Loading ...
Sorry, an error occurred while loading the content.
 

120 megapixels on APS-c

Expand Messages
  • Jeffrey Martin | 360Cities.net
    http://gizmodo.com/5620400/canon-decides-we-need-120-megapixel-sensors now we only need a lens in the world that can be used with such a sensor.......
    Message 1 of 23 , Aug 24, 2010
      http://gizmodo.com/5620400/canon-decides-we-need-120-megapixel-sensors



      now we only need a lens in the world that can be used with such a
      sensor.......


      [Non-text portions of this message have been removed]
    • Isaac Garcia
      Not that it makes that much of a difference in pixel density, but it´s APS-H, not C ;) (1.3x vs 1.6x). Now to wait for samples to check for noise. The claimed
      Message 2 of 23 , Aug 24, 2010
        Not that it makes that much of a difference in pixel density, but it´s
        APS-H, not C ;) (1.3x vs 1.6x).

        Now to wait for samples to check for noise.

        The claimed 9.5fps youd be great on my "measly" 21MP camera, though, hehe.
        Heck, I´d be more than happy with 7-8fps.


        Cheers from Madrid.




        On Tue, Aug 24, 2010 at 3:33 PM, Jeffrey Martin | 360Cities.net <
        360cities@...> wrote:

        > http://gizmodo.com/5620400/canon-decides-we-need-120-megapixel-sensors
        >
        >
        >
        > now we only need a lens in the world that can be used with such a
        > sensor.......
        >
        >
        > [Non-text portions of this message have been removed]
        >
        >
        >
        > ------------------------------------
        >
        > --
        >
        >
        >
        >


        --
        Isaac García

        http://www.coletassoft.com
        http://www.flickr.com/photos/coletassoft
        http://www.facebook.com/coletassoft
        http://twitter.com/coletassoft
        http://coletassoft.tumblr.com/
        http://coletassoft.deviantART.com

        JFGI | RTFM


        [Non-text portions of this message have been removed]
      • Isaac Garcia
        PS. So, how many shots to get a gigapixel with that? ;) On Tue, Aug 24, 2010 at 3:33 PM, Jeffrey Martin | 360Cities.net
        Message 3 of 23 , Aug 24, 2010
          PS. So, how many shots to get a gigapixel with that?

          ;)


          On Tue, Aug 24, 2010 at 3:33 PM, Jeffrey Martin | 360Cities.net <
          360cities@...> wrote:

          > http://gizmodo.com/5620400/canon-decides-we-need-120-megapixel-sensors
          >
          >
          >
          > now we only need a lens in the world that can be used with such a
          > sensor.......
          >
          >
          > [Non-text portions of this message have been removed]
          >
          >
          >
          > ------------------------------------
          >
          > --
          >
          >
          >
          >


          --
          Isaac García

          http://www.coletassoft.com
          http://www.flickr.com/photos/coletassoft
          http://www.facebook.com/coletassoft
          http://twitter.com/coletassoft
          http://coletassoft.tumblr.com/
          http://coletassoft.deviantART.com

          JFGI | RTFM


          [Non-text portions of this message have been removed]
        • Mark D. Fink
          If I read the specs correctly, you could do perspective control in video by using a very wide angle lens and choosing which part of the sensor to draw your
          Message 4 of 23 , Aug 24, 2010
            If I read the specs correctly, you could do perspective control in video by
            using a very wide angle lens and choosing which part of the sensor to draw
            your pixels from, right?

            Or, put two fisheye lenses back to back with right angle mirrors so you end
            up with two high resolution circular images on the one sensor. Never mind
            lens to sensor distance issues... Would be cool, eh?

