 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
| PhotoKB HomeDiscussion Groups Digital Photography Digital PhotoDSLR CamerasZLR CamerasPoint & Shoot CamerasFilm Photography 35 mmLarge FormatMedium formatDarkroomFilm and LabsOther EquipmentPhoto Technique Nature PhotographyPeople PhotographyTechnique GeneralGeneral Photo Topics General TopicsAustralian PhotographyUK PhotographyDirectoryPhoto Clubs |
Photo Forum / Digital Photography / DSLR Cameras / March 2006Dark Corners on Canon 5D
Has anyone else noticed that on wide angle lenses on a Canon 5D the corners are dark from max aperture up to about f5.6? Once at f8, the problem goes away. I've noticed this on a 20mm f2.8 and a 28mm f2.8. I've checked with a 50mm f1.8 and the problem is not apparent. I noticed this today after taking some pictures in the snow. >Has anyone else noticed that on wide angle lenses on a Canon 5D the corners >are dark from max aperture up to about f5.6? Once at f8, the problem goes [quoted text clipped - 4 lines] > >I noticed this today after taking some pictures in the snow. It's call vignetting and is found in most wide angle lenses to some degree. http://www.vanwalree.com/optics/vignetting.html Don't worry about it. If it bothers you, fix it in Photoshop. You must be new to photography to now know what vignetting is. ********************************************************** "A combat photographer should be able to make you see the color of blood in black and white" David Douglas Duncan Speaking on why in Vietnam he worked only in black and white http://www.hrc.utexas.edu/exhibitions/online/ddd/ >>Has anyone else noticed that on wide angle lenses on a Canon 5D the >>corners [quoted text clipped - 24 lines] > he worked only in black and white > http://www.hrc.utexas.edu/exhibitions/online/ddd/ I understand what vignetting is but I've never seen it so bad. I'll check out fixing it in Photoshop though. > I understand what vignetting is but I've never seen it so bad. > I'll check out fixing it in Photoshop though. The only thing you can do to get rid of it is to crop the picture. Gerrit >>I understand what vignetting is but I've never seen it so bad. >>I'll check out fixing it in Photoshop though. > > The only thing you can do to get rid of it is to crop the picture. No, you can mathematically brighten the corners and/or darken the middle. To get an accurate measurement of the light fall-off, you can do what is caller a flat field. For example, go to a uniformly lit wall, move up close (not too close to cause a shadow), set the focus for the picture you are trying to correct (you want the wall out of focus). Take an image. Rotate the camera about the optical axis, and take another. Keep doing this until you have gone all the way around 360 degrees. Average the frames together, and this is the flat field. Next take each image and divide by the flat field (of course scaling the image data numbers so they remain within range). 16-bit data works well for this, 32-bit floating point even better. Almost ALL lenses have light fall off (vignetting). It becomes less with longer focal lengths, and is dependent on the design. Even my 500 mm f/4 L IS lens shows it. If you have a 35mm film camera, put the lens on the camera, open the shutter, and open the back of the camera and look at the lens from different angles looking through the camera from the back. On axis the aperture appears round. Off axis, it is an ellipse, thus reducing the light. That is a fundamental property of the geometry, so all lenses have light fall off. But more important is if a side of the lens mount intrudes on the aperture. You'll likely see that with wide angle lenses. Now stop down the lens to f/8, and you'll notice that the side of the lens probably doesn't block some of the aperture. But you still have the elliptical aperture shape off axis so you still have light fall-off. Roger > Almost ALL lenses have light fall off (vignetting). But most aren't this obvious about it in the final results.  SignatureStacey >>Almost ALL lenses have light fall off (vignetting). > > But most aren't this obvious about it in the final results. Huh? I never said anything about what is obvious. It is a fact that all lenses have light fall-off. It can be measured. I have observed it in film images and in digital with the same lenses. It does not appear more or less in my DSLR images, but about the same (hard comparison due to contrast differences between film and digital). Roger >> Stacey wrote: >>> Roger N. Clark (change username to rnclark) wrote: [quoted text clipped - 12 lines] >> >> Roger In days gone by, the use of a centre spot ND filter was sometimes needed to fix the problem. In fact my Fuji GWS 617 absolutely needs one to take clear pictures. It's not big deal yet everyone seems to think it's some sort of failing. If the sensors were round, maybe the problem would not be there. Signaturewww.photosbydouglas.com www.weprint2canvas.com If you really must write,use my name at an above domain. > If the sensors were round, maybe the problem would not be there. Nitwit. It would just transform the problem from "dark corners" to "dark edges". Thinking can't be that hard, not even in Australia. >> D-Mac wrote: >> >>> If the sensors were round, maybe the problem would not be there. >> >> Nitwit. It would just transform the problem from "dark corners" to >> "dark edges". Thinking can't be that hard, not even in Australia. Well ...for the average fuckwit like yourself thinking outside the circle - out into a rectangle is beyond you grasp, I imagine? The sensor is not a square, it's a rectangle and it is this rectangular shape past the square of the thing which causes vignetting in the corners. Even us Aussies know that much. Signaturewww.photosbydouglas.com www.weprint2canvas.com If you really must write,use my name at an above domain. > Huh? I never said anything about what is obvious. > It is a fact that all lenses have light fall-off. [quoted text clipped - 5 lines] > > Roger Hi Roger: This is OT to general photography, but truly zero-falloff wide angle lenses are becoming increasingly common mainly as projection lenses. A common feature is perfect telecentricity coupled with very low pupil aberration. If you think about it, a perfectly telecentric lens with zero pupil aberration must have zero falloff because the ray bundles striking the image plane all have identical cone angles. Add a little of the right sort of pupil aberration and you can even have have a condition where the corners are actually brighter than the center. I've encountered this on many occasions, particularly in systems that include a relay, even when geometrical distortion is zero. Your "flat field" experimental approach to measuring falloff is very interesting, and I'm sure I'll use it at some point. Brian > Your "flat field" experimental approach to measuring falloff is very > interesting, and I'm sure I'll use it at some point. Flat-fielding is as common as spit in the astronomical imaging area. It is, in fact, a necessity for precision. Even with an optic that has a perfectly flat field, one can't assume that the gain of all the pixels are the same. > This is OT to general photography, but truly zero-falloff wide angle > lenses are becoming increasingly common mainly as projection lenses. A [quoted text clipped - 6 lines] > I've encountered this on many occasions, particularly in systems that > include a relay, even when geometrical distortion is zero. I do not see how telecentricity solves the light fall-off problem. The light rays entering the lens off axis must pass through the aperture at an angle, and the projected area of the front element will also be an ellipse, so the light gathered by the lens is a function of angle