 |
 | 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 / September 2008Canon 50D 5D
With the imminent appearance of the Canon 50D, the price of the 5D has fallen (there will be a new one in this line). Thus if I can get either the 50D or the 5D for a similar price, which would you suggest? One is old technology but full sensor and the other new technology dust prevention and APS-C sensor. With good but not very good lenses, I am told the full sensor is better. But one would not get the digic 4 and various functions. So what? Opinions with some explanations, would be most appreciated. Regards, Gedeon gedeon_h@tpg.com.au gedeonh@tpg.com.au > So what? Opinions with some > explanations, would be most appreciated. [quoted text clipped - 3 lines] > gedeon_h@tpg.com.au > gedeonh@tpg.com.au Write down your preferences based on what you´d like to shoot and buy the camera that suits your needs best and cheapest. Invest your savings in good glass. I had a 20D and got myself a 40D but returned that in a hurry and bought the 5D b/c i could not accept to have worse per pixel sharpness at 100% view than my old 20D. I´m happy now. Regards, Markus > With the imminent appearance of the Canon 50D, the price of the 5D has > fallen (there will be a new one in this line). [quoted text clipped - 5 lines] > But one would not get the digic 4 and various functions. So what? > Opinions with some explanations, would be most appreciated. grab the new 5D instead: http://www.canon.com/moon/en/index.html > grab the new 5D instead: > > http://www.canon.com/moon/en/index.html It's a pro model, so certainly looks that way. Also the buttons and the slope look like a 5D, as opposed to a 1D MK-IIIn. > > grab the new 5D instead: > > > > http://www.canon.com/moon/en/index.html > > It's a pro model, so certainly looks that way. Also the buttons and the > slope look like a 5D, as opposed to a 1D MK-IIIn. http://www.dpreview.com/news/0809/08090501somethingcanon.asp >> > grab the new 5D instead: >> > >> > http://www.canon.com/moon/en/index.html >> It's a pro model, so certainly looks that way. Also the buttons and the >> slope look like a 5D, as opposed to a 1D MK-IIIn. > > http://www.dpreview.com/news/0809/08090501somethingcanon.asp I am guessing that it will be revealed on 15 September then, when it's a full moon. Will also be around a week before Photokina. >> With the imminent appearance of the Canon 50D, the price of the 5D has >> fallen (there will be a new one in this line). [quoted text clipped - 9 lines] > > http://www.canon.com/moon/en/index.html I can not afford that sort of budget. What about the present 5D (digic II) compared top the new 50D. Same price for 50D plus lens as 5D body only and I have lenses for it. Regards, Gedeon Despite objective measures to the contrary some continue to insist that no APS sized sensor can have noise levels as low as or lower than "full frame" sensors. Some people believe John McCain is the candidate of change, despite the objective evidence of his voting record in the senate. The facts should never get in the way of a good opinion. if your brain says that a camera cannot be any good if it does not have the form factor of a 1936 Exacta with the same dimensions for the image capture media then in your hear you know which is the right camera for you. > Despite objective measures to the contrary some continue to insist that > no APS sized sensor can have noise levels as low as or lower than "full > frame" sensors. For low pixel counts, that may possibly be true. Now, show your objective data that for a given number of pixels, a high pixel count APS sized sensor has as little noise as a full frame. Let's say for 12 Mpix or more. >With the imminent appearance of the Canon 50D, the price of the 5D has >fallen (there will be a new one in this line). [quoted text clipped - 5 lines] >But one would not get the digic 4 and various functions. So what? >Opinions with some explanations, would be most appreciated. If you don't _know_ why to get a 5D then get the 50D, or even a 40D. Or, for that matter, an XTi.  SignatureRay Fischer rfischer@sonic.net > With the imminent appearance of the Canon 50D, the price of the 5D has > fallen (there will be a new one in this line). > Thus if I can get either the 50D or the 5D for a similar price, which > would you suggest? If price is a major point, get the 50D as you are much more likely to find affordable Canon lenses that'll support it compared to what you'll need for the 5D. > If price is a major point, get the 50D as you are much more likely to > find affordable Canon lenses that'll support it compared to what > you'll need for the 5D. Explain please? John. The full frame 5D requires either the very best film lenses you can find to maximize performance or new lenses specifically designed for a FF digital sensor. Nikon has some, Canon has few if any. >> If price is a major point, get the 50D as you are much more likely to >> find affordable Canon lenses that'll support it compared to what [quoted text clipped - 3 lines] > > John. > The full frame 5D requires either the very best film lenses you can find > to maximize performance or new lenses specifically designed for a FF > digital sensor. Nikon has some, Canon has few if any. This is a common myth. Any halfway decent lens makes lovely sharp images on the 5D. The only exception are the Canon