            Mark
            www.pinnacle-vr.com
            www.northernlight.net
            www.virtual-travels.tumblr.com


            > -----Original Message-----
            > From: PanoToolsNG@yahoogroups.com [mailto:PanoToolsNG@yahoogroups.com] On
            > Behalf Of Isaac Garcia
            > Sent: Tuesday, August 24, 2010 10:45 AM
            > To: PanoToolsNG@yahoogroups.com
            > Subject: Re: [PanoToolsNG] 120 megapixels on APS-c
            >
            > PS. So, how many shots to get a gigapixel with that?
            >
            > ;)
            >
            >
            > On Tue, Aug 24, 2010 at 3:33 PM, Jeffrey Martin | 360Cities.net <
            > 360cities@...> wrote:
            >
            > > http://gizmodo.com/5620400/canon-decides-we-need-120-megapixel-sensors
            > >
            > >
            > >
            > > now we only need a lens in the world that can be used with such a
            > > sensor.......
            > >
            > >
            > > [Non-text portions of this message have been removed]
            > >
            > >
            > >
            > > ------------------------------------
            > >
            > > --
            > >
            > >
            > >
            > >
            >
            >
            > --
            > Isaac García
            >
            > http://www.coletassoft.com
            > http://www.flickr.com/photos/coletassoft
            > http://www.facebook.com/coletassoft
            > http://twitter.com/coletassoft
            > http://coletassoft.tumblr.com/
            > http://coletassoft.deviantART.com
            >
            > JFGI | RTFM
            >
            >
            > [Non-text portions of this message have been removed]
            >
            >
            >
            > ------------------------------------
            >
            > --
            >
            >
            >
          • robert
            Does no one read the details anymore? This is NOT a commercial product, even close to one and will be years if not a decade before something like this see the
            Message 5 of 23 , Aug 24, 2010
              Does no one read the details anymore?

              This is NOT a commercial product, even close to one and will be years if not a decade before something like this see the light of day in any commercial available camera.

              Note, Canon demonstrated a 50 megapixel sensor back in 2007 very similar to this. Do we have a 50 megapixel DSLR, - NO. When will we see a 50 megapixel DSLR ? Who knows, but it's probably not this year or the next. There is a lot of work to be done on this type of advancement to make it economical and commercially viable.

              I think Canon presented this to ensure no one forgets they are the current king of imaging sensors given that Nikon appears to be finally making or controlling their own sensors and Sony is starting to rock the two big guys boats big time.

              Now if only Red would actually produce a new camera - sigh.

              Regards,

              Robert
            • Keith Martin
              ... I don t think Red will ever be a mainstream manufacturer. The company is doing very well as a high-end niche specialist, and it would have to change in
              Message 6 of 23 , Aug 24, 2010
                Sometime around 24/8/10 (at 17:23 +0000) robert said:

                >Now if only Red would actually produce a new camera - sigh.

                I don't think Red will ever be a mainstream manufacturer. The company
                is doing very well as a high-end niche specialist, and it would have
                to change in huge ways to shift across to a more mass-market business.
                And if it did... how much would it cost, once you pick the various
                parts you'd need? (And if you take it out of the video arena what's
                so amazing about the resolution?)

                k
              • prague
                Robert, regardless of whether this is in the lab or in my local shop, the fact remains - there can t possibly be a lens in existence that allows this sensor to
                Message 7 of 23 , Aug 25, 2010
                  Robert, regardless of whether this is in the lab or in my local shop, the fact remains - there can't possibly be a lens in existence that allows this sensor to be used effectively.... is there? I wonder if there is any practical limit to the resolution of lenses? Or is it just a matter of money?

                  --- In PanoToolsNG@yahoogroups.com, "robert" <image360@...> wrote:
                  >
                  >
                  > Does no one read the details anymore?
                  >
                  > This is NOT a commercial product, even close to one and will be years if not a decade before something like this see the light of day in any commercial available camera.
                  >
                  > Note, Canon demonstrated a 50 megapixel sensor back in 2007 very similar to this. Do we have a 50 megapixel DSLR, - NO. When will we see a 50 megapixel DSLR ? Who knows, but it's probably not this year or the next. There is a lot of work to be done on this type of advancement to make it economical and commercially viable.
                  >
                  > I think Canon presented this to ensure no one forgets they are the current king of imaging sensors given that Nikon appears to be finally making or controlling their own sensors and Sony is starting to rock the two big guys boats big time.
                  >
                  > Now if only Red would actually produce a new camera - sigh.
                  >
                  > Regards,
                  >
                  > Robert
                  >
                • Ian Wood
                  ... If you limit it to APS/35mm equivalent cameras, no. 50MP digital SLRS? Sure there are. A digital back on a Hasselblad or similar is still an SLR. ;-) Ian
                  Message 8 of 23 , Aug 25, 2010
                    On 24 Aug 2010, at 18:23, robert wrote:

                    > Do we have a 50 megapixel DSLR, - NO.

