 |
 | 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 2006Canon Response to 5D Dark Corners
For those of you that have been following the thread entitled "Dark Corners on Canon 5D" here's the response I got back from Canon: "This is not a product fault, but an unavoidable limitation of the camera technology when used with lenses set at extreme aperture settings. The Vignetting occurs as the camera's sensor exceeds the lens visible image circle at these extreme aperture settings. We hope this information is of use to you. However, should you require any further technical assistance, please do not hesitate to contact us quoting the incident number displayed in the subject of this email." > For those of you that have been following the thread entitled "Dark > Corners on Canon 5D" here's the response I got back from Canon: [quoted text clipped - 7 lines] > contact us quoting the incident number displayed in the subject of > this email." This is why it's totally and utterly foolish to buy anything other than a Nikon D2x or D200. Nikon wouldn't dream of creating this distress towards their customers by offering mediocre lenses. Canon should have never rushed to production with the 5D until they had suitable lenses to work with it. Rita >> For those of you that have been following the thread entitled "Dark >> Corners on Canon 5D" here's the response I got back from Canon: [quoted text clipped - 12 lines] >their customers by offering mediocre lenses. Canon should have never rushed >to production with the 5D until they had suitable lenses to work with it. Oh, I would say Canon has very good lenses which work very well with both the 5D and the 1DsMkII. It's a shame Nikon doesn't have full frame sensor for their lenses so we could see how they work when they don't just crop out of the sweet spot. ***************************************************** "Vietnam is what we had instead of happy childhoods." Tim Page in "Dispatches" by Michael Herr > Oh, I would say Canon has very good lenses which work very well with > both the 5D and the 1DsMkII. It's a shame Nikon doesn't have full > frame sensor for their lenses so we could see how they work when they > don't just crop out of the sweet spot. I would like to see Nikon go FF as well, but Nikon realizes the inherent problems with it using current technology. Nikon refuses to release a substandard format that is guaranteed to cause everyone grief. My comment wasn't meant to be a slam on Canon, it was meant to demonstrate my aggravation for a company that would write a piss poor response to a customer that paid a premium price for an FF body and professional lenses. It's bullshit to expect a customer to massage an image in Photoshop to correct hardware deficiencies that shouldn't be there in the first place. Rita >> Oh, I would say Canon has very good lenses which work very well with >> both the 5D and the 1DsMkII. It's a shame Nikon doesn't have full [quoted text clipped - 11 lines] > > Rita I agree. I've no preference between Canon or Nikon. Canon have let me down by not advising consumers of the design defect they have with the 5D. As long as I can get a refund, and I've spent nearly £4,000 on new Canon products in the last 2 months, then I'll revert to the Nikon D-200. Hi, The vignetting can be removed in Adobe Camera RAW by eye but it would be nicer if this could be done in a more rigorous way ..... i.e. ACR determining the lens type and applying a defined correction according to the focal length used. The concept would be similar to the lens-dependent correction for pincushion/barrel distortion that PTLens does. James http://www.sns.ias.edu/~jns/wp/category/photography/ > I agree. > I've no preference between Canon or Nikon. [quoted text clipped - 3 lines] > Canon products in the last 2 months, then I'll revert to the Nikon > D-200. As others have mentioned, this is a limitation of the lenses and was present when using film bodies. That being said, it's poor justification to try to pass these deficiencies off in a *newer* high tech system when one could have simply used their old film body and scanned the images afterwards. Canon should have ironed out these problems prior to the 5D's release. Doing otherwise is a great disservice to their loyal customers. Rita Rita Ä Berkowitz FUDs away: > As others have mentioned, this is a limitation of the lenses and was present > when using film bodies. That being said, it's poor justification to try to > pass these deficiencies off in a *newer* high tech system when one could > have simply used their old film body and scanned the images afterwards. > Canon should have ironed out these problems prior to the 5D's release. > Doing otherwise is a great disservice to their loyal customers. So you argue that Canon must deep-six their entire lens set and re-tool for replacements prior to the release of a full-frame sensor ... just to solve a problem that (a) everyone had been dealing with with film cameras for, literally, decades, and (b) has trivial solutions with a digital camera? > So you argue that Canon must deep-six their entire lens set and > re-tool for replacements prior to the release of a full-frame sensor > ... just to solve a problem that (a) everyone had been dealing with > with film cameras for, literally, decades, and (b) has trivial > solutions with a digital camera? I would argue that *ANY* company that continually tries to pass off decades old problems when they are charging premium prices for supposedly new technology. If Nikon did this I would be just as aggravated. Why should one buy a 5D when they can have the same problems with a film body and scan the image afterwards? So, yes, Canon should re-tool for improved lens design or should use a sensor that can deal with the problem. Then again, with current technology Canon could have buried this shortcoming in firmware and nobody would have