off-axis. The light fall-off problem begins at the front element before any light enters a lens. Roger >I do not see how telecentricity solves the light fall-off problem. >The light rays entering the lens off axis must pass through [quoted text clipped - 3 lines] >problem begins at the front element before any light enters a >lens. That is one reason why the front element of many wide angles is so large and curved - not all of the front element is collecting light for every point in the image, as is often the case with longer focal lengths. For example, a 16mm f/2.8 lens only needs a front aperture of 6mm, but all 16mm f/2.8 lenses are much larger. The lens stop is the defining aperture, not the objective diameter, and the front of the lens images this stop onto a near hemispherical 'surface', so that it appears to have the same size irrespective of field angle, preventing light fall-off at the front of the lens. At the rear of a telecentric lens, the stop sits directly above each point in the image, again being smaller than the diameter of the rear element which is, of necessity, rather large for a telecentric design. However, the same effect can be achieved without telecentric designs using similar techniques to those for the front of the lens. Of course, such alternative approaches only compensate for the solid angle of the lens stop at each field position, it does not compensate for the alleged difference in sensitivity between digital sensors and film to incidence angle - but we all know that is negligible bunk anyway. SignatureKennedy 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) > That is one reason why the front element of many wide angles is so large > and curved - not all of the front element is collecting light for every [quoted text clipped - 10 lines] > achieved without telecentric designs using similar techniques to those > for the front of the lens. I understand the large front element, but light from different parts of the scene MUST go through the aperture at different angles and that causes the light fall-off. I have never seen a lens design that didn't have light fall-off for off-axis rays. If all rays went through the aperture perpendicular to the aperture stop, they couldn't be angularly separated on the other side. Roger > I understand the large front element, but light from different > parts of the scene MUST go through the aperture at different > angles and that causes the light fall-off. I have never seen > a lens design that didn't have light fall-off for off-axis rays. > If all rays went through the aperture perpendicular to the > aperture stop, they couldn't be angularly separated on the other side. The image of the stop does not always have to be perpendicular to the optical axis. That is, the entrance pupil, which is the image of the stop formed by the elements in front of it, does not have to be a perfect ellipse when seen from an off-axis angle. I think this type of pupil tilting is one of the things Brian was referring to when he mentioned pupil aberrations. Stuff like this gets more important as you go farther off axis and paraxial assumptions don't hold. On looking at some wide angle lenses, I can't persuade myself that pupil tilt is actually measurable, although in a fisheye lens the apparent position of the pupil is clearly moving around. The foreshortening of the pupil only contributes one power of cos(theta) to the light fall off, which is one reason real lenses can do better than the cos^4(theta) rule-of-thumb. Neblette's book "Photographic Lenses" says of retrofocus wide-angles, "This type of lens consists of a negative component ... followed by a positive, image-forming component ..... the negative power of the front component helps greatly in reducing the difference in illumination between the margins and the center of the image." This isn't a mathematical book and he doesn't provide details. I wrote in reponse to Roger Clark: > On looking at some wide angle lenses, I can't persuade myself > that pupil tilt is actually measurable, although in a fisheye lens > the apparent position of the pupil is clearly moving around. I tried this again and could definitely see pupil tilting, or something like it. The first time, I had the lens wide open, and the largest effect is that the round pupil turns into a cat eye shape as I view off axis. But this is mostly due to physical vignetting. If I stop the lens down to about f/5.6-8, the pupil seems to rotate to follow the off axis angle, reducing foreshortening. Try it - take a retrofocus wideangle (I used 21-28mm), stop it down a few stops, and look through the front while illuminating it from the back. When I tried a non-retrofocus WA, the pupil seemed to foreshorten in a more normal way, which supports the idea that the large negative front element of a retrofocus WA is doing something that reduces light falloff. >>I understand the large front element, but light from different >>parts of the scene MUST go through the aperture at different [quoted text clipped - 28 lines] > This isn't a mathematical book and he doesn't provide > details. I agree, but the key point is it reduces the effect, but does not eliminate it. All lens designs I have seen have measurable light fall-off. Roger >> That is one reason why the front element of many wide angles is so >>large and curved - not all of the front element is collecting light [quoted text clipped - 17 lines] >If all rays went through the aperture perpendicular to the >aperture stop, they couldn't be angularly separated on the other side. Of course they can Roger. The stop is just that, a defining aperture. In a simple lens system it would be, as you note, presenting a different solid angle at the field corners from the solid angle it presents on axis. If this was the only solution then all lenses of a particular field of view would have the same light fall off pattern. In addition, the fall-off would be the same for all apertures of all lenses with the same field of view. Neither is true, even a single lens has a different light fall off as the aperture changes. Contrary to what you say above, there is no optical requirement for stop to be perpendicular to the optic axis for every field angle, and it is fairly easy to check just by looking at the aperture on a wide angle lens as you position your eye at different field angles. Depending on the lens, at the widest apertures some other areas in the lens body will restrict the stop size, so you might have to close down by a stop or two. However you can easily see, particularly with extreme wide angle lenses, that the stop keeps its shape and size much better as you view off axis than a simple aperture would. SignatureKennedy 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) >it is fairly easy to check just by looking at the aperture on a wide >angle lens as you position your eye at different field angles. [quoted text clipped - 3 lines] >extreme wide angle lenses, that the stop keeps its shape and size much >better as you view off axis than a simple aperture would. Just as an update to that, Roger, I was referring to the shape and size of the stop when viewed from the front of the lens. At the time of writing I had checked it on my old Olympus 18mm f/3.5 and saw the stop "rotating" with field angle to keep its solid angle the same. I have just checked it with my Canon 16-35 f/2.8 zoom, set at 16mm. With that lens, the stop actually appears to *increase* in area at the extremes of field angle due to the effects of the front elements, becoming quite elliptical with major (not minor!) axis in the direction of the field angle. This probably partially offsets other light fall off effects at the rear of the lens, such as the departure from telecentricity. If I