                    If you limit it to APS/35mm 'equivalent' cameras, no. 50MP digital SLRS? Sure there are.

                    A digital back on a Hasselblad or similar is still an SLR. ;-)

                    Ian
                  • mrjimbo
                    Yes their seems to be a limit.. Kodak made a 10k linear chip that was to be used in a better light scan back among other ruses.. Betterlight delayed their back
                    Message 9 of 23 , Aug 25, 2010
                      Yes their seems to be a limit.. Kodak made a 10k linear chip that was to be used in a better light scan back among other ruses.. Betterlight delayed their back for a couple of years hoping that a lens would get developed to use it they finally came out with it using existing lens technology....Later Kodak discontinued the chip as no lenses were available nor could be made to effectively resolve it. Both Rodenstock and Schneider were involved.. Their are a few of these out there now but from what I understand don't really work anybetter then the 8k betterlight version.. This sensor would deliver a native file approx 768 MB if I recall but it was used in a 4x5. using less then the 4x5's film plane area.. so it's possible that 120 MP in an APS size sensor is at or beyond the resolving power of lenses. I'll follow that up with Mike Collette at Betterlight.. he'll know.. as he was on top of it for the back..

                      jimbo
                      ----- Original Message -----
                      From: prague
                      To: PanoToolsNG@yahoogroups.com
                      Sent: Wednesday, August 25, 2010 1:57 AM
                      Subject: [PanoToolsNG] Re: 120 megapixels on APS-c



                      Robert, regardless of whether this is in the lab or in my local shop, the fact remains - there can't possibly be a lens in existence that allows this sensor to be used effectively.... is there? I wonder if there is any practical limit to the resolution of lenses? Or is it just a matter of money?

                      --- In PanoToolsNG@yahoogroups.com, "robert" <image360@...> wrote:
                      >
                      >
                      > Does no one read the details anymore?
                      >
                      > This is NOT a commercial product, even close to one and will be years if not a decade before something like this see the light of day in any commercial available camera.
                      >
                      > Note, Canon demonstrated a 50 megapixel sensor back in 2007 very similar to this. Do we have a 50 megapixel DSLR, - NO. When will we see a 50 megapixel DSLR ? Who knows, but it's probably not this year or the next. There is a lot of work to be done on this type of advancement to make it economical and commercially viable.
                      >
                      > I think Canon presented this to ensure no one forgets they are the current king of imaging sensors given that Nikon appears to be finally making or controlling their own sensors and Sony is starting to rock the two big guys boats big time.
                      >
                      > Now if only Red would actually produce a new camera - sigh.
                      >
                      > Regards,
                      >
                      > Robert
                      >





                      [Non-text portions of this message have been removed]
                    • Bjørn K Nilssen
                      ... What exactly is it with lenses that prevents them from delivering the necessary resolution? Is it the glass composition? Glass finish/polish? Size of
                      Message 10 of 23 , Aug 25, 2010
                        On 25 Aug 2010 at 7:21, mrjimbo wrote:

                        > Yes their seems to be a limit.. Kodak made a 10k linear chip that was to be used in a
                        > better light scan back among other ruses.. Betterlight delayed their back for a couple
                        > of years hoping that a lens would get developed to use it they finally came out with it
                        > using existing lens technology....Later Kodak discontinued the chip as no lenses were
                        > available nor could be made to effectively resolve it. Both Rodenstock and Schneider
                        > were involved.. Their are a few of these out there now but from what I understand don't
                        > really work anybetter then the 8k betterlight version.. This sensor would deliver a
                        > native file approx 768 MB if I recall but it was used in a 4x5. using less then the
                        > 4x5's film plane area.. so it's possible that 120 MP in an APS size sensor is at or
                        > beyond the resolving power of lenses. I'll follow that up with Mike Collette at
                        > Betterlight.. he'll know.. as he was on top of it for the back..

                        What exactly is it with lenses that prevents them from delivering the necessary
                        resolution?
                        Is it the glass composition?
                        Glass finish/polish?
                        Size of photons?
                        Or something else?