noticed. Canon rushed the 5D out the door and some people rightfully don't want to have this problem. And yes, fixing in post is easy, but why should one have to when the problem shouldn't be there in the first place. Rita Rita Ä Berkowitz FUDs on and on: > > So you argue that Canon must deep-six their entire lens set and > > re-tool for replacements prior to the release of a full-frame sensor [quoted text clipped - 5 lines] > old problems when they are charging premium prices for supposedly new > technology. Feel free to argue thus. It's a stupid position, but hey, it's a free country. > If Nikon did this I would be just as aggravated. Of course, of course. > Why should > one buy a 5D when they can have the same problems with a film body and scan > the image afterwards? Don't you remember? You even said it yourself: film cameras have exactly the same problem. Why scan a vignetted, noisy, slide that costs money to buy and process, when you can capture a vignetted image and trivially correct it post-exposure? > So, yes, Canon should re-tool for improved lens > design or should use a sensor that can deal with the problem. Always easy spending other people's money, eh? > Then again, > with current technology Canon could have buried this shortcoming in firmware > and nobody would have noticed. Canon rushed the 5D out the door and some > people rightfully don't want to have this problem. And yes, fixing in post > is easy, but why should one have to when the problem shouldn't be there in > the first place. <insert sound of violin playing sad, sad, music> Oh, cry me a river! The only people complaining are a bunch of ignorant wannabes who appear to have no clue whatsoever. > Don't you remember? You even said it yourself: film cameras have > exactly the same problem. Why scan a vignetted, noisy, slide that > costs money to buy and process, when you can capture a vignetted image > and trivially correct it post-exposure? Because it's foolish to waste $3,000 on a "new technology" body and get crippled with 1960s results in your images. > Always easy spending other people's money, eh? It's easier spending money on something that should work better than a film camera of thirty years ago. > Oh, cry me a river! The only people complaining are a bunch of > ignorant wannabes who appear to have no clue whatsoever. I'm not crying anything since I'm currently using a Nikon platform. I could be using Canon since the decision for me was a mere coin toss since I feel Canon an Nikon are equally great and both have their own set of shortcomings. What wannabes are complaining? If I wanted a 5D I can simply pick up the old phone and call B&H tomorrow and it will be in my hands the following day. What will the 5D buy me that the Nikon D70, D200, and D2x don't already offer me? Rita Rita Ä Berkowitz FUDs on and on ad nauseum: > > Don't you remember? You even said it yourself: film cameras have > > exactly the same problem. Why scan a vignetted, noisy, slide that [quoted text clipped - 3 lines] > Because it's foolish to waste $3,000 on a "new technology" body and get > crippled with 1960s results in your images. I see. You are a moron then. Unless, of course, your suggestion of scanning film was not to be taken seriously, in which case you are an intellectually dishonest Nikon Nutcase. Par for the course. > > Oh, cry me a river! The only people complaining are a bunch of > > ignorant wannabes who appear to have no clue whatsoever. > > I'm not crying anything since I'm currently using a Nikon platform. Of course you are! > I could > be using Canon since the decision for me was a mere coin toss since I feel > Canon an Nikon are equally great and both have their own set of > shortcomings. Yes, yes, equally great, but you still feel that urge to justify your Nikon Nutcaseness by babbling on at odds with physical reality. You sound like the Rich kook: he can't buy a camera because they have that plastic stuff in them. Oh, he can afford them -- he is in the 36% tax bracket!! He just doesn't _want_ one! Now along comes you and other Nikon Nutcases and your tedious, wishywashy, FUD rationalizations about the 5D (or whatever object of envy has consumed your soul). Can't buy a camera that has vignetting, can we? Yes, yes! Vignetting! That's the reason! I'm sure of it! > What wannabes are complaining? I can't read USENET for you. > If I wanted a 5D I can simply pick up the old phone and call B&H tomorrow > and it will be in my hands the following day. What will the 5D buy me that > the Nikon D70, D200, and D2x don't already offer me? I don't give a sh.t what you buy or use. You are the one FUDstering loudly about how Canon should conduct its business. Is Canon telling you how to conduct yours? Hm? >> Don't you remember? You even said it yourself: film cameras have >> exactly the same problem. Why scan a vignetted, noisy, slide that [quoted text clipped - 3 lines] > Because it's foolish to waste $3,000 on a "new technology" body and get > crippled with 1960s results in your images. You know, Rita, I'm absolutely sick and tired of various and sundry people telling me, by implication, that I'm foolish to have spent this amount of money on a camera. That's one reason you and other (primarily) Nikon users get such a harsh reaction from Canon owners. You and others use inflammatory language calculated to get a reaction. This camera produces excellent images, the light fall off is a minor concern, and is not present on any of the "L" lenses except for the 16-35, which, admittedly, could be a bit better for the price. >> Always easy spending other people's money, eh? > > It's easier spending money on something that should work better than a > film > camera of thirty years ago. What about the Nikon F6, that just came out late last year.  