superwides, but even the superwides make sharp images out to the corners at f/16. I wonder about the 50D, though. The 5D's Nyquist frequency is 60 lp/mm, so decent contrast in the 40 to 50 lp/mm range is required for sharp images. But the 50D's Nyquist frequency is 105 lp/mm. How many lenses have decent contrast in the 70 to 90 lp/mm range? SignatureDavid J. Littleboy Tokyo, Japan : > The full frame 5D requires either the very best film lenses you can find : > to maximize performance or new lenses specifically designed for a FF [quoted text clipped - 8 lines] : But the 50D's Nyquist frequency is 105 lp/mm. How many lenses have decent : contrast in the 70 to 90 lp/mm range? I'm having trouble grasping your point, presumably because I understand the underlying physics so poorly. Are you saying that with a lens of average quality, aliasing will be a more serious problem on the 50D than on the 5D because of the finer granularity of the 50D's sensor? And that as a result the 50D requires better lenses than the 5D does? Or am I missing it altogether? Bob > : > The full frame 5D requires either the very best film lenses you can > find [quoted text clipped - 16 lines] > > I'm having trouble grasping your point, I'm saying that it appears that the 50D requires better lenses than the 5D. The higher pixel density sensor requires 1.75 times better lenses, because, being higher density, it is "looking much more closely" at the image projected by the lens. Most 35mm lenses are having trouble at 40 lp/mm, even in the center. How does anyone expect then to cough up 70 lp/mm? It's a mystery.... SignatureDavid J. Littleboy Tokyo, Japan > On Mon, 8 Sep 2008 09:45:38 +0900, "David J. Littleboy" <davidjl@gol.com> >: The 5D's Nyquist frequency is 60 lp/mm, so >: decent contrast in the 40 to 50 lp/mm range is required for sharp images. >: But the 50D's Nyquist frequency is 105 lp/mm. How many lenses have decent >: contrast in the 70 to 90 lp/mm range? > I'm having trouble grasping your point, presumably because I understand the > underlying physics so poorly. Are you saying that with a lens of average > quality, aliasing will be a more serious problem on the 50D than on the 5D > because of the finer granularity of the 50D's sensor? No, to the contrary. He said that you need *very* high resolution lenses (70-90 cycles/mm)for 100% pixel view on the 50D opposed to merely high resolution ones (40-50 cycles/mm)for the 5D. Contrast is what makes a zebra crossing bvlack&white versus gray on gray --- and the 50D wants lenses that show even very tiny patterns (very high resolution) at good contrast. Few lenses can deliver. In other words, the 50D outperforms most lenses, thus low contrast and/or low resolution of the lens will be very visible at 100%. (It won't be visible in normal print unless you print well beyond, say, 20x30cm, nor will it be visible in web presentations.) Aliasing occurs if the sampling rate (pixel density) is not high enough to really capture the detail offered (by the lens), and then a 'false' pattern of a lower frequency can be recorded. AA-filters guard against this, as they optically (try to) remove too high frequency patterns. To have such high frequency patterns, you need lenses capable of resolving them for the given sensor. -Wolfgang >> On Mon, 8 Sep 2008 09:45:38 +0900, "David J. Littleboy" <davidjl@gol.com> > [quoted text clipped - 26 lines] >too high frequency patterns. To have such high frequency patterns, >you need lenses capable of resolving them for the given sensor. So if the 50D is better than the lenses, would there be any point in upgrading from the 40D? Is the better sensor in the 50D throwing away quality? Archibald > So if the 50D is better than the lenses, If. Depends on the lenses. A consumer 70-300mm at 30mm is a different beast than an L prime lens. > would there be any point in upgrading from the 40D? - All the other features the 50D has which the 40D has not. - Using your glass to the max. (Many people have many times the body price in lenses.) > Is the better sensor in the 50D throwing away quality? Nope, unless you compare different enlargements. It'll get the maximum out of your lens. If your lens can't deliver, you have wasted space (for the extra MPix that don't carry information) but not quality. Though the 50D *might* have increased noise (or noise supression) at high ISO. If so and how much remains to be seen; ist also depends on if you actually use high ISO settings. That's the only real thing In can think of that may reduce image quality. -Wolfgang : : > The full frame 5D requires either the very best film lenses you can find : : > to maximize performance or new lenses specifically designed for a FF [quoted text clipped - 14 lines] : because of the finer granularity of the 50D's sensor? And that as a result the : 50D requires better lenses than the 5D does? Or am I missing it altogether? Whereupon David responded: > I'm saying that it appears that the 50D requires better lenses than the 5D. > The higher pixel density sensor requires 1.75 times better lenses, because, > being higher density, it is "looking much more closely" at the image > projected by the lens. Most 35mm lenses are having trouble at 40 lp/mm, even > in the center. How does anyone expect then to cough up 70 lp/mm? It's a > mystery.... I'm still having trouble getting my head around