                        --
                        Bjørn K Nilssen - http://bknilssen.no - panoramas and 3D
                      • Roger D. Williams
                        On Wed, 25 Aug 2010 22:30:58 +0900, Bjørn K Nilssen ... There are several factors at work. One of them will be familiar to you. You probably
                        Message 11 of 23 , Aug 25, 2010
                          On Wed, 25 Aug 2010 22:30:58 +0900, Bjørn K Nilssen <bk@...>
                          wrote:

                          > What exactly is it with lenses that prevents them from delivering the
                          > necessary
                          > resolution?
                          > Is it the glass composition?
                          > Glass finish/polish?
                          > Size of photons?
                          > Or something else?

                          There are several factors at work. One of them will be familiar to
                          you. You probably know that it is counter-productive to stop down
                          a lens with a very short focal length in the hope of increasing
                          the depth of field in focus. In fact, diffraction at the aperture
                          diaphragm causes blurring that represents a physical limit not
                          just for a specific lens but inherently for ALL lenses, although
                          some get closer to the theoretical limit than others. There are
                          other theoretical limits, like the physical consistency of the
                          glass used.

                          Roger W.

                          --
                          Business: www.adex-japan.com
                          Pleasure: www.usefilm.com/member/roger
                        • mrjimbo
                          I don t know the technical answer in the way that you ask it.. What I was told was that lens technology cannot resolve at that level of information.. the
                          Message 12 of 23 , Aug 25, 2010
                            I don't know the technical answer in the way that you ask it.. What I was told was that lens technology cannot resolve at that level of information.. the phrase that was used was "not enough resolving power".. I have a question in to someone that should know but their response may not be as technical as you'd like.

                            jimbo
                            ----- Original Message -----
                            From: Bjørn K Nilssen
                            To: PanoToolsNG@yahoogroups.com
                            Sent: Wednesday, August 25, 2010 7:30 AM
                            Subject: Re: [PanoToolsNG] Re: 120 megapixels on APS-c



                            On 25 Aug 2010 at 7:21, mrjimbo wrote:

                            > Yes their seems to be a limit.. Kodak made a 10k linear chip that was to be used in a
                            > better light scan back among other ruses.. Betterlight delayed their back for a couple
                            > of years hoping that a lens would get developed to use it they finally came out with it
                            > using existing lens technology....Later Kodak discontinued the chip as no lenses were
                            > available nor could be made to effectively resolve it. Both Rodenstock and Schneider
                            > were involved.. Their are a few of these out there now but from what I understand don't
                            > really work anybetter then the 8k betterlight version.. This sensor would deliver a
                            > native file approx 768 MB if I recall but it was used in a 4x5. using less then the
                            > 4x5's film plane area.. so it's possible that 120 MP in an APS size sensor is at or
                            > beyond the resolving power of lenses. I'll follow that up with Mike Collette at
                            > Betterlight.. he'll know.. as he was on top of it for the back..

                            What exactly is it with lenses that prevents them from delivering the necessary
                            resolution?
                            Is it the glass composition?
                            Glass finish/polish?
                            Size of photons?
                            Or something else?

                            --
                            Bjørn K Nilssen - http://bknilssen.no - panoramas and 3D





                            [Non-text portions of this message have been removed]
                          • mrjimbo
                            Got an answer back..not sure if I can share it properly but here goes.. The smallest pixel size to date is on the Canon G9.. they are about 2 microns.. They
                            Message 13 of 23 , Aug 25, 2010
                              Got an answer back..not sure if I can share it properly but here goes.. The smallest pixel size to date is on the Canon G9.. they are about 2 microns.. They made a lens specifically for this camera and it is capable of resolving the 2 micron sq pixels.. for the thumbnail size sensor. They figure this is the practical limit. The length of green light waves is about a half a micron so this will be the theoretical limit of any optic designed to work with visible light.
                              Regarding the larger sensors used in scan backs typically used in 4x5's or med format, the high end sensors their are typically 7 micron .. They use LF lenses and it is apparently presently not feasible to achieve that level of optical quality over that size lens's image circle.
                              So if I'm reading between the lines the answer is to go backwards.. larger sensor using larger formats and larger lenses with bigger image circles.