SignatureSkip Middleton http://www.shadowcatcherimagery.com > You know, Rita, I'm absolutely sick and tired of various and sundry > people telling me, by implication, that I'm foolish to have spent [quoted text clipped - 5 lines] > except for the 16-35, which, admittedly, could be a bit better for > the price. I don't understand why you are sick of people telling you "X" about anything? This is Usenet and everyone has an opinion that may or may not agree with yours or mine. That said, you have a 5D and I know it is an extremely fine camera that is working great for tens of thousands of photographers around the world. I have various Nikon dSLR bodies and they do exactly what I need them to do. Is your Canon better than my Nikon or visa versa? Who cares since you and I both get the job done? Lighten up since none of this means sh.t to a tree. Enjoy life and have fun with this great technology, I know I am no matter what anybody's opinion of which platform I use. > What about the Nikon F6, that just came out late last year. I don't have one. Rita >> You know, Rita, I'm absolutely sick and tired of various and sundry >> people telling me, by implication, that I'm foolish to have spent [quoted text clipped - 21 lines] > > I don't have one. An opinion does not deride someone else's action or thought. You said, "Because it's foolish to waste $3,000 on a "new technology" body and get crippled with 1960s results in your images." So, by implication, you are saying that anyone who bought a 5D is foolish. This is not the first time that has been expressed. And it is calculated to bring a harsh response. You are not expressing an opinion of the camera, you are deriding the purchasers of that camera. That's personal. The point I'm making with the F6 is that it has the same issues with light fall off that you are taking Canon to task for, but it's ok, because it's a Nikon, or because it's film? SignatureSkip Middleton http://www.shadowcatcherimagery.com > The point I'm making with the F6 is that it has the same issues with light > fall off that you are taking Canon to task for, but it's ok, because it's a > Nikon, or because it's film? Because it's film. The problem is worse with digital sensors than it ever was with film, by all accounts. It seems that a reasonable "solution" is to use lenses with an image circle significantly larger than the sensor. Indeed, the only lenses I've seen noticeable falloff from on a Nikon are DX lenses. It's just one more reason not to hurry toward 35mm digital sensors. I don't see the need for them and there are so many disadvantages that I frankly hope Nikon just sticks to what works. SignatureJeremy | jeremy@exit109.com >> The point I'm making with the F6 is that it has the same issues with >> light [quoted text clipped - 14 lines] > see the need for them and there are so many disadvantages that I frankly > hope Nikon just sticks to what works. Not "by all accounts." I'm one who's unconvinced, and I have pretty extensive experience with film and a few of these same lenses. And many others feel the same way I do, most of them with the cameras in question, the 5D and 1Ds mkII. Most everybody else who is chiming in with "it's worse with digital than film" seems to not have one of the cameras in question. SignatureSkip Middleton http://www.shadowcatcherimagery.com > Most everybody else who is chiming in > with "it's worse with digital than film" seems to not have one of the > cameras in question. Yep, that's me. I don't have a 5D, but the recent pictures from the OP and other people complaining about the same problem makes me wonder. I guess, to be fair, it's in the same noise category as the D200 banding issues we are all subjected to endure? Rita >> Most everybody else who is chiming in >> with "it's worse with digital than film" seems to not have one of the [quoted text clipped - 5 lines] > to be fair, it's in the same noise category as the D200 banding issues we > are all subjected to endure? Pretty much. SignatureSkip Middleton http://www.shadowcatcherimagery.com >Because it's film. The problem is worse with digital sensors than it ever >was with film, by all accounts. No it isn't - see the thread entitled "Light fall off on dSLRs - an experiment" for proof. 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) >> Because it's film. The problem is worse with digital sensors than it ever >> was with film, by all accounts. > > No it isn't - see the thread entitled "Light fall off on dSLRs - an > experiment" for proof. Proof of absolutely nothing, since you didn't use a lens. The thing is that digital sensors are more sensitive to the angle of incidence of light than film. The extreme angle of incidence is caused by the lens design, and happens more with wide-angle lenses. Any kind of test of this would need to use a lens or it means nothing at all. SignatureJeremy | jeremy@exit109.com >>> Because it's film. The problem is worse with digital sensors than it ever >>> was with film, by all accounts. [quoted text clipped - 3 lines] > >Proof of absolutely nothing, since you didn't use a lens. That was the whole point! Eliminate the effect of the lens and measure the sensitivity of the sensor to angle of incidence. >The thing is that digital sensors are more sensitive to the angle of >incidence of light than film. That is exactly what the test disproved - I used two angles: perpendicular to the sensor and as extreme an angle of incidence as the Canon lens mount permits. There was essentially no difference in sensor sensitivity between these two angles of incidence. So, unless you are arguing that film has a rather remarkable property of increasing its sensitivity with angle of incidence to compensate for lens vignette (which would have caused huge problems with lenses which didn't vignette!) then this alleged difference in sensitivity between film and digital is complete bunk. >The extreme angle of incidence is caused >by the lens design, and happens more with wide-angle