this. As I understand (if I dare use that word) it, aliasing results when the resolving power of the lens and that of the sensor are about equal but the lines presented by the lens are out of phase with those detected by the sensor. And if the resolving power of the lens is poor, the image recorded by the sensor will be blurry, but not necessarily aliased. As the resolving power of the lens increases, so does the sharpness of the recorded image until the phase angle between the lens and the sensor represents twice the Nyquist frequency, at which point (absent specific anti-aliasing measures) aliasing effectively cancels out further increases in image quality. By this reasoning, the Nyquist frequency represents an upper bound on the advantage conveyed by a sharper lens, but doesn't imply that a poorer lens will perform worse than it would if the Nyquist frequency were lower. So rather than saying that the 50D "requires" better lenses, wouldn't it be more accurate to say merely that the 50D *benefits* more from a better lens than the 5D does? My knowledge of physics, such as it is, is more of an interest than a skill. So I'd be happy to have someone set me straight. Bob > I'm still having trouble getting my head around this. Forget pixels. Assume both the 5D and 50D have an infinite number of pixels and perfectly capture the image that the lens projects. Say you want to make a 12x18 print. Assume you want a quality print. What does "quality" mean. Well, the Leica types claim "8 lp/mm", but they're full of it. Lets say 4 lp/mm. Since 12 x 18 is 305 x 457 mm, you need 4 x 305 lines to be resolved in the short direction. That's 1220 line pairs. This is a requirement on the image file. To get an image file with that much information, an image with that much information must be projected onto the sensor. Since the 5D sensor is 24mm in the short direction, our lens needs to resolve 51* line pairs per milimeter to meet our quality requirements. Again, the _lens_ must project an image with 51 lp/mm onto the sensor. *: That's 51 = (1220/24). What about the 50D? The sensor is 14.9 x 22.3 mm. So to get those 1220 line pairs you need (1220/14.9), or 82 lp/mm. So the lens must project an image with 82 lp/mm onto the sensor. You will note that 82 lp/mm is a much bigger number than 51 lp/mm<g>. So to make a print with the same 4 lp/mm image quality, the 5D needs 51 lp/mm and the 50D needs 21 lp/mm from the lens. But 82 lp/mm is a lot of lp/mm for Canon/Nikon lenses. Even the better ones. > As I understand (if I dare use that word) it, aliasing results when the > resolving power of the lens > and that of the sensor are about equal but the lines presented by the lens > are > out of phase with those detected by the sensor. Aliasing is irrelevant to this discussion. But for the record, aliasing occurs when the lens projects an image onto the sensor that has better resolution than the sensor. That is, if the sensor sees (samples) an image that has information above the Nyquist frequency of the sensor, that information will be incorrectly reported as frequencies that are within the pass band of the sensor. > So rather than saying that the 50D "requires" better lenses, wouldn't > it be more accurate to say merely that the 50D *benefits* more from a > better > lens than the 5D does? That would be true if the 50D were the same size as the 5D, but it's 1.6 times smaller. So to provide the same 12x18 print quality, it requires better lenses. SignatureDavid J. Littleboy Tokyo, Japan > Say you want to make a 12x18 print. Assume you want a quality print. What > does "quality" mean. Well, the Leica types claim "8 lp/mm", but they're full > of it. Lets say 4 lp/mm. Since 12 x 18 is 305 x 457 mm, you need 4 x 305 > lines to be resolved in the short direction. That's 1220 line pairs. This is > a requirement on the image file. And at *least* 2440 pixel in that dimension, so you can /resolve/ 1220 line pairs. That means 9Mpix is enough. [...] > You will note that 82 lp/mm is a much bigger number than 51 lp/mm<g>. That's basically a fancy way of saying "when the sensor is smaller (as in APS, or worse, compact cameras), you've got to enlarge more for the same final result. Enlarging more also enlarges any lens problems and limitations." :-) However, I understood the problem as "you have a sensor with 15MPix(!), which lens will give enough information so you have an advantage over, 8, 10, 12 MPix?" > So to make a print with the same 4 lp/mm image quality, the 5D needs 51 > lp/mm and the 50D needs 21 lp/mm from the lens. But 82 lp/mm is a lot of > lp/mm for Canon/Nikon lenses. Even the better ones. s/21/82/ However, the 5D needs the 51 lp/mm even in the corner of the EF lens for full quality, whereas the 50D doesn't exploit the (usually weaker) outmost part of the image circle of an EF lens. Thus _just_ comparing the lp/mm can be misleading. -Wolfgang >> Say you want to make a 12x18 print. Assume you want a quality print. What >> does "quality" mean. Well, the Leica types claim "8 lp/mm", but they're [quoted text clipped - 6 lines] > And at *least* 2440 pixel in that dimension, so you can > /resolve/ 1220 line pairs. That means 9Mpix is enough. No, it means 9MP would be enough if you had a perfect AA filter and an RGB (or monochrome) sensor. In real life, getting 1220 lp/height from a 12MP Bayer sensor is about what one