recall correctly from previous discussions, you have this lens yourself, so you should be able to easily confirm it. SignatureKennedy 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) >>> That is one reason why the front element of many wide angles is so >>> large and curved - not all of the front element is collecting light [quoted text clipped - 36 lines] > lenses, that the stop keeps its shape and size much better as you view > off axis than a simple aperture would. The point is that optical designs can reduce the light fall-off problem, but not eliminate it. If you know of a lens design with no light fall-off, I would sure be interested, as I have yet to see one. Please provide a reference. Roger >>>> That is one reason why the front element of many wide angles is so >>>>large and curved - not all of the front element is collecting light [quoted text clipped - 40 lines] >I would sure be interested, as I have yet to see one. >Please provide a reference. I didn't say that any optical design eliminated it - but there are several ways of *reducing* the light fall off due to the cosine mechanism of the front aperture that you were referring to. As I suggested in a follow-up post, pick up a super-wide angle lens and look at how the lens stop appears from the front at different field angles and you will physically see one method of how the effect you raised is compensated for. SignatureKennedy 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) Roger: You have to think in terms of the exit pupil. When I mentioned zero pupil aberrations I should have said zero exit pupil aberrations. I can't say it any plainer: a telecentric lens with zero *exit pupil* aberrations must have exactly zero illumination falloff because the converging ray cones have exactly the same cone angle regardless of image position. What does "zero exit pupil aberrations" mean? In this context it means that the exit pupil has exactly the same size and shape regardless of image height. What happens with the entrance pupil is another matter. In a distortionless wide angle lens without falloff the entrance pupil gets bigger, mostly in the tangential direction. I've designed a number of high-end projection lenses of this general type that have gone into production, and they really do have extremely uniform illumination. > Roger: > You have to think in terms of the exit pupil. When I mentioned zero [quoted text clipped - 13 lines] > type that have gone into production, and they really do have extremely > uniform illumination. You say extremely uniform illumination, but how uniform? I have yet to see a lens with no light fall-off, and I would be quite interested if one actually exists (scientific applications). Do you have a reference to such a lens? I can find references to many lenses with "minimal light fall-off," but not zero. Roger "You say extremely uniform illumination, but how uniform? I have yet to see a lens with no light fall-off, and I would be quite interested if one actually exists (scientific applications). Do you have a reference to such a lens? I can find references to many lenses with "minimal light fall-off," but not zero. " Roger: Last year I published a paper showing falloff data for an ultrawide-ultrafast lens: J. B. Caldwell, "Lens Systems with Large Optical Invariant", SPIE 5865, pp. 126-135. This design actually had an increase in illumination in the corners. If you have lens design software I can give you examples rectilinear lenses with a few percent or less illumination variation across a field of 80 degrees or so. Its possible to get the variation under a percent, although slight pupil aberrations tend to make the illumination curve depart from absolute flatness. Zemax shows a perfectly flat illumination curve for a paraxial lens with the stop located one focal length in front of the lens to make it telecentric. Many telecentric projections lenses for DMD or LCD light engines will have better than 90% illumination uniformity. Any store that sells DMD rear projector sets will have them on display. Of course, the screen may wind up dominating the illumination characteristics in this case. Brian > Roger: > Last year I published a paper showing falloff data for an > ultrawide-ultrafast lens: > > J. B. Caldwell, "Lens Systems with Large Optical Invariant", SPIE 5865, > pp. 126-135. Brian, This is good news. Can you send me a pdf or a reprint? Are any lenses with less than 1% light fall-off in commercial production? Roger "Brian, This is good news. Can you send me a pdf or a reprint? Are any lenses with less than 1% light fall-off in commercial production? " Roger: PDF sent. Most of the low-falloff lenses I've been involved with have all been small volume production for special applications, so they would be hard to get and probably over-priced. One of my zero-distortion designs that probably will be produced in higher volume has about a 3% falloff over an 80-degree diagonal field of view, which may be too much for your application. Years ago USPL (since sold a couple of times) produced projection lens for the now-extinct two-chip DLP light engine, and it had very low illumination falloff. I assume that similar lenses for single-chip and 3-chip projectors have similar falloff characteristics. One issue that I haven't really considered is slight falloff variations due to changes in coating efficiency with incidence angle. If you really do need a flat illumination profile then this could be a significant problem, and it is a very difficult thing to model accurately. Why do you need lenses with less than 1% falloff? Brian > "Brian, > This is good news. Can you send me a pdf or a reprint? Are any lenses [quoted text clipped - 23 lines] > > Brian Brian, Thanks, I got the pdf. That is a very interesting paper. A lens covering 35mm, of say 15 mm focal length looks like it would be quite large, rivaling perhaps large telephoto lenses. It also looks like the back focus would be difficult for a 15 mm lens on a 35 mm camera, but perhaps you have a design for that situation. But I see how such lenses are now possible. Nice job. I have a couple of applications. One, for example, is I measured the spectral reflectance of the color targets on the Mars rovers in my lab in an environment chamber at martian temperatures. Uniformity of the light field was a major issue in the tests and made testing setup and measurement very complex. lenses with no light falloff would help, both for the illuminator and for the light source. Then, sometimes (actually a lot) I work in the infrared, so the optical system must work from 0.3 to 2.5 microns, and sometimes from 0.2 to 200 microns, so we use mirrors. So can you do it with mirrors? ;-) So I guess I should modify my original statement that ALL lenses show light fall off to all currently available consumer camera lenses show light fall off, but it is scientifically possible to design a zero fall-off lens, and some low fall-off lenses have been built. Would you agree with that? Roger > Brian, > Thanks, I got the pdf. That is a very interesting paper. A lens [quoted text clipped - 3 lines] > lens on a 35 mm camera, but perhaps you have a design for that situation. > But I see how such lenses are now possible. Nice job. Thanks for your comments. I have used a relay approach for wide angle lenses covering nearly 24x36mm, but size is a major issue. It makes most large telephoto lenses look pretty small. > I have a couple of applications. One, for example, is I measured > the spectral reflectance of the color targets on the Mars rovers in my [quoted text clipped - 6 lines] > 0.2 to 200 microns, so we use mirrors. So can you do it with > mirrors? ;-) There must be some good all-mirror solutions available, but I have less experience