                              As a note the 10k sensor used in the Betterlight back was used over an image area of 72x96 mm and produced a file size of 794 MB.. Kodak pulled the plug on the sensor as the output could not apparently be effectively used so it had no market.
                              Realistically a lens that would step up to the plate or close that would work in an image circle to resolve 250 line pairs ... to us ea 2 micron pixel size is not feasible or it's viewed as cost prohibitive.. On the bright side I understand Obama has two of them on order.. (sorry just couldn't help it)



                              jimbo


                              ----- Original Message -----
                              From: mrjimbo
                              To: PanoToolsNG@yahoogroups.com
                              Sent: Wednesday, August 25, 2010 8:32 AM
                              Subject: Re: [PanoToolsNG] Re: 120 megapixels on APS-c



                              I don't know the technical answer in the way that you ask it.. What I was told was that lens technology cannot resolve at that level of information.. the phrase that was used was "not enough resolving power".. I have a question in to someone that should know but their response may not be as technical as you'd like.

                              jimbo
                              ----- Original Message -----
                              From: Bjørn K Nilssen
                              To: PanoToolsNG@yahoogroups.com
                              Sent: Wednesday, August 25, 2010 7:30 AM
                              Subject: Re: [PanoToolsNG] Re: 120 megapixels on APS-c

                              On 25 Aug 2010 at 7:21, mrjimbo wrote:

                              > Yes their seems to be a limit.. Kodak made a 10k linear chip that was to be used in a
                              > better light scan back among other ruses.. Betterlight delayed their back for a couple
                              > of years hoping that a lens would get developed to use it they finally came out with it
                              > using existing lens technology....Later Kodak discontinued the chip as no lenses were
                              > available nor could be made to effectively resolve it. Both Rodenstock and Schneider
                              > were involved.. Their are a few of these out there now but from what I understand don't
                              > really work anybetter then the 8k betterlight version.. This sensor would deliver a
                              > native file approx 768 MB if I recall but it was used in a 4x5. using less then the
                              > 4x5's film plane area.. so it's possible that 120 MP in an APS size sensor is at or
                              > beyond the resolving power of lenses. I'll follow that up with Mike Collette at
                              > Betterlight.. he'll know.. as he was on top of it for the back..

                              What exactly is it with lenses that prevents them from delivering the necessary
                              resolution?
                              Is it the glass composition?
                              Glass finish/polish?
                              Size of photons?
                              Or something else?

                              --
                              Bjørn K Nilssen - http://bknilssen.no - panoramas and 3D

                              [Non-text portions of this message have been removed]





                              [Non-text portions of this message have been removed]
                            • Erik Krause
                              ... There are lenses (and not necessarily the most expensive ones) which are limited by diffraction only. One of them is a russian 1000mm mirror lens, but this
                              Message 14 of 23 , Aug 25, 2010
                                Am 25.08.2010 15:43, schrieb Roger D. Williams:
                                > There are several factors at work. One of them will be familiar to
                                > you. You probably know that it is counter-productive to stop down
                                > a lens with a very short focal length in the hope of increasing
                                > the depth of field in focus. In fact, diffraction at the aperture
                                > diaphragm causes blurring that represents a physical limit not
                                > just for a specific lens but inherently for ALL lenses, although
                                > some get closer to the theoretical limit than others.

                                There are lenses (and not necessarily the most expensive ones) which are
                                limited by diffraction only. One of them is a russian 1000mm mirror
                                lens, but this one has only f/10 and hence isn't very sharp in total.

                                How good a lens actually is in terms of resolution is shown by the MTF
                                graph. Photodo.com publishes independent (AFAIK) lens tests with MTF
                                since years - real tests on an optical bench not image based ones like
                                most magazines and camera web sites do.

                                The best lens they tested was the Canon EF 200mm f/1.8. It reproduces
                                between 70% and 80% contrast at 40 line pairs per mm (lp/mm) at f/8
                                throughout the whole image (most lenses usually are softer towards the
                                edges). If I roughly extrapolate from the MTF lines for 10, 20 and 40
                                lp/mm this lens will still resolve 80 lp/mm at about 50% contrast (which
                                is still a good value. Some consumer zoom lenses resolve less than 50%
                                at 10 or 20 lp/mm).