should be expecting. >> You will note that 82 lp/mm is a much bigger number than 51 lp/mm<g>. > > That's basically a fancy way of saying "when the sensor is smaller > (as in APS, or worse, compact cameras), you've got to enlarge > more for the same final result. Enlarging more also enlarges > any lens problems and limitations." :-) Yes, but, it's giving a specific technical requirement on lenses you'll need on your 15MP camera to match 12MP camera performance. > However, I understood the problem as "you have a sensor with > 15MPix(!), which lens will give enough information so you have > an advantage over, 8, 10, 12 MPix?" Right, which is why one needs to provide a lower limit. > However, the 5D needs the 51 lp/mm even in the corner of the > EF lens for full quality, whereas the 50D doesn't exploit the > (usually weaker) outmost part of the image circle of an EF lens. > Thus _just_ comparing the lp/mm can be misleading. But hoping that an FF 20mm lens will be 1.6 times better in the center than an FF 35mm lens will be overall is rather problematic as well. Lenses get worse in performance as they get wider. And there's also the problem that you can't use these lenses anywhere near wide open on a cropped camera: they are already funky soft on FF, and will be really gross on high-MP-count APS-C. In real life, anything 50mm or longer in the Canon line provides lovely sharp images at the corners with no trouble whatsoever, and the 24TSE does nicely in the corners, too. So there's a lot of photography that can be done with sharp corners. And if you can put up with f/16, even the superwides are sharp in the corners. (The 5D + 17-40 is sharper at every f stop from f/5.6 to f/11 at every point in the frame than the 20D + 10-22. At the center, the 50D might close up the difference, but it won't at the edges and corners, and that's without even looking at the 17-40 at f/16.) Again in real life, one sees a lot of sharp images from the 5D. Of course, if the 5DII is 24MP, I'm going to be buying the Zeiss ZF 18/3.5 and an adaptor... SignatureDavid J. Littleboy Tokyo, Japan >> And at *least* 2440 pixel in that dimension, so you can >> /resolve/ 1220 line pairs. That means 9Mpix is enough. > No, it means 9MP would be enough if you had a perfect AA filter and an RGB > (or monochrome) sensor. > In real life, getting 1220 lp/height from a 12MP Bayer sensor is about what > one should be expecting. I stand corrected. >> However, the 5D needs the 51 lp/mm even in the corner of the >> EF lens for full quality, whereas the 50D doesn't exploit the >> (usually weaker) outmost part of the image circle of an EF lens. >> Thus _just_ comparing the lp/mm can be misleading. > But hoping that an FF 20mm lens will be 1.6 times better in the center than > an FF 35mm lens will be overall is rather problematic as well. True. At least from the (sensible) 'same field of view' point of view. It can be different from a 'same lens' POV --- at least dpreview says that that the Nikon AF-S VR 70-200mm F2.8 G ED has quite a different performance on FX than on DX. That may be the odd one out, but I still think Nikon tweaked it for maximum DX performance. -Wolfgang What a wonderful sounding discussion for physic nurds and totally unhelpful for us plebeians. Now here is a question which if answered would help me: How would the pictures compare taken with - Canon 5D fitted with a Tamron 28-300 and Canon 50D fitted with a Canon 18-200 (and or Tamron 18-250) in daylight and in low light at 1600 ISO each? So we compare EF-s lenses and a non EF-s lens on both cameras and test higher ISO shooting even at 3200 if the 5D has that capability. If still interested add the "cheaper" L lenses such as f4 17-40 on the cameras. The answers witout the "why"s i.e. no technical terms would be revealing. Regards, Gedeon >>> And at *least* 2440 pixel in that dimension, so you can >>> /resolve/ 1220 line pairs. That means 9Mpix is enough. [quoted text clipped - 25 lines] > > -Wolfgang > What a wonderful sounding discussion for physic nurds and totally > unhelpful for us plebeians. [quoted text clipped - 13 lines] > > The answers witout the "why"s i.e. no technical terms would be revealing. Both Tamron lenses are very, very, very mediocre. I wouldn't insult the camera makers of any brand by usesing them. >> What a wonderful sounding discussion for physic nurds and totally >> unhelpful for us plebeians. [quoted text clipped - 13 lines] >> >> The answers witout the "why"s i.e. no technical terms would be revealing. > Both Tamron lenses are very, very, very mediocre. The later Tamron is so much superior to the earlier that you should at least drop one of those three "very"s when describing it. It's a new generation with a new standard of performance in that range. > I wouldn't insult the camera makers of any brand by usesing them. Fair enough. You clearly have very high standards of lens performance. But you do seem to be implying that you might nevertheless be prepared to use the Canon 18-200 zoom, despite it sharing the inevitable performance deficits of even the most recent and highest quality zooms of such a range. Can you clarify? SignatureChris Malcolm, IPAB, School of Informatics, Informatics Forum, 10 Crichton Street, Edinburgh EH8 9AB >>> What a wonderful sounding discussion for physic nurds and totally >>> unhelpful for us plebeians. [quoted