with pure catoptric systems. For 1:1 imaging a Dyson catadioptric system can be very good, although it may not fit your test setup and it does require some refractive elements. Perhaps a small angle unobscured system might work pretty well. Falloff in Cassegrain type telescope objectives is mainly due to vignetting caused by the obscuration. > So I guess I should modify my original statement that ALL > lenses show light fall off to all currently available consumer > camera lenses show light fall off, but it is scientifically possible > to design a zero fall-off lens, and some low fall-off lenses > have been built. Would you agree with that? I would agree, with the addition that consumer projection lenses also have low falloff. Its one of those high performance optical things that many people own without realizing it, like the diffraction-limited high-NA read lenses in CD players. > Roger Brian >> "Brian, >> This is good news. Can you send me a pdf or a reprint? Are any lenses [quoted text clipped - 50 lines] > >Roger Unless there is some savings to be had by doing this, they might as well just do the correct thing, build lenses with coverage large enough not to vignette a given sensor. Canon has not done this, they are stuck with lenses made for film. I don't understand what the big deal is since they had lenses that cover huge areas of film since the 1800s. Some Petzval lenses covered plates a couple feet in width. > Unless there is some savings to be had by doing this, they might > as well just do the correct thing, build lenses with coverage large [quoted text clipped - 3 lines] > they had lenses that cover huge areas of film since the 1800s. Some > Petzval lenses covered plates a couple feet in width. And they ALL had light fall off, as do ALL camera lenses made for ALL consumer cameras from ALL manufacturers. Roger >> Unless there is some savings to be had by doing this, they might >> as well just do the correct thing, build lenses with coverage large [quoted text clipped - 6 lines] >And they ALL had light fall off, as do ALL camera lenses >made for ALL consumer cameras from ALL manufacturers. Not just 'consumer cameras". http://www.luminous-landscape.com/reviews/cameras/expensive-ps.shtml All wide angle lenses vignette (dark corners) and the Digitar 35mm XL is no exception. There are two solutions. One is to fix it via software, such as Vignette from The Imaging Factory. This is the least expensive approach, but if possible it is preferable to use a center filter. These are expensive, but if you can afford it, and the two stops of light lost, it's likely the preferable approach. ***************************************************** "Vietnam is what we had instead of happy childhoods." Tim Page in "Dispatches" by Michael Herr > Not just 'consumer cameras". > [quoted text clipped - 6 lines] > filter. These are expensive, but if you can afford it, and the two > stops of light lost, it's likely the preferable approach. By consumer, I meant everything except scientific, so I meant to include pro lenses, like Canon's L series. Roger >> Unless there is some savings to be had by doing this, they might >> as well just do the correct thing, build lenses with coverage large [quoted text clipped - 8 lines] > >Roger Sure, but how much light fall off? If the image circle is large enough (for eg) to encompass a 45mm area then its likely that coverage over 35mm would not be vignetted. -Rich >>>Unless there is some savings to be had by doing this, they might >>>as well just do the correct thing, build lenses with coverage large [quoted text clipped - 13 lines] > coverage over 35mm would not be vignetted. > -Rich Image circle has little to do with the amount of light fall off; i.e. a large image circle does not mean no light fall off. Roger >>>>Unless there is some savings to be had by doing this, they might >>>>as well just do the correct thing, build lenses with coverage large [quoted text clipped - 17 lines] > >Roger This would seem to be true. I just carried out a fairly informal test photographing a plain pale coloured wall on my newly-acquired 5D using a 17-35mm f/2.8L and a 24mm f/3.5 TS-E. Both showed similar levels of vignetting wide open (which disappeared 2 stops down). Since the 24 TS-E has a (for 35mm) truly enormous image circle, to cope with the +/- 12mm or so of shift, I found this quite surprising, but the results speak for themselves. All were quite easily correctable using the Canon software to the point of near-invisibility, which in a real-life picture would mean total invisibility. David SignatureDavid Littlewood SNIP > I just carried out a fairly informal test photographing a plain pale > coloured wall on my newly-acquired 5D using a 17-35mm f/2.8L and a > 24mm f/3.5 TS-E. Both showed similar levels of vignetting wide open > (which disappeared 2 stops down). I also conducted some tests, in this case between an EF 50mm f/1.4 set to f/2.8 and a TS-E 45mm f/2.8 wide open. The latter exhibited more light fall-off. The design compromise apparently didn't allow for low fall-off, wide open. I have yet to determine how far to stop that TS-E down for 'best' fall-off achievable. > Since the 24 TS-E has a (for 35mm) truly enormous image circle, to > cope with the +/- 12mm or so of shift, I found this quite > surprising, but the results speak for themselves. All three (24, 45 and 90mm) TS-E lenses have an image circle of 58.6mm at infinity focus. That should be more than adequate for the 43.3 full-frame diagonal, even if shifted. > All were quite easily correctable using the Canon software to the > point of near-invisibility, which in a real-life picture would mean > total invisibility. How do you do that with with the Canon software (DPP?)? Maybe it's something to do with the 5D, but I don't know of an anti-vignetting tool. Bart >SNIP >> I just carried out a fairly informal test photographing a plain pale [quoted text clipped - 7 lines] >fall-off, wide open. I have yet to determine how far to stop that TS-E >down for 'best' fall-off achievable. I mostly don't find the vignetting a problem on either the 24 or 45 TS-E lenses, since my normal procedure with interiors is to use f/5.6 or f/8 and a solid support (often the floor in churches where tripods are prohibited or discouraged). The vignetting has gone by then. I think of the f/3.5 as a nice bright image for composing and focussing, just as with a view camera. Haven't actually repeated the tests with the 45 or other lenses. >> Since the 24 TS-E has a (for 35mm) truly enormous image circle, to >>cope with the +/- 12mm or so of shift, I found this quite surprising, [quoted text clipped - 13 lines] > >Bart Sorry, slip of the brain, it was the PS CS2 Bridge RAW converter I was using. Not had time to investigate the Canon software yet. Needed +60 correction and the mid-point pushing down to zero (the default overcorrects the corners but leaves a dark mid-region). Not a facility I have used before I did test, as I don't have a problem with the images in practice. I find that most churches have naturally dark corner regions so vignetting is not apparent. Outside, I feel a little shading in the corners helps hold the picture together. Nice to have an easy way of reducing or removing it though. David SignatureDavid Littlewood SNIP > I mostly don't find the vignetting a problem on either the 24 or 45 > TS-E lenses, since my normal procedure with interiors is to use > f/5.6 or f/8 and a solid support (often the floor in churches where > tripods are prohibited or discouraged). The vignetting has gone by > then. I can now confirm that for the 45mm, as I just completed the (un-shifted) fall-off analysis of my TS-E 45 mm (f/4.5 and up won't get better): <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff45mm.png> and