                                You need at least two pixels to reproduce a line which makes 4 pixels
                                for a line pair or 320 pixels per mm at 80 lp/mm. Hence an APS-H sized
                                sensor (17x30mm) would need 52 monochrome megapixels to satisfy this
                                lens. Given the sensor is used with a bayer pattern you need triple this
                                number, which gives more than 150 megapixels.

                                This means that a 120 MP APS-H sensor isn't as good as the best lenses...

                                --
                                Erik Krause
                                http://www.erik-krause.de
                              • Bjørn K Nilssen
                                ... Interesting read :) But what is it that makes that lens so good? Glass formula, design, coating, surface finish etc ? -- Bjørn K Nilssen -
                                Message 15 of 23 , Aug 25, 2010
                                  On 25 Aug 2010 at 21:06, Erik Krause wrote:

                                  > Am 25.08.2010 15:43, schrieb Roger D. Williams:
                                  > > There are several factors at work. One of them will be familiar to
                                  > > you. You probably know that it is counter-productive to stop down
                                  > > a lens with a very short focal length in the hope of increasing
                                  > > the depth of field in focus. In fact, diffraction at the aperture
                                  > > diaphragm causes blurring that represents a physical limit not
                                  > > just for a specific lens but inherently for ALL lenses, although
                                  > > some get closer to the theoretical limit than others.
                                  >
                                  > There are lenses (and not necessarily the most expensive ones) which are
                                  > limited by diffraction only. One of them is a russian 1000mm mirror
                                  > lens, but this one has only f/10 and hence isn't very sharp in total.
                                  >
                                  > How good a lens actually is in terms of resolution is shown by the MTF
                                  > graph. Photodo.com publishes independent (AFAIK) lens tests with MTF
                                  > since years - real tests on an optical bench not image based ones like
                                  > most magazines and camera web sites do.
                                  >
                                  > The best lens they tested was the Canon EF 200mm f/1.8. It reproduces
                                  > between 70% and 80% contrast at 40 line pairs per mm (lp/mm) at f/8
                                  > throughout the whole image (most lenses usually are softer towards the
                                  > edges). If I roughly extrapolate from the MTF lines for 10, 20 and 40
                                  > lp/mm this lens will still resolve 80 lp/mm at about 50% contrast (which
                                  > is still a good value. Some consumer zoom lenses resolve less than 50%
                                  > at 10 or 20 lp/mm).
                                  >
                                  > You need at least two pixels to reproduce a line which makes 4 pixels
                                  > for a line pair or 320 pixels per mm at 80 lp/mm. Hence an APS-H sized
                                  > sensor (17x30mm) would need 52 monochrome megapixels to satisfy this
                                  > lens. Given the sensor is used with a bayer pattern you need triple this
                                  > number, which gives more than 150 megapixels.
                                  >
                                  > This means that a 120 MP APS-H sensor isn't as good as the best lenses...

                                  Interesting read :)
                                  But what is it that makes that lens so good?
                                  Glass formula, design, coating, surface finish etc ?

                                  --
                                  Bjørn K Nilssen - http://bknilssen.no - panoramas and 3D
                                • michael crane
                                  ... At a guess the accuracy of the chemical recipe and the annealing to make ideal crystal structure. -- mickiwiki.com
                                  Message 16 of 23 , Aug 25, 2010
                                    On Wed, August 25, 2010 10:26 pm, Bjørn K Nilssen wrote:

                                    > But what is it that makes that lens so good?
                                    > Glass formula, design, coating, surface finish etc ?

                                    At a guess the accuracy of the chemical recipe and the annealing to make
                                    ideal crystal structure.