text clipped - 20 lines] > least drop one of those three "very"s when describing it. It's a new > generation with a new standard of performance in that range. Okay, dropping one :-) >> I wouldn't insult the camera makers of any brand by usesing them. > [quoted text clipped - 3 lines] > sharing the inevitable performance deficits of even the most recent > and highest quality zooms of such a range. Can you clarify? I have a EOS 5D & 40D with EF 24-105 f/4 L IS USM EF 70-200m f/4 IS USM EF 100 f/2,8 USM macro. I can't imagine me buying the EF-S 18-200 unless Canon found a magic way. Even then ...... > Now here is a question which if answered would help me: > How would the pictures compare taken with - > Canon 5D fitted with a Tamron 28-300 and > Canon 50D fitted with a Canon 18-200 (and or Tamron 18-250) > in daylight and in > low light at 1600 ISO each? Let me put it this way: you got a 50.000 horse power engine --- and all you can think about is putting it in a econobox using wheels rated for at most 45 miles/hour? Use good lenses. Always use good lenses. The body can only record what the lens can deliver. With superzooms you don't need much of a sensor. If you *really* need a 28-300, use the EF28-300mm f/3.5-5.6L IS USM. It's not great in any way, but it certainly is better than the Tamron. -Wolfgang : What a wonderful sounding discussion for physic nurds and totally : unhelpful for us plebeians. [quoted text clipped - 13 lines] : : The answers witout the "why"s i.e. no technical terms would be revealing. You've already gotten two responses strongly disparaging the Tamrons, one explicitly and the other in a blanket denunciation of any non-Canon lens. But I can't help noticing that neither responder acknowledged ever having seen, much less used, either of those lenses. ;^) Bob > You've already gotten two responses strongly disparaging the Tamrons, one > explicitly and the other in a blanket denunciation of any non-Canon lens. > But > I can't help noticing that neither responder acknowledged ever having > seen, > much less used, either of those lenses. ;^) FWIW, I have the Tamron 28-75/2.8, and I had been quite happy with it. It has produced some of the sharpest images I've seen from both my 300D and 5D. But that's because I almost always shoot in the 28 to 35mm range of the lens where it really is gloriously sharp. I was fooling around with it (test shots to compare the 5D to 6x7 film) and found that at 50mm and over it's mush. Oops. (It may be out of adjustment; it's old and the extending section doesn't seem very strong when extended.) But I'm planning on moving to all primes, starting this trip. I'm only taking the 24TSE and 50/1.4. SignatureDavid J. Littleboy Tokyo, Japan : > You've already gotten two responses strongly disparaging the Tamrons, one : > explicitly and the other in a blanket denunciation of any non-Canon lens. [quoted text clipped - 13 lines] : But I'm planning on moving to all primes, starting this trip. I'm only : taking the 24TSE and 50/1.4. And which body? If the 5D, then you're planning to wing it with no telephoto capability at all?? Bob > : But I'm planning on moving to all primes, starting this trip. I'm only > : taking the 24TSE and 50/1.4. > > And which body? If the 5D, then you're planning to wing it with no > telephoto > capability at all?? Telephoto? Do people actually use that gimmick? (I'd take the 100/2.0 along too if I thought I'd use it. Maybe I will. But I like to travel light.) SignatureDavid J. Littleboy Tokyo, Japan >>> But I'm planning on moving to all primes, starting this trip. I'm >>> only taking the 24TSE and 50/1.4. [quoted text clipped - 7 lines] > (I'd take the 100/2.0 along too if I thought I'd use it. Maybe I > will. But I like to travel light.) I use the 70-300mm on my Nikon D40 quite a lot of the time, and many of the pictures are at 300mm (450mm eq.)! And yet - I find the 16mm end of the 16-85mm used a lot as well. <G> David > Telephoto? Do people actually use that gimmick? Yes. In a perfect world, you could step right up to the lion (or just a shy animal) and use your 28mm. In this world, lions sometimes get slightly upset and give you the proper papparazzi treatment. In a perfect world, there never are any inconvenient barriers, walls, fences, etc. In a perfect world, you'll also have as much --- or as little --- DOF as you wish. In this world, many people aren't relaxed when a camera points in their direction, and just look plainly bad in their stiff poses --- and sometimes stepping back and photographing them when they don't have their camera face on, is the answer (and may require a telephoto lens). Of course, 'get closer' is still a good idea, but not at the cost of disturbing these fine fellows: http://shooting4joy.com/photos/377895924_8fyKA-X2-1.jpg (768x768 pixel, Original, 2050x2050 pixel: http://shooting4joy.com/photos/377895924_8fyKA-O-1.jpg ) Feel free to guess the lens and aperture used. -Wolfgang >> Telephoto? Do people actually use that gimmick? > > Yes. > In a perfect world, you could step right up to the lion That's one seriously defective sense of humor you've got there. I'd recommend taking it in for a tune up, if not replacement. SignatureDavid J. Littleboy Tokyo, Japan >>> Telephoto? Do people actually use that gimmick? >> Yes. >> In a perfect world, you could step right up to the lion > That's one seriously