05-29-2011, 12:56 PM
[quote name='fall' timestamp='1306671360' post='8897']
i read that the focus shifting problem of the canon 50L was because of how the lens was designed, and to fix this the overall design had to change which would give us a mkII. was that not right?
good news if they found a way to fix this issue, may i know the date code on your 50L, just so that i know which batch doesn't have this issue. <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/smile.gif' class='bbc_emoticon' alt='
' />
[/quote]
Well, let me put this politely: there are too many people out there on the interweb who don't know what they are talking about, and/or parrot each other.
Essentially, a very fast lens by definition suffers from aspherical aberrations - that's Optics 201 (rather than 101). These aberrations are the same ones which generally provide the great, smooth bokeh many people are after. To some degree this can be mitigated by aspherical surfaces. However, aspherical surfaces really only work well specifically for a single aperture.
So, is this a design flaw? IMO, no, it is inherent to fast (large aperture) optics. However, what Canon engineers did here is create a lens that not only has very good bokeh behind the focal plane, but also in front of it, which is optically very, very difficult, if not virtually impossible. In short, this lens, from a bokeh POV and controlling aspherical aberrations, is at the bleeding edge of optical design. Now, in a lab environment, where each lens is hand calibrated to the n-th degree, it is probably possible to adjust each lens so that it is perfect. In an assembly plant, where 1000s of lenses are manufactured at speed, this may be a bit more compicated to obtain. It requires quite small tolerances to get such a design working properly, and once one realises what the tolerances involved are, it is really amazing that it is possible to create high quality lenses consistently in general.
It looks like it took some time before the plant was at a similar quality level as the labs, and it looks it has been there now for a few years with regard to the 50L.
As to the design 'flaws" that people talk about, they often refer to the 'floating element' that the 85L and 85L II have. Well, first of all, the 85L has floating element[size="3"]s[/size], plural IOW, not one, but all but one. The rear element is a static element, all other elements move ('float') when focusing. This is a design that has been used in the past mostly to prevent mechanical vignetting by the lens barrel and rear optics getting to far from the exit point when focusing up close, but according to Canon, it is used in the 85L to correct for coma, and likely also for astigmatism, which are abberrations with fast, (modified or not) double Gaussian designs such as the 50L and 85L are known to suffer from. In short, this, as in, floating elments, has nothing to do with correcting spherical aberrations - that is what the aspheric element(s) are for, and high refracive index glass elements.
Furthermore, after my own 50L was calibrated, I tested it against my own 85L II for residual focus shift (note: it didn't have noticeable focus shift to begin with, only consistent backfocus), as all fast lenses, even well corrected ones, will still have residual spherical aberrations and therefore residual focus shift. What I found is that both lenses had their strongest focus shift at F/2, at close focus distances (where it doesn't disappear in the DoF), where the 50L had about 1/6 stop of shift in the total DoF area, and the 85L II 1/4 stop. Both these values are well within the tolerances of fast AF points and the AF tolerances Canon promises (1/3 f-stop accuracy). But interestingly, the 85L II also has some focus shift, and even more than my 50L... <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/biggrin.gif' class='bbc_emoticon' alt='
' />
HTH, kind regards, Wim
i read that the focus shifting problem of the canon 50L was because of how the lens was designed, and to fix this the overall design had to change which would give us a mkII. was that not right?
good news if they found a way to fix this issue, may i know the date code on your 50L, just so that i know which batch doesn't have this issue. <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/smile.gif' class='bbc_emoticon' alt='
' />[/quote]
Well, let me put this politely: there are too many people out there on the interweb who don't know what they are talking about, and/or parrot each other.
Essentially, a very fast lens by definition suffers from aspherical aberrations - that's Optics 201 (rather than 101). These aberrations are the same ones which generally provide the great, smooth bokeh many people are after. To some degree this can be mitigated by aspherical surfaces. However, aspherical surfaces really only work well specifically for a single aperture.
So, is this a design flaw? IMO, no, it is inherent to fast (large aperture) optics. However, what Canon engineers did here is create a lens that not only has very good bokeh behind the focal plane, but also in front of it, which is optically very, very difficult, if not virtually impossible. In short, this lens, from a bokeh POV and controlling aspherical aberrations, is at the bleeding edge of optical design. Now, in a lab environment, where each lens is hand calibrated to the n-th degree, it is probably possible to adjust each lens so that it is perfect. In an assembly plant, where 1000s of lenses are manufactured at speed, this may be a bit more compicated to obtain. It requires quite small tolerances to get such a design working properly, and once one realises what the tolerances involved are, it is really amazing that it is possible to create high quality lenses consistently in general.
It looks like it took some time before the plant was at a similar quality level as the labs, and it looks it has been there now for a few years with regard to the 50L.
As to the design 'flaws" that people talk about, they often refer to the 'floating element' that the 85L and 85L II have. Well, first of all, the 85L has floating element[size="3"]s[/size], plural IOW, not one, but all but one. The rear element is a static element, all other elements move ('float') when focusing. This is a design that has been used in the past mostly to prevent mechanical vignetting by the lens barrel and rear optics getting to far from the exit point when focusing up close, but according to Canon, it is used in the 85L to correct for coma, and likely also for astigmatism, which are abberrations with fast, (modified or not) double Gaussian designs such as the 50L and 85L are known to suffer from. In short, this, as in, floating elments, has nothing to do with correcting spherical aberrations - that is what the aspheric element(s) are for, and high refracive index glass elements.
Furthermore, after my own 50L was calibrated, I tested it against my own 85L II for residual focus shift (note: it didn't have noticeable focus shift to begin with, only consistent backfocus), as all fast lenses, even well corrected ones, will still have residual spherical aberrations and therefore residual focus shift. What I found is that both lenses had their strongest focus shift at F/2, at close focus distances (where it doesn't disappear in the DoF), where the 50L had about 1/6 stop of shift in the total DoF area, and the 85L II 1/4 stop. Both these values are well within the tolerances of fast AF points and the AF tolerances Canon promises (1/3 f-stop accuracy). But interestingly, the 85L II also has some focus shift, and even more than my 50L... <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/biggrin.gif' class='bbc_emoticon' alt='
' />HTH, kind regards, Wim
Gear: Canon EOS R with 3 primes and 2 zooms, 4 EF-R adapters, Canon EOS 5 (analog), 9 Canon EF primes, a lone Canon EF zoom, 2 extenders, 2 converters, tubes; Olympus OM-D 1 Mk II & Pen F with 12 primes, 6 zooms, and 3 Metabones EF-MFT adapters ....