lenses. The angle of incidence at the focal plane cannot exceed the angle from the edge of the lens mount - in fact, it has to be less than this due to the physical mount itself. >Any kind >of test of this would need to use a lens or it means nothing at all. Having the lens present is what causes the confusion, since the lens itself also introduces light fall-off, simple vignetting that has been present in every lens since before Galileo first put two of them together to create his telescope. The argument that many, including yourself above, have made is that the sensor itself has some sensitivity to the angle of incidence of the light. That is exactly what the test was intended to measure - how sensitive is the sensor to this angle of incidence? That is why the LED was positioned almost a metre from the focal plane - so that variation of the angle of incidence across the focal plane was negligible and controlled purely by rotating the camera on the tripod head. The answer appears to be is that sensor isn't sensitive to angle of incidence at all! To be honest, I was expecting this to be a lot more complicated and have to take into account gamma response and all sorts of other effects to compare the results quantitatively. However, the numbers came out so close that any other response effects would have made no difference. In short, there is *NO* response variation due to angle of incidence on the sensor at all. Light fall-off at the edges and corners is *all* due to lens vignetting. This is exactly what others who have experience of using the Canon full frame sensors have been saying throughout this and other threads on the topic. All my test has done is quantify that - removing the effect of lens vignetting and proving once and for all that the sensor is no different to film in this respect. And it is so simple to do that anyone with a dSLR can validate it - you don't need an LED with a current source, and fixed point source of light will do, even a flashgun at a suitably large distance. 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 exactly what the test disproved - I used two angles: > perpendicular to the sensor and > as extreme an angle of incidence as the Canon lens mount permits. > There was essentially no difference in sensor sensitivity between these > two angles of incidence. But you didn't simulate the light that falls on the sensor from a lens. You actually seem to have tried to make it a point light source from the sensor's perspective, which is not correct. I don't think you understand the problem, let alone how to test for it. > Having the lens present is what causes the confusion, since the lens > itself also introduces light fall-off, simple vignetting that has been > present in every lens since before Galileo first put two of them > together to create his telescope. That's why the "control" should be an image shot on film with the same lens. SignatureJeremy | jeremy@exit109.com >> That is exactly what the test disproved - I used two angles: >> perpendicular to the sensor and [quoted text clipped - 3 lines] > >But you didn't simulate the light that falls on the sensor from a lens. Exactly, because the argument that you have made is that it isn't the lens that causes the problem, but that the *sensor* somehow has a different response to angle of incidence than film has. So the test attempted to measure how different that response actually is in a real sensor. Whether the light comes from a lens, the sun or a point source is irrelevant. Your, and others, argument has been that there is a fundamental angular response variation in the sensor. The test proved that if there is it is irrelevant and that essentially *all* of the light fall off at the edges and corners of the field *MUST* be due to the lens and therefore no worse on full frame dSLRs than it was in full frame film dSLRs. >You actually seem to have tried to make it a point light source from the >sensor's perspective, which is not correct. Yes it is - a point source placed a significant distance from the focal plane ensures that ALL of the light is incident at every pixel on the focal plane at almost the same angle. This angle can then be changed by rotating the camera relative to the light source and any difference in response measured. > I don't think you understand >the problem, let alone how to test for it. I understand the problem and how to separate the two components of it. Nobody is arguing that dSLR sensors and lens combinations do not produce vignetting at the corners and edges of the field - that is obvious and there are web sites where this is quantified for different lenses and apertures. However the question is how much of that is due to the lens (and therefore common to the format, whether film or digital) and how much is due to the angular sensitivity of the sensor (and therefore possibly different between film and digital on the same format). The test assessed the latter comparison, showing negligible difference, and thus showed that all of the light fall off *is* due to the lens, not the sensor. Since you have consistently argued that the sensor is somehow defective, it is clear that you don't understand the problem or its cause! >> Having the lens present is what causes the confusion, since the lens >> itself also introduces light fall-off, simple vignetting that has been >> present in every lens since before Galileo first put two of them >> together to create his telescope. > >That's why the "control" should be an image shot on film with the same lens. I don't think you understand what you are claiming. You have said the sensor has a different response to angle of incidence of light. The test shows, quite specifically, that this is wrong, there is no fall-off in response within the angles of incidence that a lens mounted on a Canon body can produce. It doesn't matter to the sensor where the light comes from, only its angle of incidence. The "control" is therefore a measurement at perpendicular incidence - this can be achieved in a long focal length lens which is stopped down to an extreme level, or with a point source a large distance from the focal plane. The "test" is whether changing that angle of incidence produces a different response from the sensor. you cannot change that angle of incidence in the lens without introducing other effects, such as the variation in lens pupil across the focal plane. The point source can be positioned at any angle. Between the two extreme cases of perpendicular and as deep as any lens mounted on a Canon SLR can produce, there was effectively no difference. 