northumberland.telescope

    The website describes the old f20 Cauchoix lens as "not very good by
    modern standards" which is a little unfair. When it was new the glass
    was a lot clearer and it was still pretty impressive in the 1980's even
    though misting of the lens surface was detracting from image quality.

    --
    Regards,
    Martin Brown
     
    Martin Brown, Apr 12, 2013
    #11
  12. Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    > On 10/04/2013 22:44, Wolfgang Weisselberg wrote:
    >> Alfred Molon <> wrote:


    >>> Perhaps another question: would they make a mirror lens for a camera and
    >>> the shape of the mirrors were perfect (assume zero error), no
    >>> displacement errors etc., would such a lens be perfect?


    >> Take a look at the bokeh of a mirror lens to be cured of that
    >> idea. :)


    > Only a problem where the image formed is out of focus.


    Shooting brick walls and flat daily papers on the wall ...
    well, if you shoot only such flat objects ...

    > Donuts can look sometimes OK on the right subject.


    Yep, but they're eatable.


    >> Of course, if your light rays are coming in perfectly parallel
    >> (i.e. infinite distance) it means there's no closer or further
    >> and thus no bokeh problems.


    > If the target is all in focus then bokeh is not a problem.


    see above.

    > It is worth pointing out here that almost all professionally used
    > astronomical telescopes these days are obstructed aperture reflectors.


    It is worth pointing out that even the closest astronomical
    object, the moon, is very very far away --- practically
    infinitely far away --- compared to terrestrial photography
    objects.


    > The largest lens is in the Yerkes observatory made from a pair of 42"
    > blanks the history of which includes land price bubbles and intrigue.


    | The Great Paris Exhibition Telescope of 1900, with an objective lens
    | of 1.25 m (49.2 inches) in diameter, was the largest refracting
    | telescope ever constructed.
    http://en.wikipedia.org/wiki/Great_Paris_Exhibition_Telescope_of_1900

    49" > 42", if that source is correct.


    > You can't realistically make lenses any bigger - they sag under their
    > own weight and cannot be mounted without stressing the component.


    Diffractive optics. Look up Canon's DO lenses. That may
    well be a way to build larger lenses ... but they're no longer
    needed for telescopes, are they?

    Not that that matters in the least:
    Who would use such heavy and huge monsters for terrestrial
    photography? Telescopes, yes ... but in a photo newsgroup
    "a mirror lens for a camera" (OP) does not really hint at that.

    -Wolfgang
     
    Wolfgang Weisselberg, Apr 13, 2013
    #12
  13. Martin Brown

    Martin Brown Guest

    On 13/04/2013 00:01, Wolfgang Weisselberg wrote:
    > Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    >>
    >> It is worth pointing out here that almost all professionally used
    >> astronomical telescopes these days are obstructed aperture reflectors.

    >
    > It is worth pointing out that even the closest astronomical
    > object, the moon, is very very far away --- practically
    > infinitely far away --- compared to terrestrial photography
    > objects.


    Still close to infinity but the closest things that get routinely imaged
    are the space junk in near Earth orbit just a few tens or hundreds of
    miles up. Norad provides the orbital ephemeris.

    >> The largest lens is in the Yerkes observatory made from a pair of 42"
    >> blanks the history of which includes land price bubbles and intrigue.

    >
    > | The Great Paris Exhibition Telescope of 1900, with an objective lens
    > | of 1.25 m (49.2 inches) in diameter, was the largest refracting
    > | telescope ever constructed.
    > http://en.wikipedia.org/wiki/Great_Paris_Exhibition_Telescope_of_1900
    >
    > 49" > 42", if that source is correct.


    I was not aware that the lenses of the Great Paris scope had been found.
    They were still badly compromised by their own weight and the company
    that made it went bust leaving little trace of its work.

    --
    Regards,
    Martin Brown
     
    Martin Brown, Apr 16, 2013
    #13
  14. Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:
    > On 13/04/2013 00:01, Wolfgang Weisselberg wrote:
    >> Martin Brown <|||newspam|||@nezumi.demon.co.uk> wrote:


    >>> It is worth pointing out here that almost all professionally used
    >>> astronomical telescopes these days are obstructed aperture reflectors.


    >> It is worth pointing out that even the closest astronomical
    >> object, the moon, is very very far away --- practically
    >> infinitely far away --- compared to terrestrial photography
    >> objects.


    > Still close to infinity but the closest things that get routinely imaged
    > are the space junk in near Earth orbit just a few tens or hundreds of
    > miles up. Norad provides the orbital ephemeris.


    A few tens of miles? How about 62?


    >>> The largest lens is in the Yerkes observatory made from a pair of 42"
    >>> blanks the history of which includes land price bubbles and intrigue.


    >> | The Great Paris Exhibition Telescope of 1900, with an objective lens
    >> | of 1.25 m (49.2 inches) in diameter, was the largest refracting
    >> | telescope ever constructed.
    >> http://en.wikipedia.org/wiki/Great_Paris_Exhibition_Telescope_of_1900


    >> 49" > 42", if that source is correct.


    > I was not aware that the lenses of the Great Paris scope had been found.
    > They were still badly compromised by their own weight and the company
    > that made it went bust leaving little trace of its work.


    Still, compromised or not, the size is what matters.
    (Doesn't it always? "My lens is larger than your lens!")

    -Wolfgang
     
    Wolfgang Weisselberg, Apr 17, 2013
    #14
    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. A Cowan

    FireFox would be perfect if....

    A Cowan, Mar 8, 2005, in forum: Firefox
    Replies:
    6
    Views:
    401
    A Cowan
    Mar 9, 2005
  2. James Silverton

    Re: Perfect lens

    James Silverton, Apr 2, 2013, in forum: Digital Photography
    Replies:
    6
    Views:
    176
    Helpful person
    Apr 16, 2013
  3. RichA

    Re: Perfect lens

    RichA, Apr 2, 2013, in forum: Digital Photography
    Replies:
    0
    Views:
    149
    RichA
    Apr 2, 2013
  4. Paul Ciszek

    Re: Perfect lens

    Paul Ciszek, Apr 7, 2013, in forum: Digital Photography
    Replies:
    3
    Views:
    182
    J. Clarke
    May 8, 2013
  5. Mark Sieving

    Re: Perfect lens

    Mark Sieving, Apr 10, 2013, in forum: Digital Photography
    Replies:
    3
    Views:
    164
    Doug McDonald
    Apr 18, 2013
Loading...

Share This Page