defective sense of humor you've got there. Straight faced delivery and all that. -Wolfgang > > The full frame 5D requires either the very best film lenses you can find > > to maximize performance or new lenses specifically designed for a FF [quoted text clipped - 3 lines] > the 5D. The only exception are the Canon superwides, but even the superwides > make sharp images out to the corners at f/16. Would that it was true. A 17-40L is incapable at f5.6 (as an example) of rendering decent edge quality with that camera. Having to stop down to f16 to achieve this is pretty bad. > I wonder about the 50D, though. The 5D's Nyquist frequency is 60 lp/mm, so > decent contrast in the 40 to 50 lp/mm range is required for sharp images. [quoted text clipped - 3 lines] > -- > David J. Littleboy Take the best Canon wide lens and try it on the 40D and the 5D and look at the corners. The centre area (inner 2/3rds) of both images will likely be good even at relatively wide apertures, but the 5D's edges will not look as good as the 40Ds simply because the glass cannot support the image circle the way it should. It's not the height of resolution that should concern people, it's the aberration at the edges. > Tokyo, Japan On Sep 7, 8:45 pm, "David J. Littleboy" <davi...@gol.com> wrote: > This is a common myth. Any halfway decent lens makes lovely sharp images > on > the 5D. The only exception are the Canon superwides, but even the > superwides > make sharp images out to the corners at f/16. Would that it was true. A 17-40L is incapable at f5.6 (as an example) of rendering decent edge quality with that camera. Having to stop down to f16 to achieve this is pretty bad. <<<<<<<<<<<<< You might want to read what I wrote before you reply... It is irritating, but it works fine for landscape work where you need f/16 for DOF. > I wonder about the 50D, though. The 5D's Nyquist frequency is 60 lp/mm, so > decent contrast in the 40 to 50 lp/mm range is required for sharp images. > But the 50D's Nyquist frequency is 105 lp/mm. How many lenses have decent > contrast in the 70 to 90 lp/mm range? Take the best Canon wide lens and try it on the 40D and the 5D and look at the corners. <<<<<<<<<<<< I wasn't talking about the 40D. Also, you have to compare apples to apples. That is, you have to compare, say, the 20/2.8 on the 50D, with the 35/2.0 on the 5D. I suspect that you will find the 35/2.0 on the 5D better than the 20/2.8 on the 50D. (The "sweet spot" myth has always been rather bogus, since people conveniently forget that lenses tend to get worse as they get wider, and conveniently forget that the APS-C image is enlarged 1.6x as much when you actually make prints.) Oh, yes. You'll find the 17-40 on the 5D better than the 10-22 on the 50D, especially when you stop down an extra stop on the 5D for the same DOF. (There's a review somewhere that demonstrated how suprisingly funky the 10-22 is on the 20D; funkyness that's going to be a lot more disturbing on the 50D than it was on the 20D. (The 5D + 17-40 was sharper at every point in the frame at every f stop (it compared the same f stop with both) and the review didn't even try f/16 on the 17-40.)) SignatureDavid J. Littleboy Tokyo, Japan > The full frame 5D requires either the very best film lenses you can find to > maximize performance or new lenses specifically designed for a FF digital > sensor. Nikon has some, Canon has few if any. Let me give you a little hint about reality. The 5D is now fairly old. Many Canon 1.6 crop cameras have much, mush smaller pixels. Those cameras require much better lenses than does the 5D. And there is no such thing as a "digital" lens. Given the constraint of the back focus for SLR lenses, and the mount diameter, and normally designed microlenses, talk of "the exit pupil is too close" is silly. Resolution and MTF are still king and queen. Now when you start talking about 5D successors with 36 megapixels, now you're talking the very best lenses. And canon has those in droves, especially in the critical (to sports pros) tele range. Yes, it does have some lenses that are not up to the capability of a 36 megapixel FF camera .... but so does Nikon. Doug McDonald > > The full frame 5D requires either the very best film lenses you can find to > > maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 17 lines] > > Doug McDonald Canon does have some good telephoto lenses. Too bad about the wider ones. The idea of the digital lens is that the light beam is as perpendicular to the plane of the sensor as possible to avoid optical problems created when it is not. Olympus (granted, easier for them owing to sensor size) has made lenses to this "ideal" since they released their E-1 in 2003. Nikon is now doing the same. You can SEE the differences in their images compared to the old film lenses, even some so-called lenses designed for digital sensors (to cover a 1.5 crop instead of FF) are not ideally suited to the task. If a lens is incapable of supporting the edge when fully open or near fully open, IMO, it is not a good lens for digital. So, this example I've posted before. Shows a Nikon 18-70mm lens (DX lens) and an Olympus 14-42mm lens, both wide open on an Olympus body. Notice that despite being made for a 1.5 sensor, the Nikon cannot even support the edge properly on the 4/3rds sensor? Because it wasn't produced as a digital lens should