                                    --
                                    mickiwiki.com
                                  • Jeff Smith
                                    This discussion of the limitations of optics is interesting, but doesn t address the problem inherent in very small pixel well site sizes - the fact that these
                                    Message 17 of 23 , Aug 26, 2010
                                      This discussion of the limitations of optics is interesting, but
                                      doesn't address the problem inherent in very small pixel well site sizes
                                      - the fact that these very small buckets collect very small samples to
                                      measure light intensities. The smaller the sample, the smaller the
                                      dynamic range capable of being produced. Of course, you can write
                                      algorithms to compensate until your blue in the face, but you're still
                                      creating data rather than recording it. Kind of like the Bayer filters
                                      themselves. We record only one data set for one color at each pixel
                                      site, the other two colors are created out of whole cloth by sampling
                                      the adjacent sites.
                                      If you've ever worked with true color captures (all 3 colors captures
                                      at each pixel site) you've seen the difference. It can be substantial.
                                      I'm not sure you'd really like the look that a sensor with well sites
                                      smaller than 2 microns would generate. Personally, I don't like the look
                                      generated by wellsites as small as 5 microns, but then I'm spoiled
                                      shooting in the studio with MF backs using 9 to 12 microns.
                                      --


                                      [Non-text portions of this message have been removed]
                                    • Erik Krause
                                      ... I guess the amount of correction needed to make a 200mm lens of that speed. At f/1.8 the entrance pupil diameter is 111mm! It contains three lenses made of
                                      Message 18 of 23 , Aug 26, 2010
                                        Am 25.08.2010 23:26, schrieb Bjørn K Nilssen:
                                        > Interesting read:)
                                        > But what is it that makes that lens so good?
                                        > Glass formula, design, coating, surface finish etc ?

                                        I guess the amount of correction needed to make a 200mm lens of that
                                        speed. At f/1.8 the entrance pupil diameter is 111mm! It contains three
                                        lenses made of ultra-low dispersion glass eliminating all chromatic
                                        aberration. I guess most expensive Zeiss or Contax lenses are at least
                                        as good. The second best lens in the photodo list is a Contax G Planar
                                        45mm f/2.

                                        I choose the Canon lens as an example because it a measured MTF graph is
                                        available and it is from Canon - showing that they are able to make
                                        lenses which are good enough for that sensor...

                                        --
                                        Erik Krause
                                        http://www.erik-krause.de
                                      • Erik Krause
                                        ... Well, the 120 MP sensor has 13,000 pixels at the long side. APS-H has 30mm, which gives still 2.3 microns pixel size... -- Erik Krause
                                        Message 19 of 23 , Aug 26, 2010
                                          Am 26.08.2010 18:56, schrieb Jeff Smith:

                                          > I'm not sure you'd really like the look that a sensor with well sites
                                          > smaller than 2 microns would generate.

                                          Well, the 120 MP sensor has 13,000 pixels at the long side. APS-H has
                                          30mm, which gives still 2.3 microns pixel size...

                                          --
                                          Erik Krause
                                          http://www.erik-krause.de
                                        • hd_de_2000
                                          ... and later ... If both the Canon at f/8 and the Russian lens (I guess you mean the Maksutov MTO 100/1000mm which I happen to own) were diffraction limited,
                                          Message 20 of 23 , Aug 30, 2010
                                            --- In PanoToolsNG@yahoogroups.com, Erik Krause <erik.krause@...> wrote:
                                            >
                                            > There are lenses (and not necessarily the most expensive ones) which are
                                            > limited by diffraction only. One of them is a russian 1000mm mirror
                                            > lens, but this one has only f/10 and hence isn't very sharp in total.

                                            and later

                                            > the Canon EF 200mm f/1.8. It reproduces
                                            > between 70% and 80% contrast at 40 line pairs per mm (lp/mm) at f/8

                                            If both the Canon at f/8 and the Russian lens
                                            (I guess you mean the Maksutov MTO 100/1000mm which I happen to own)
                                            were diffraction limited, there would be little difference
                                            since f/8 is not much wider than f/10. However, from my own experience I doubt that the MTO really is diffraction limited.


                                            Most lenses become diffraction limited if stopped down sufficiently, and f/8 is not very unusual.


                                            >
                                            > You need at least two pixels to reproduce a line which makes 4 pixels
                                            > for a line pair ...

                                            A line pair is a black and a white line (= one cycle). You need two pixels for that if the line is oriented parallel to the pixel pattern.


                                            > Given the sensor is used with a bayer pattern you need triple this
                                            > number, which gives more than 150 megapixels.
                                            >
                                            Depends on the color. The green pixels alone provide half the required
                                            resolution (i.e. you need twice the pixel count, not triple). A black/white pattern should be resolved even better, however with color errors.