my EF 50mm f/1.4 shows (marginally) less fall-off (also f/4.5 and up, although f/4.0 is almost as good): <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff.png> Interestingly, my 45mm lens also peaks early in resolution (versus diffraction) at about f/4.0 , which means that the 45mm T&S is really useful to get top performance (apparent DOF and true resolution), besides the creative potential, and at relatively wide apertures, although that may differ some between sensors (I tested with a 1Ds Mk II which has a different sensel pitch). SNIP > Sorry, slip of the brain, it was the PS CS2 Bridge RAW converter I > was using. Not had time to investigate the Canon software yet. That's okay, I was just wondering if I missed something in the latest DPP (version 2.2.1). Bart > "David Littlewood" <david@nospam.demon.co.uk> wrote > [quoted text clipped - 7 lines] > better): > <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff45mm.png> Oh that's a tilt/shift lens so I wouldn't expect much fall-off. > and my EF 50mm f/1.4 shows (marginally) less fall-off (also f/4.5 and > up, although f/4.0 is almost as good): > <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff.png> That speaks well for the 50/1.4 if it matches the wider circle of the tilt/shift lens. So that same half stop difference in the worst corner is due to the sensor angle or the allowed engineered fall-of of the lens? Either way 1/2 stop isn't much. > Interestingly, my 45mm lens also peaks early in resolution (versus > diffraction) at about f/4.0 , which means that the 45mm T&S is really > useful to get top performance (apparent DOF and true resolution), > besides the creative potential, and at relatively wide apertures, > although that may differ some between sensors (I tested with a 1Ds Mk II > which has a different sensel pitch). SNIP >> <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff45mm.png> > > Oh that's a tilt/shift lens so I wouldn't expect much fall-off. That's why I mentioned it, as one would hope for less fall-off. I'm working on an additional full shift analysis, which will probably show a bit more fall-off. Not to obsess over it, it can be handled in post-processing if it isn't to much. >> and my EF 50mm f/1.4 shows (marginally) less fall-off (also f/4.5 >> and up, although f/4.0 is almost as good): >> <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff.png> > > That speaks well for the 50/1.4 if it matches the wider circle of > the tilt/shift lens. No, it won't match the TS-E image circle, but it also doesn't have to. > So that same half stop difference in the worst corner is due to the > sensor angle or the allowed engineered fall-of of the lens? It's even a bit less than half a stop, and again less for the average of all 4 corners. It is also measured at the most extreme pixels of the corner and is already better if you move in a millimetre or so (quickly lost when mounting a print). It is the combined result of an approximately circular aperture turning slightly elliptical when viewed from an angle, plus the higher magnification (due to approx. 9% longer projection distance to the corner at infinity focus) which could produce 0.25 stops loss, plus whatever remaining effect the AA-filter and microlenses and sensor have. > Either way 1/2 stop isn't much. I agree, and most of it is caused by lens design and physics. The type of sensor in this case has nothing to do with most of the light fall-off, and I expect something similar from the 5D. That's one of the reasons why I did the "full system" tests, it's what the actual results are from shooting with such a system. The aperture variation fall-off series will also allow to produce very good anti-vignetting settings at any aperture in postprocessing (even if it is hard to estimate in some images due to lack of smooth gradients). Because the series was shot in Raw, I can re-use them with any Raw converter if I copy the tonemapping parameters that are to be used for an actual image. Bart >It is the combined result of an approximately circular aperture turning >slightly elliptical when viewed from an angle, plus the higher [quoted text clipped - 10 lines] >That's one of the reasons why I did the "full system" tests, it's what >the actual results are from shooting with such a system. So by combination of both your results and mine it is clear that the light fall-off is no worse with digital than it can be on film. Your measurements show that the total effect is close to what would be expected for the lens alone, whilst my measurements show that the contribution of the sensor is negligible within the constraints of the lens mount. SignatureKennedy 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) SNIP > So by combination of both your results and mine it is clear that the > light fall-off is no worse with digital than it can be on film. Well, there's still the possibility that film is worse ;-) Afterall, it doesn't have microlenses ... On a more serious note, there may still be differences with other sensor arrays. As I interpret Canon's information at: <http://web.canon.jp/Imaging/cmos/technology-e/light_gathering.html> , part of the solution is in reducing the circuit area of the sensels. It'll allow a larger unobstructed area for the microlenses to focus their light on (thus also reducing hor/ver roll-off angle differences). FF-sensor arrays that pre-date the 5D and 1DS Mark II might respond differently. > Your measurements show that the total effect is close to what would > be expected for the lens alone, whilst my measurements show that the > contribution of the sensor is negligible within the constraints of > the lens mount. Yes, IMHO you covered your (sensor only) end well enough, so I thought it might be useful to approach the same question from another angle (pun intended). Both approaches strongly indicate (although the sample size of 2 different sensors is smallish they *were* mass-produced) a negligible influence from the sensor array under the limitations to the angle of incidence posed by the Canon mount. As a personal interest I will also test a fully shifted TS-E 45mm f/2.8, although those results will have little bearing on normal/un-shifted photography. I expect having to use a smaller aperture before an unobstructed cone of light can be produced. Current (e.g. Photoshop) de-vignetting tools don't allow to offset the center of vignetting, so I'll have to verify if the addition of 'some' additional 'canvas' on one side will allow the tools to perform their intended 'magic' on 'shifted' images. After that the 'tilt' needs to be factored in ... My quest hasn't ended yet, but the topic at hand seems to have been adequately debunked. Just wondering how often will this thread will need to be linked to in the future? Bart > My quest hasn't ended yet, but the topic at hand seems to have been > adequately debunked. Just wondering how often will this thread will need > to be linked to in the future? Heh, but there was some pretty convincing arguments on the other side too! eawckyegcy@yahoo.com wrote: > I repeated McEwen's experiment last night with a Canon 1DMkII and a > PrincetonTec "Impact" LED flashlight (it has a very clean pattern). [quoted text clipped - 12 lines] > +20 850.8 897.3 320.3 848.8 > +25 539.8 570.1 189.8 543.0 That's 50% fall-off if I'm reading correctly... well maybe that's the same as a stop... <scratching head> SNIP > That's 50% fall-off if I'm reading correctly... well maybe that's > the same as a stop... <scratching head> Yes -1.09 stops, which is much much more fall-off than Kennedy could measure from his sensor without lens. Maybe if "eawckyegcy" could take a Raw of a (50mm?) lens mounted, infinity focus shot of a uniformly lit