be. The Olympus lens does support the edge very well. http://www.pbase.com/andersonrm/image/93686092 > Canon does have some good telephoto lenses. Too bad about the wider > ones. The idea of the digital lens is that the light beam is as > perpendicular to the plane of the sensor as possible to avoid optical > problems created when it is not. There is a BIG downside to such lenses. That is that these lenses (telecentric) have the apparent exit pupil at a "infinite" distance from the camera. This means that the light from the center of the exit pupil is perpendicular to the sensor. Now think about that: say you have a lens mount that equals the sensor diagonal in diameter. At the corners the light from the center of the lens just misses hitting teh flange, so you get huge vignetting. Consider a full frame camera with a 45 mm diagonal sensor. To get no vignetting with an f/2 lens you would need a flange inner diameter of 45 + (flange distance from sensor) / 2 millimeters. For your normal full frame SLR that a very big flange, and a very big final element for the lens! $$$$$$$$$$$$$$ Doug McDonald On Sep 8, 2:12 pm, u...@domain.invalid wrote: > > Canon does have some good telephoto lenses. Too bad about the wider > > ones. The idea of the digital lens is that the light beam is as [quoted text clipped - 11 lines] > the light from the center of the lens just misses hitting teh flange, > so you get huge vignetting. It hasn't worked out that way with some brands (notably Olympus) but for FF, Canon has a large bayonet so it shouldn't be an issue. >> > The full frame 5D requires either the very best film lenses you can find to >> > maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 36 lines] >the edge very well. >http://www.pbase.com/andersonrm/image/93686092 There is something wrong with the Nikon image. Did the camera move during the exposure? Eric Stevens >>>> The full frame 5D requires either the very best film lenses you can find to >>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 37 lines] > There is something wrong with the Nikon image. Did the camera move > during the exposure? It's the usual fallacy of conducting a test to /prove/ a conclusion that you've already formed. The Nikkor @18mm ~= 27mm (equivalent focal length on 1:1.5 crop) The Olympus lens @17mm ~= 34mm (equivalent focal length on 4/3) The Nikkor is at the extreme of zoom range, the Olympus lens isn't. So calling that an Olympus vs Nikon "kit lens test" is just a very silly thing to do. >>>>> The full frame 5D requires either the very best film lenses you can find to >>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 45 lines] >So calling that an Olympus vs Nikon "kit lens test" is just a very silly >thing to do. There is more than that to it in this case. All the lights in the Nikon image have a small tail. All the tails seem to be of the same length and pointing in the same direction. If the tails were caused by a problem in the lens I would expect them be pointing to the centre of the lens and and of shorter length closer to the centre. Neither of these aspects are present. Camera movement seems to be the logical explanation. Eric Stevens > >>>>> The full frame 5D requires either the very best film lenses you can find to > >>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 55 lines] > > Eric Stevens Read the caption, the image is cropped from the lens center to one edge, not across the field. >> >>>>> The full frame 5D requires either the very best film lenses you can find to >> >>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 58 lines] >Read the caption, the image is cropped from the lens center to one >edge, not across the field. I did read it. My comments still stand, unless the image is taken from a small segment a long way from the centre. Eric Stevens >>>>>>>> The full frame 5D requires either the very best film lenses you can find to >>>>>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 54 lines] > I did read it. My comments still stand, unless the image is taken from > a small segment a long way from the centre. It is an edge crop. SignaturePaul Furman www.edgehill.net www.baynatives.com all google groups messages filtered due to spam >>>>>> The full frame 5D requires either the very best film lenses you can find to >>>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 52 lines] > these aspects are present. Camera movement seems to be the logical > explanation. I believe that is a lens aberration: coma, or 'batwing' sagittal oblique spherical aberration (yeah I googled that). I was fiddling with a 45mm Nikkor (on FX) yesterday with a 14mm extension tube for closeup and saw something I thought was motion blur but in fact I think it was coma: http://edgehill.net/Misc/misc-photos/9-7-08-sushi/full-set/pg1pc2 -click for full crop: http://edgehill.net/1/Misc/misc-photos/2008-09-07-sushi/full-set/crop/_0000251.jpg My samples are pushing the intended use of this lens pretty far and it's an ancient lens design. The 58mm f/1.2 noct was specially designed to overcome this issue and it costs a ridiculous price these days (discontinued). But yeah, I think Nikkors tend to have this issue: something in their designs that optimizes sharpness at normal apertures at the cost of these sorts of aberrations. And, yeah there is a big difference in Rich's comparison, being at the edge of