                                            Helmut Dersch
                                          • Erik Krause
                                            ... Doesn t the mirror lens produces additional diffraction because of the donut shaped aperture? ... That was the statement of an employee of the ESO in
                                            Message 21 of 23 , Aug 30, 2010
                                              Am 30.08.2010 20:53, schrieb hd_de_2000:
                                              > If both the Canon at f/8 and the Russian lens (I guess you mean the
                                              > Maksutov MTO 100/1000mm which I happen to own) were diffraction
                                              > limited, there would be little difference since f/8 is not much wider
                                              > than f/10.

                                              Doesn't the mirror lens produces additional diffraction because of the
                                              donut shaped aperture?

                                              > However, from my own experience I doubt that the MTO really is
                                              > diffraction limited.

                                              That was the statement of an employee of the ESO in Munich. I didn't
                                              dare to doubt...

                                              > Most lenses become diffraction limited if stopped down sufficiently,
                                              > and f/8 is not very unusual.

                                              True. Nevertheless is the resolving power of the Canon lens exceptional.

                                              >>> You need at least two pixels to reproduce a line which makes 4
                                              >>> pixels for a line pair ...
                                              > A line pair is a black and a white line (= one cycle). You need two
                                              > pixels for that if the line is oriented parallel to the pixel
                                              > pattern.

                                              And if the line exactly hits a pixel line. If it is in between you get a
                                              uniform grey area instead of line pairs. Hence you need more pixels to
                                              reproduce the contrast of a line properly. That's how I understand it. I
                                              think it has something to do with Nyquist–Shannon sampling theorem but
                                              the math is too complicated for me...

                                              --
                                              Erik Krause
                                              http://www.erik-krause.de
                                            • hd_de_2000
                                              Sorry for starting a new thread but there is a problem with the reply function in Yahoo s software. ... No, you get a pattern with reduced contrast. It
                                              Message 22 of 23 , Aug 31, 2010
                                                Sorry for starting a new thread but there is a problem with the reply function in Yahoo's software.

                                                >And if the line exactly hits a pixel line. If it is in between you get a
                                                >uniform grey area instead of line pairs.

                                                No, you get a pattern with reduced contrast. It dissapears only if it is exactly in the center.

                                                >Hence you need more pixels to
                                                >reproduce the contrast of a line properly. That's how I understand it. I
                                                >think it has something to do with Nyquist–Shannon sampling theorem but
                                                the math is too complicated for me...

                                                The sampling frequency must be more than twice the base frequency, i.e. you need more than two samples (=pixels) per cycle (=linepair). But not four; 2.1 or 2.0001 or any epsilon above 2 is sufficient (in theory, there is a lot to say about the applicability to digital photography). Example: Digital audio uses 44.1kHz sampling frequency
                                                for 20kHz base frequency, i.e. a factor of ~2.2

                                                What we were discussing above is the pathological case 2.0

                                                Helmut Dersch
                                              • Erik Krause
                                                ... Ok, now I understand. Thanks for the clarification. However, my example with 80 lp/mm was far too low. Meanwhile I found evidence that many lenses can
                                                Message 23 of 23 , Aug 31, 2010
                                                  Am 31.08.2010 10:45, schrieb hd_de_2000:
                                                  > The sampling frequency must be more than twice the base frequency,
                                                  > i.e. you need more than two samples (=pixels) per cycle (=linepair).
                                                  > But not four; 2.1 or 2.0001 or any epsilon above 2 is sufficient (in
                                                  > theory, there is a lot to say about the applicability to digital
                                                  > photography). Example: Digital audio uses 44.1kHz sampling frequency
                                                  > for 20kHz base frequency, i.e. a factor of ~2.2

                                                  Ok, now I understand. Thanks for the clarification.

                                                  However, my example with 80 lp/mm was far too low. Meanwhile I found
                                                  evidence that many lenses can resolve 200 lp/mm and above. That alone
                                                  would justify a 120 MP APS-H sensor.

                                                  --
                                                  Erik Krause
                                                  http://www.erik-krause.de
                                                Your message has been successfully submitted and would be delivered to recipients shortly.