surface, we could see if that linear raw conversion shows a similar deviation compared to my shots. I assume that a shot an f/5.6 would eliminate potential specimen variations. If his shot doesn't deviate much from mine, then it must be some methodology issue with his earlier test. Bart >As a personal interest I will also test a fully shifted TS-E 45mm >f/2.8, although those results will have little bearing on [quoted text clipped - 6 lines] >intended 'magic' on 'shifted' images. After that the 'tilt' needs to be >factored in ... I wonder if the answer would be to take a photo of an even grey or white wall at various settings, reverse it and use it as an adjustment layer. I'm sure someone more expert in PS than I am could comment further. A set of a dozen or two correction layers would be quite useful. David SignatureDavid Littlewood >Well, there's still the possibility that film is worse ;-) >Afterall, it doesn't have microlenses ... [quoted text clipped - 7 lines] >FF-sensor arrays that pre-date the 5D and 1DS Mark II might respond >differently. Possibly. I was looking at the Philips/Dalsa full frame sensors while back, as a possible home made digital back for a 35mm SLR (which didn't get off the ground). These are CCDs rather than CMOS chips and they have no microlenses at all, just Bayer filters on the colour version. The datasheet for the 11Mpix colour chip at http://www.dalsa.com/pi/documents/FTF4027C_datasheet_20040419.pdf shows a chart in Figure 8 for angular response. At 30deg horizontal (the limit of the Canon mount) it is 85% of the on axis response. This is a little lower than I measured with the microlensed Canon sensor, but still only 1.6% lower than the cos(30) expected, making it negligible. I am sure it is the microlenses on the Canon chip which produce the better angular responseof their full frame sensors. SignatureKennedy 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) SNIP > I am sure it is the microlenses on the Canon chip which produce the > better angular responseof their full frame sensors. For sure, the microlenses are the most important roll-off prevention factor (eventhough primarily designed for a better apparent fill-factor). IMHO next in importance is (especially on CMOS) the miniaturization of per sensel circuits to avoid directional (left/right/hor/ver) variation. Bart >>Well, there's still the possibility that film is worse ;-) >>Afterall, it doesn't have microlenses ... [quoted text clipped - 21 lines] >is a little lower than I measured with the microlensed Canon sensor, but >still only 1.6% lower than the cos(30) expected, making it negligible. So are you saying that a 15% difference in response would not be visible? I thought that even a 10% difference would be immediately apparent. -Rich >>The datasheet for the 11Mpix colour chip at >>http://www.dalsa.com/pi/documents/FTF4027C_datasheet_20040419.pdf [quoted text clipped - 6 lines] >visible? I thought that even a 10% difference would be immediately >apparent. Well actually, what I am saying is the 1.6% would not be visible, since the 13.3% would be present on any detector, whether film or digital. Nevertheless, 15% certainly wouldn't be visible as a response fall off, since this is a linear response variation and what you perceive is logarithmic. So even 15% would be negligible in terms of perception. SignatureKennedy 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) >>> <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff45mm.png> >> [quoted text clipped - 16 lines] >> So that same half stop difference in the worst corner is due to the >> sensor angle or the allowed engineered fall-of of the lens? On second thought, the lens design of a tilt/shift (unshifted) should not drop off the same as a fixed angle lens so this probably is sensor fall-off. I think 1/2 stop is not a big deal though maybe slightly noticeable? I'm not experienced enough to say. > It's even a bit less than half a stop, and again less for the average of > all 4 corners. It is also measured at the most extreme pixels of the [quoted text clipped - 24 lines] > > Bart >SNIP >> I mostly don't find the vignetting a problem on either the 24 or 45 [quoted text clipped - 9 lines] >up, although f/4.0 is almost as good): ><http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/LightFalloff.png> Very interesting - but I had trouble reconciling your graphs with the above words. On the charts, the f/1.4 seems to fall off slightly more at f/2.8 than the TS lens, but you say above it shows marginally less fall off. Not that the difference is material though. >Interestingly, my 45mm lens also peaks early in resolution (versus >diffraction) at about f/4.0 , which means that the 45mm T&S is really >useful to get top performance (apparent DOF and true resolution), >besides the creative potential, and at relatively wide apertures, >although that may differ some between sensors (I tested with a 1Ds Mk >II which has a different sensel pitch). Normally the sign of a good lens. I have to confess that since I got my 45mm I have shamefully under-utilised it. Not often I can justify taking both it and the 24mm in a bag to be carried around all day! David SignatureDavid Littlewood SNIP > ... I had trouble reconciling your graphs with the above words. On > the charts, the f/1.4 seems to fall off slightly more at f/2.8 than > the TS lens, but you say above it shows marginally less > fall off. I've structured the graphs in such a manner that higher means better==less fall-off==higher Luminocity relative to the center. At f/2.8 the 1.4 is trailing a bit, but from f/4.0 on the 1.4 takes the lead > Not that the difference is material though. Indeed. They both soon reach the point where stopping down further won't help to reduce the fall-off. Once the physical barrel and/or front/rear lens diameter obstructions are eliminated, there's no more gain, just a narrower cone of light. Bart > SNIP >> I just carried out a fairly informal test photographing a plain pale [quoted text clipped - 7 lines] > wide open. I have yet to determine how far to stop that TS-E down for > 'best' fall-off achievable. This is cos**4 effect. In a wide angle lens the difference in path between straight on and the edge of the image is significant, so that distance is more to the corner of the image and the light (proportional to 1/r**2 is less). You will easily be able to see this through an SLR lens in the 16-24mm range. It will also show in digital images (but can be electronically corrected). With film the correction comes about to considerable effect when the negatives are printed, because the corners in the negative are brighter and so the cos**4 is somewhat cancelled out. >> I understand what vignetting is but I've never seen it so bad. >> I'll check out fixing it in Photoshop though. > >The only thing you can do to get rid of it is to crop the picture. What turnip truck did you fall off of? It's a standard correction in CS2. http://www.dpreview.com/reviews/adobephotoshopcs2/page2.asp You can do it in other versions with the gradient tool. Cropping is not required. ********************************************************** "A combat photographer should be able to make you see the color of blood in black and white" David Douglas Duncan Speaking on why in Vietnam he worked only in black and white http://www.hrc.utexas.edu/exhibitions/online/ddd/ > The only thing you can do to get rid of it is to crop the picture. Oh dear. Wrong answer. >>> The only thing you can do to get rid of it is to crop the picture. >> >> Oh dear. >> >> Wrong answer. In her typically female mindset, Helen has in two words, exemplifired he ability to make seemingly "expert" comments in a pure satirist method known as a "one liner" without actually