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) >> But you didn't simulate the light that falls on the sensor from a lens. > > Exactly, because the argument that you have made is that it isn't the > lens that causes the problem, but that the *sensor* somehow has a > different response to angle of incidence than film has. It's not the argument I have made; it's the argument I have thus far accepted. There is a distinction. I cannot prove it. I go on the statements of people who know more about it than I do, mixed in with simple observation. > So the test attempted to measure how different that response actually > is in a real sensor. Whether the light comes from a lens, the sun or > a point source is irrelevant. Why? The light is falling on the sensor entirely differently in your test than it does in actual usage. Yes, your test eliminates the effect of falloff from the lens, but it does so by changing too many variables for me to accept it as valid, when observation shows something else. Eliminate lens falloff by comparing a shot on film using the same lens. We *know* there is falloff in wide-angle lenses (especially), and we know why it happens, so its effect can be subtracted from a test by changing only the variable we're trying to test -- the sensor. The light hitting a pixel on the sensor is not all coming from the same angle, in actual usage. The angle varies; you know this. So, in order to accept your results I'd like either an explanation, from a sensor design perspective, of why this doesn't matter with regard to the problem at hand (angle of incidence), or I'd like a test that doesn't change that fact, done using a lens. > Your, and others, argument has been that there is a fundamental angular > response variation in the sensor. The test proved that if there is it > is irrelevant and that essentially *all* of the light fall off at the > edges and corners of the field *MUST* be due to the lens and therefore > no worse on full frame dSLRs than it was in full frame film dSLRs. Except that simple observation says otherwise -- the problem appears to be more pronounced with digital than with film. And everyone has been saying that the angle of incidence has more effect with digital than with film, which nicely matches up with real-world observation. Your test has not convinced me otherwise. > Yes it is - a point source placed a significant distance from the focal > plane ensures that ALL of the light is incident at every pixel on the > focal plane at almost the same angle. But that's exactly my point. This doesn't happen in real-world usage with a lens, and thus we have another change between the test and the real world, which has not been accounted for in the test. > However the question is how much of that is due to the lens (and > therefore common to the format, whether film or digital) and how much > is due to the angular sensitivity of the sensor (and therefore possibly > different between film and digital on the same format). So why not compare film with digital on the same format? > Since you have consistently argued that the sensor is somehow defective, > it is clear that you don't understand the problem or its cause! Not defective, but simply different. If the contention is that the difference between film and digital causes the visible effect, the valid test is film versus digital, not something else entirely; you've tested something else entirely, with no justification to discount the additional variables. SignatureJeremy | jeremy@exit109.com SNIP > Nobody is arguing that dSLR sensors and lens combinations do not > produce vignetting at the corners and edges of the field - that is > obvious and there are web sites where this is quantified for > different lenses and apertures. Indeed, and I already posted an example of quantified vignetting-analysis some 6 months ago: <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/Vignetting_50mm.png> <http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/Vignetting_50mm_8.png> This is (obviously) a combined response of lens and camera body (1Ds Mark II in this case). The f/8 chart shows that there is some residual vignetting that won't reduce much with even smaller apertures, despite the (by now) mostly circular base shape of the lens' exit-pupil cone-of- light to the corners. > However the question is how much of that is due to the lens (and > therefore common to the format, whether film or digital) and how > much is due to the angular sensitivity of the sensor (and therefore > possibly different between film and digital on the same format). IMHO there are three major factors involved, in order of significance: 1. Lens design 2. Projection path length (=magnification) 3. Angle of incidence to capture medium. Ad 1. It is obvious from looking at the lens' exit pupil at an angle, that the wide-open image is restricted to a sort of an oval/ellipse by design. Stopping down will improve vignetting performance to a more circular exit pupil even when viewed from an off-axis angle. Ad 2. The distance of, say, an average cone of light from a 50mm lens (like in the example above) would have to travel 50mm on axis, and an average 54.5mm to the extreme corners, in other words it would lose 15.8% (square of delta distance due to magnification), or -0.249 stops (very close to the -0.244 stops mean of the f/8 linked above shot). Seems like an inevitable boundary for full 24x36mm frames to me. Ad 3. This is expected to be small on sensor arrays fitted with micro-lenses. Bart SNIP > It seems that a reasonable "solution" is to use lenses with an > image circle significantly larger than the sensor. That by itself doesn't reduce light fall-off. In fact my TS-E 45mm f/2.8 (image circle diameter of 58.6mm) produces more vignetting than my EF 50mm f/1.4 (image circle of approx. 