the zoom's range and same focal length with a different crop factor (the Nikkor is magnified): surely 17mm & 18mm are not comparable? Here's an example of motion blur with a very similar shape: http://edgehill.net/Misc/misc-photos/9-7-08-sushi/full-set/pg1pc1 -click for full crop: http://edgehill.net/1/Misc/misc-photos/2008-09-07-sushi/full-set/crop/_0000250.jpg SignaturePaul Furman www.edgehill.net www.baynatives.com all google groups messages filtered due to spam > I believe that is a lens aberration: coma, or 'batwing' sagittal oblique > spherical aberration (yeah I googled that). I was fiddling with a 45mm [quoted text clipped - 3 lines] > -click for full crop: > http://edgehill.net/1/Misc/misc-photos/2008-09-07-sushi/full-set/crop/_0000251.jpg Motion blur? Signature-- r.p.e.35mm user resource: http://www.aliasimages.com/rpe35mmur.htm -- r.p.d.slr-systems: http://www.aliasimages.com/rpdslrsysur.htm -- [SI] gallery & rulz: http://www.pbase.com/shootin -- e-meil: Remove FreeLunch. -- usenet posts from gmail.com and googlemail.com are filtered out. > >>>> The full frame 5D requires either the very best film lenses you can find to > >>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 45 lines] > So calling that an Olympus vs Nikon "kit lens test" is just a very silly > thing to do. Not really, the Nikon is supposed to be able to support a 1.5 crop sensor to its edge, which is far wider than a 4/3rds crop sensor which was asked to support, but couldn't very well. The extreme of the zoom range is no guarantee of poor performance, often, the poorest performance is on the long end. The Olympus lenses just happen to work very well at controlling aberrations including CA, coma and spherical aberration. > >> > The full frame 5D requires either the very best film lenses you can find to > >> > maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 41 lines] > > Eric Stevens That coma you see, not movement, it turns the edge point sources into teardrops or comet-shaped things. >>>>> The full frame 5D requires either the very best film lenses you can find to >>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 38 lines] > That coma you see, not movement, it turns the edge point sources into > teardrops or comet-shaped things. And it's not something you see with the 18-70 Nikkor, with several (3?) ED elements. Your "test" is a fraud. > >>>>> The full frame 5D requires either the very best film lenses you can find to > >>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 42 lines] > ED elements. > Your "test" is a fraud. You brainless bumpkin, ED glass ONLY reduces effects colour error it has no effect on other aberrations! Moron. >>>>>>> The full frame 5D requires either the very best film lenses you can find to >>>>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 42 lines] > You brainless bumpkin, ED glass ONLY reduces effects colour error it > has no effect on other aberrations! Moron. My apologies. Aspherical elements reduce such aberrations. There's colour aberrations as well as coma spherical aberration in your sample shot. That's _not_ typical of the Nikkor 17-80. That lens has a few faults for sure, but not those faults. In any case, it's nuts comparing two wide zooms designed for different systems when comparing one at the extreme wide end, the other toward middle of the range. Are you nuts? > You brainless bumpkin, ED glass ONLY reduces effects colour error it > has no effect on other aberrations! Moron. You are not necessarily a moron, but that statement is not true. I actually have designed (and had constructed!) custom lenses for scientific purposes. ED glass effects EVERYTHING. This is because it has strange refractive index/dispersion and dispersion/secondary dispersion ratios, and generally (not always) is very low refractive index. Changing the refractive index always requires a major lens redesign. **IF** ED glass had the same refractive index and dispersion as regular glass, and only a different secondary dispersion, then, indeed, what you say would be true. But that is not the case. Doug McDonald >> >> > The full frame 5D requires either the very best film lenses you can find to >> >> > maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 44 lines] >That coma you see, not movement, it turns the edge point sources into >teardrops or comet-shaped things. But why does it appear to be constant in magnitude and direction across the image? Eric Stevens >>>>>> The full frame 5D requires either the very best film lenses you can find to >>>>>> maximize performance or new lenses specifically designed for a FF digital [quoted text clipped - 40 lines] > But why does it appear to be constant in magnitude and direction > across the image? Because it's an edge crop. The right side doesn't show as much. SignaturePaul Furman www.edgehill.net www.baynatives.com all google groups messages filtered due to spam > And there is no such thing as a "digital" lens. But there is! Take a very long, thin tube. Place a grounded screen at one end and tape that to your camera instead of an "analog" lens. Point the other end towards your target. Afterwards, reduce the resulting image to either 0% (black) or 100% (white) and a pixel size of 1. That's a digital lens! -Wolfgang |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2008 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.