saying anything at all. Helen, your most intelegent statements are when your mouth is shut. The 5D sensor is a trifle wider than most 35mm frames and is prone to producing more viginetting than say an EOS 5. Photoshop will indeed remove it during development of a RAW file into an image file. That is providing you have the latest CS2 version. Otherwise, you'll have to obtain an aftermarket action or plugin if you want to do it with an earlier version or JPEG file. Alternatively, there is a lens correction program called DxO. Provided you use a lens on their list (why are so many lenses producing faulty images) you can use this program to correct all known faults cause by lenses, from your images. Surprisingly good results too. Even images from the lowly "plastic" element Canon lenses can look surprisingly good. Signaturewww.photosbydouglas.com www.weprint2canvas.com If you really must write,use my name at an above domain. >>>> The only thing you can do to get rid of it is to crop the picture. >>> [quoted text clipped - 13 lines] >aftermarket action or plugin if you want to do it with an earlier version or >JPEG file. You can do it older versions using gradients... here's an example. http://robertdfeinman.com/tips/tip22.html ********************************************************** "Take the glamour out of war! I mean how the bloody Hell can you do that? Go and take the glamour out of a Huey, go take the glamour out of a Sheridan....Can you take the glamour out of a Cobra or getting stoned on China Beach?... Oh war is good for you, you can't take the glamour out of of that. It's like trying to take the glamour out of sex, trying to take the glamour out of the Rolling Stones...I mean, you know that, it just can't be done." Tim Page to editor asking him to "finally take the glamour out of war.". From "Nam" by Tim Page >>> Oh dear. >>> [quoted text clipped - 5 lines] > as a "one liner" without actually saying anything at all. Helen, your most > intelegent statements are when your mouth is shut. Were you born as a sexist twerp or is it something you have studied since that regrettable and no doubt mistaken event? >>>>> Oh dear. >>>>> [quoted text clipped - 8 lines] >> Were you born as a sexist twerp or is it something you have studied >> since that regrettable and no doubt mistaken event? Actually all the females in my life are both respected and loved. Fortunately they don't attempt to bellittle people to come across as some sort of smart bitch. If you actually had something of value - no, just something relevant to a thread would do, you'd go in your interest with my blessing but for some reason known only to you, you post remarks intended to not as a reply but a barb to offend people genuine in their intent. Why do you do that? Signaturewww.photosbydouglas.com www.weprint2canvas.com If you really must write,use my name at an above domain. > The 5D sensor is a trifle wider than most 35mm frames 35mm film is 35mm wide from side to side, but the image area is actually 36mm x 24mm (the sprockets make up the difference between 24mm and 35mm). The 5D's sensor is 35.8mm x 23.9mm, according to http://www.dpreview.com/reviews/canoneos5d/page2.asp i.e. a trifle _narrower_ than the standard 35mm film image. > Alternatively, there is a lens correction program called DxO. Provided you > use a lens on their list (why are so many lenses producing faulty images) If you want lighter, smaller, more convenient (zoom!) and/or cheaper lenses, you have to make allowances somewhere, so you get slower, more flareprone, less well corrected lenses with lotsa plastic. If you want something better, you pay lots and lots more. -Wolfgang >>>Has anyone else noticed that on wide angle lenses on a Canon 5D the >>>corners [quoted text clipped - 27 lines] > I understand what vignetting is but I've never seen it so bad. > I'll check out fixing it in Photoshop though. FWIW, different lenses are supposed to have it to varying degrees depending on the exit pupil of the lens. You might want to look for a lens with a larger exit pupil (too bad mfgs don't spec this for photo lenses). Photoshop does have tools for dealing with this as does Nikon Capture (ironic isn't it since Nikon doesn't make a FF dslr to really have this problem in the first place? maybe it's a sign). > I understand what vignetting is but I've never seen it so bad. > I'll check out fixing it in Photoshop though. sensors are very sensitive to the angle of incoming light. Most (all?) wide lenses made for film will project light at the very sharp angle at the edges of imaging circle. This is called telecentricity (or rather lack of thereof). It does not matter for film, but matters for sensors a lot, mostly due to microlenses. The closer edge of sensor is to the edge of imaging circle, the more visible vignetting and blur is (for wide lenses). IMVHO this is also the biggest problem of 24x36mm sensors (price is another problem). You could have see this on some sample photos Canon published for 5D, actually. B. >> I understand what vignetting is but I've never seen it so bad. >> I'll check out fixing it in Photoshop though. [quoted text clipped - 8 lines] >(price is another problem). You could have see this on some sample >photos Canon published for 5D, actually. Price is not a problem. All you need to make more money. Problems are things which money can't fix. ********************************************************** "A combat photographer should be able to make you see the color of blood in black and white" David Douglas Duncan Speaking on why in Vietnam he worked only in black and white http://www.hrc.utexas.edu/exhibitions/online/ddd/ > ********************************************************** > > "A combat photographer Have you been flamed today for your "illegally"-long SIG? If not, I'll cheerfully do it. Read once, it's fine. I might have even followed the link it includes. Reading it FIVE times in 13 articles (including those of lazy quoters) in the SAME thread requires that I NEVER visit the site. Did you know that, if you insert two hyphens and a single space at the beginning of a blank line, standard-conforming newsreaders will not quote what follows? Do you care? I suspect not. > Speaking on why in Vietnam he worked only in black and white Just like Woody Allen, artistic merit had NOTHING to do with it. It's easier and CHEAPER. Ansel Adams would probably have used COLOR had it been feasible using the back of a horse-drawn wagon as a darkroom. (Rosebud THIS!!) SignatureJR -- :) JR >> ********************************************************** >> [quoted text clipped - 10 lines] >beginning of a blank line, standard-conforming newsreaders will not quote what >follows? Do you care? I suspect not. No, I don't care but maybe some might learn something. >> Speaking on why in Vietnam he worked only in black and white > >Just like Woody Allen, artistic merit had NOTHING to do with it. It's easier >and CHEAPER. Ansel Adams would probably have used COLOR had it been feasible >using the back of a horse-drawn wagon as a darkroom. (Rosebud THIS!!) I take it you know nothing of Duncan's work or career. Others were working in color in Vietnam at that time and Duncan was the best combat photographer in the world then and he made an explicit decision to work in B&W. Cost was not an issue as he had all the resources of Time-Life behind him. But if you had heard his speech at the giving of his archive to the Ransom you would know that. You could have even talked with him at the reception afterwards. Well I'm sure the Ransom will never want your photo archive. ************************************************************** "There has always been war. War is raging throughout the world at the present moment. And there is little reason to believe that war will cease to exist in the future. As man has become increasingly civilized, his means of destroying his fellow man have become ever more efficient, cruel and devastating. Is it possible to put an end to a form of human behavior which has existed throughout history by means of photography? The proportions of that notion seem ridiculously out of balance. Yet, that very idea has motivated me. James Nachtwey War Photographer http://www.jamesnachtwey.com/ > No, I don't care That's as I expected. I recall your cavalier attitude on prior occasions where your interminable sig was discussed. > but maybe some might learn something. Then my words were worthwhile as I had no expectation that YOU would learn anything. > I take it you know nothing of Duncan's work or career. Your presumptuous snobbery is well rehearsed. > Duncan was the best combat photographer in the world then Your fawning adoration is obvious. > and he made an explicit decision to work in B&W. Of course he did. He is, after all, an ARTEEST. > Cost was not an issue as he had all the resources of Time-Life behind him. Cost, my sheltered "friend", is ALWAYS an issue. > You could have even talked with him at the reception afterwards. Little, old ME? I wouldn't presume... > Well I'm sure the Ransom will never want your photo archive. Sir, you cut me to the quick! I'll have you know we are in negotiation as I type. You can soon expect a traveling exhibit featuring messy-faced babies, precocious grandchildren and still-life that has been shot by a million others. And get this: They're considering COLOR work! Hehehehehe! Signature:) JR >> I understand what vignetting is but I've never seen it so bad. >> I'll check out fixing it in Photoshop though. [quoted text clipped - 10 lines] > > B. Thanks for that explanation. The lets me understand why I've not noticed it on film SLRs over the past 30 years nor on my Canon 300D. >>> I understand what vignetting is but I've never seen it so bad. >>> I'll check out fixing it in Photoshop though. [quoted text clipped - 14 lines] > The lets me understand why I've not noticed it on film SLRs over the past > 30 years nor on my Canon 300D. Heresy! :-) You MUST have noticed this on film because all the canon guys claim this as fact! SignatureStacey >> I understand what vignetting is but I've never seen it so bad. >> I'll check out fixing it in Photoshop though. [quoted text clipped - 10 lines] > >B. It wouldn't be a problem if Canon didn't insist on using old lens designs they've amortized 1000x over. -Rich >> On Sun, 12 Mar 2006 15:15:29 +0000, Bronek Kozicki <brok@rubikon.pl> >> wrote: [quoted text clipped - 19 lines] >> designs they've amortized 1000x over. >> -Rich That's not very fair Rich... They must have done it at least 5000 times! Signaturewww.photosbydouglas.com www.weprint2canvas.com If you really must write,use my name at an above domain. >> I understand what vignetting is but I've never seen it so bad. >> I'll check out fixing it in Photoshop though. [quoted text clipped - 8 lines] >(price is another problem). You could have see this on some sample >photos Canon published for 5D, actually. Not much of a factor for DSLRs since the closest sensor-rear element distance and hence the angle of the rays is determined by the mount and mirror clearance considerations. KS >>Has anyone else noticed that on wide angle lenses on a Canon 5D the >>corners [quoted text clipped - 24 lines] > he worked only in black and white > http://www.hrc.utexas.edu/exhibitions/online/ddd/ I see how to adjust this in Photoshop. However, if I have to manually adjust all my images then I going to think about sending the camera back and getting a Nikon D-200. I hardly ever use Photoshop. Can you advise if I can use the Photoshop batch option to automatically adjust images for me? > I see how to adjust this in Photoshop. > However, if I have to manually adjust all my images then I going to think [quoted text clipped - 3 lines] > Can you advise if I can use the Photoshop batch option to automatically > adjust images for me? ImagesPlus has a calibration mode where you make a flat field, then assign the flat in a calibration window, then do batch corrections on all your images. And it does 64-bit math, and works very well. Roger > Has anyone else noticed that on wide angle lenses on a Canon 5D the > corners are dark from max aperture up to about f5.6? Once at f8, the [quoted text clipped - 4 lines] > > I noticed this today after taking some pictures in the snow. I've now noticed that there is a problem with the 50mm - it's just not so bad. >> Has anyone else noticed that on wide angle lenses on a Canon 5D the >> corners are dark from max aperture up to about f5.6? Once at f8, the [quoted text clipped - 7 lines] > I've now noticed that there is a problem with the 50mm - it's just not so > bad. It was also a problem with the same lenses on film cameras, not too much of a surprise, since the "sensor" was the same size as that on the 5D (and 1Ds/1Ds mkII, for that matter.) It really only rears its ugly head under certain circumstances, but shooting in snow would be one of them. SignatureSkip Middleton http://www.shadowcatcherimagery.com >>> Has anyone else noticed that on wide angle lenses on a Canon 5D the >>> corners are dark from max aperture up to about f5.6? Once at f8, the [quoted text clipped - 9 lines] >> > It was also a problem with the same lenses on film cameras, Uh notice he said he's never seen this on film. Have you actually bothered to test this for yourself or just afraid of what the results might be.. > -- Stacey >>>> Has anyone else noticed that on wide angle lenses on a Canon 5D the >>>> corners are dark from max aperture up to about f5.6? Once at f8, the [quoted text clipped - 18 lines] > > Stacey Stacey, be a dear and re-read his first post. (The one to which he added the comment to which I replied.) He said nothing about not seeing it on film. He said, in his second post, that he'd never seen it so bad. He did not specify the medium upon which he'd never seen it so bad. The post that he'd not seen it on film came in AFTER the post you are replying to, so I never had a chance to read it, much less comment on it. He also doesn't say what lenses, or system he used. I never noticed the vignetting on my old FD mount Canon stuff, but did with the 28-135 IS and (rented) 17-35 f2.8L and 28-70 f2.8L, among others. And, Stacey, I don't need to test it, I've seen it in the past when I shot a lot of film with some of the same lenses I use now. Including the above mentioned 28-135 IS. By the way, that $20 on the upgrade bet? You said a year or less, on Oct 2. So, if you insist on taking it to the bloody end, you have a little over 6 months... SignatureSkip Middleton http://www.shadowcatcherimagery.com >>>>> Has anyone else noticed that on wide angle lenses on a Canon 5D the >>>>> corners are dark from max aperture up to about f5.6? Once at f8, the [quoted text clipped - 22 lines] > the comment to which I replied.) He said nothing about not seeing it on > film. He said, in his second post, that he'd never seen it so bad. Exactly. > He > did [quoted text clipped - 4 lines] > say > what lenses, or system he used. "Thanks for that explanation. The lets me understand why I've not noticed it on film SLRs over the past 30 years nor on my Canon 300D." > By the way, that $20 on the upgrade bet? You said a year or less, on Oct > 2. So, if you insist on taking it to the bloody end, you have a little > over 6 months... Why wouldn't I? You think half way through has proven anything? Just wait till they come out with their next generation sensor, you won't be able to resist.. SignatureStacey |
Enable EMail Alerts
Start New Thread
Reply to this Message