44? mm) at similar apertures and centered on optical axis. I'll repeat; it's the lens *design* that determines the amount of light fall-off, within the physical/economical boundaries of a given lens. Bart >> What about the Nikon F6, that just came out late last year. > >I don't have one. You don't have a 5D either, but that didn't stop you shooting your mouth off about an issue for which it is no better or worse. 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) >You know, Rita, I'm absolutely sick and tired of various and sundry people >telling me, by implication, that I'm foolish to have spent this amount of >money on a camera. In the scheme of things, $3000 is almost a trivial amount of money, especially if, as in your case, you are using it for your business. To put it in perspective, it is less than one year's depreciation on one of my cars. I think the 5D represents outstanding value for money. I have had the focusing problems fixed under warranty (I took the 5D and the lens to the repairer at 8:30 AM and collected it at 4:00 PM) and it now works faultlessly. >>You know, Rita, I'm absolutely sick and tired of various and sundry people >>telling me, by implication, that I'm foolish to have spent this amount of [quoted text clipped - 9 lines] > the repairer at 8:30 AM and collected it at 4:00 PM) and it now works > faultlessly. That's part of my point, it does great for what it is. It does it better, for me and my wife, than any other camera on the market, with the possible exception of the 1Ds mkII, certainly better than any camera at its price point or less. The so-called light fall off "issue" is a non issue in everyday usage. What focusing problems were you having, Tony? SignatureSkip Middleton http://www.shadowcatcherimagery.com >>>You know, Rita, I'm absolutely sick and tired of various and sundry people >>>telling me, by implication, that I'm foolish to have spent this amount of [quoted text clipped - 16 lines] >everyday usage. >What focusing problems were you having, Tony? I had focusing problems with the 50mm f/1.4. Focus was erratic so I took it, the 24-105 and the 5D to my friendly repairer. He checked all three, obtained a replacement 50mm lens (under warranty) and re-calibrated the body AF system. Now it focuses just fine. I use manual focus most of the time, but it's nice to know that the AF system works for those occasions when I might need it. I didn't ask precisely what was wrong, sorry. >>What focusing problems were you having, Tony? > [quoted text clipped - 6 lines] > it's nice to know that the AF system works for those occasions when I > might need it. I didn't ask precisely what was wrong, sorry. That's close enough, at least for my curiosity... SignatureSkip Middleton http://www.shadowcatcherimagery.com Rita D Berkowitz wrote: > > So you argue that Canon must deep-six their entire lens set and > > re-tool for replacements prior to the release of a full-frame sensor [quoted text clipped - 5 lines] > decades old problems when they are charging premium prices for > supposedly new technology. Physics can be a real PITA at times. SignatureAlbert Nurick | Nurick + Associates - Web Design albert@nurick.com | eCommerce - Content Management www.nurick.com | Web Applications - Hosting > I would argue that *ANY* company that continually tries to pass off > decades [quoted text clipped - 11 lines] > is easy, but why should one have to when the problem shouldn't be there in > the first place. But Nikon does _exactly_ this when they sell a film camera. Remember, the F6 is a new design. So they, and everybody else who manufactures 35mm film bodies are just as "bad" as Canon! Sheesh, get a grip. The people who don't want this "problem" have options. Canon 30D for one, or even the 1D mkIIn. Or a camera not made by Canon. SignatureSkip Middleton http://www.shadowcatcherimagery.com > But Nikon does _exactly_ this when they sell a film camera. Yeah, I'm just checking the nikon 28/2.8 I it's about like Edward's shots on a comparable canon lens altrhough the nikon lens seems about half the price, only $250. http://www.olegnovikov.com/technical/28mmf28d/28mmf28d.shtml I'm so used to APS dSLRs that I'm not really familiar with the problem. >> But Nikon does _exactly_ this when they sell a film camera. > [quoted text clipped - 3 lines] > http://www.olegnovikov.com/technical/28mmf28d/28mmf28d.shtml > I'm so used to APS dSLRs that I'm not really familiar with the problem. Wow, that looks bad at f2.8. I wonder if Rita will look at it... SignatureSkip Middleton http://www.shadowcatcherimagery.com >Wow, that looks bad at f2.8. I wonder if Rita will look at it... The article doesn't seem to specify which slide film was used and whether the scans were corrected for the rather high contrast produced by most slides. It does look a lot worse then I am used to seeing in Nikon lenses. Usually, Nikon seems to be willing to trade resolution for even illumination. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency > Wow, that looks bad at f2.8. I wonder if Rita will look at it... Yes, I did. It's time for Nikon to re-tool their lens line as well if they want to offer FF. As I said earlier, this is probably why Nikon won't offer FF. The nice thing about APS is it makes a mediocre lens acceptable. Rita >> Wow, that looks bad at f2.8. I wonder if Rita will look at it... > >Yes, I did. It's time for Nikon to re-tool their lens line as well if they >want to offer FF. As I said earlier, this is probably why Nikon won't offer >FF. The nice thing about APS is it makes a mediocre lens acceptable. Canon sure proved that! >>> Wow, that looks bad at f2.8. I wonder if Rita will look at it... >> [quoted text clipped - 3 lines] > >Canon sure proved that! The way some people talk about center crops, you'd think that they believe that lenses have infinite resolution in the center, and become finite near the edges of a 43mm-diameter frame. The resolution in the center is finite, and loses 37.5% in terms of the total image with a 1.6x crop. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >> Canon rushed the 5D out the door and some >> people rightfully don't want to have this problem. And yes, fixing in [quoted text clipped - 4 lines] >But Nikon does _exactly_ this when they sell a film camera. Remember, the >F6 is a new design. The F6 is for people who want to shoot film. I don't think Nikon has to fix problems with film. The lack of quality of commonly used films is not Nikon's problem. :-) But what find strange about the comment about the 5D is the suggestion that the 5D is somehow different when it comes to light fall-off than the 1Ds. As far as I know, the light fall-off is a lens problem. And the 1Ds shows the same problems. It doesn't have anything to do with the design of the 5D. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency >>> Canon rushed the 5D out the door and some >>> people rightfully don't want to have this problem. And yes, fixing in [quoted text clipped - 9 lines] > fix problems with film. The lack of quality of commonly used films is not > Nikon's problem. :-) But the quality of the film is not the issue, it's the ability of the lens to cover the film frame that is. I don't have a problem with Nikon having light fall off on film cameras (that's the only "full frame" camera they build, currently) what I'm trying to do is point out to Rita that while she's on her high horse, in a high dudgeon about how Canon is mistreating their customers by not offering a full line of lenses that address a problem that has occured with film cameras for decades, Nikon is doing the exact same thing with the F6. They have not intruduced a single lens that addresses this problem, probably because they feel that the purchasers of that camera and the matching lenses won't feel it's a problem severe enough to wail about. And they're probably right. That's the way many 5D and 1Ds/mkII owners feel, too. > But what find strange about the comment about the 5D is the suggestion > that [quoted text clipped - 4 lines] > > It doesn't have anything to do with the design of the 5D. The whole thing is a tempest in a teapot. I think that one reason there was so little bitching about the 1Ds is that camera's users had more, and more recent, experience with film, and the images looked pretty much like what they were used to, in the past. I know photographers for whom that was their first digital camera, or at most, their second. Moving from a 1Ds to a 1Ds mkII produced no changes, so it was no big deal. I think that many of the 5D users, and the predominance of those either contemplating such a move, or commenting on it, are coming from crop sensor cameras, and have little to no experience with film SLRs. I shot for many years with film cameras, and see little difference between what I get with wide angle lenses and the 5D and what I used to get with my Exacta, Canon FD mount and Canon EF mount film cameras. SignatureSkip Middleton http://www.shadowcatcherimagery.com >I think that many of >the 5D users, and the predominance of those either contemplating such a >move, or commenting on it, are coming from crop sensor cameras, and have >little to no experience with film SLRs. I agree, and have just conducted a simple test that anyone can try which, as far as I am concerned, proves beyond any further argument that the issue is entirely lens based and therefore exactly the same with these full frame digital sensors as it was with full frame 35mm film. See the thread entitled: "Light fall-off with dSLRs - an experiment". The problem, in as much as it exists, is entirely lens vignetting and has been present in 35mm cameras since Oscar Barnack invented the format. 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) > Rita Ä Berkowitz FUDs away: > [quoted text clipped - 8 lines] > So you argue that Canon must deep-six their entire lens set and re-tool > for replacements prior to the release of a full-frame sensor ... Yep. But then they couldn't claim to have a "deep line" of lenses. SignatureStacey >> I agree. >> I've no preference between Canon or Nikon. [quoted text clipped - 12 lines] > Canon should have ironed out these problems prior to the 5D's release. > Doing otherwise is a great disservice to their loyal customers. If this was a problem with the film bodies too then it is not the camera or sensor at fault the fault was in the lense the whole time. The only cure would be a sensor smaller than full frame which sort of defeats the object. > If this was a problem with the film bodies too then it is not the > camera or sensor at fault the fault was in the lense the whole time. The > only cure would be a sensor smaller than full frame which sort of > defeats the object. BINGO!!!! We have a winner! As much as I would like to see Nikon go FF they won't because they realize they will have the same shortcomings as Canon's 5D. I respect a company that won't trash their customers for the sake of a novelty. Rita >> If this was a problem with the film bodies too then it is not the >> camera or sensor at fault the fault was in the lense the whole time. The [quoted text clipped - 5 lines] >Canon's 5D. I respect a company that won't trash their customers for the >sake of a novelty. Some Nikon lenses show hardly any light fall-off. For example this is the 24/2.0 (wide open): <http://misc.hq.phicoh.net/tmp/m.3-96-6-28.jpg> It would be stupid not to make a full-frame DSLR just because some lenses show a bit of a problem wide open. With argument, you just as well drop all high resolution sensors as well, because there are plenty of lenses that are not all that sharp wide open. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency >Some Nikon lenses show hardly any light fall-off. > >For example this is the 24/2.0 (wide open): ><http://misc.hq.phicoh.net/tmp/m.3-96-6-28.jpg> Oh my God; please tell me this was a joke! You chose a subject that is brighter at the edges to demonstrate? Regardless; I see dark corners. Don't forget, anything that lowers contrast will hide roll-off as well. white highlights are often compressed, relative to the actual capture. IOW, based on this image, it is still possible that this lens still has significant roll-off. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< > If this was a problem with the film bodies too then it is not the camera or > sensor at fault the fault was in the lense the whole time. Yes, except that the problem I've seen from the Canon 35mm-sensor cameras is far, far worse than I ever saw on film. So it seems there is something to the idea that it matters more with digital than film. > The only cure would be a sensor smaller than full frame which sort of > defeats the object. And solves the problem. SignatureJeremy | jeremy@exit109.com >> If this was a problem with the film bodies too then it is not the camera or >> sensor at fault the fault was in the lense the whole time. > >Yes, except that the problem I've seen from the Canon 35mm-sensor cameras is >far, far worse than I ever saw on film. So it seems there is something to >the idea that it matters more with digital than film. Of course it is far worse. Film will accept light from almost any angle a lens will supply to it, whereas digital sensors, even with microlenses, suffer from a reduction in light accepted by the sensor as the angle of incidence departs from 90 degrees. Think of it like an LCD monitor which has a certain maximum viewing angle, beyond which almost no light gets transmitted. That's why a Canon 24mm or 28mm wide angle lens (for example) will display far greater light fall-off on a full frame digital sensor than it will on film. There is still some light fall-off with film, but it is far less significant. The answer is to use lenses that are near-telecentric in design, such as the Canon EF 24-105mm f/4L. The light fall-off is significantly less on the 5D than the 24mm or 28mm lenses, because its optical design was optimised for digital use. > Of course it is far worse. Film will accept light from almost any > angle a lens will supply to it, whereas digital sensors, even with > microlenses, suffer from a reduction in light accepted by the sensor > as the angle of incidence departs from 90 degrees. I know that; but it's been argued here that this *isn't* true, especially with microlenses. SignatureJeremy | jeremy@exit109.com >> Of course it is far worse. Film will accept light from almost any >> angle a lens will supply to it, whereas digital sensors, even with [quoted text clipped - 3 lines] >I know that; but it's been argued here that this *isn't* true, especially >with microlenses. The argument is usually that fast normal lenses (50/1.4) are not really different from retro-focus wide angle lenses when it comes to the angles of light. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency >> Of course it is far worse. Film will accept light from almost any >> angle a lens will supply to it, whereas digital sensors, even with [quoted text clipped - 3 lines] >I know that; but it's been argued here that this *isn't* true, especially >with microlenses. Then it has been argued incorrectly. Microlenses will improve the acceptance angle of a receptor, but only up to a point. Beyond that point, where the angle of incidence is greater than the microlenses can cope with, the fall-off is dramatic. And that's what the OP is seeing here. For some reason, the sensor EOS 5D appears more vulnerable to light fall-off than the EOS 1D Mk II. Why that should be, when they both have full frame sensors, I don't know. I'm just pleased that I chose to buy the 24-105mm f/4L because the light fall-off at 24mm on the 5D is very well controlled. Of course Nikon will have to deal with exactly the same problems when the Nikon full frame DSLR is rolled out. >>> Of course it is far worse. Film will accept light from almost any >>> angle a lens will supply to it, whereas digital sensors, even with [quoted text clipped - 9 lines] >can cope with, the fall-off is dramatic. And that's what the OP is >seeing here. This is the thing I'm not clear on. The light rays are coming in at a non-perpendicular angle. They hit a positive microlens. Wouldn't the light ray be made even less perpendicular to the sensor plane? Why aren't microlenses negative lenses since that will make the rays more perpendicular? -Rich >>>> Of course it is far worse. Film will accept light from almost any >>>> angle a lens will supply to it, whereas digital sensors, even with [quoted text clipped - 15 lines] >Why aren't microlenses negative lenses since that will make the >rays more perpendicular? The lens material has a positive refractive index which is greater than that of air. Look up Snell's Law at http://en.wikipedia.org/wiki/Snell%27s_Law Typically the refractive index of the microlens material is around 1.6-1.8, depending on the colour and dyes used (these double as the Bayer filter array). However, the issue is *not* sensitivity of the silicon sensor itself to off-axis light, in any case. Microlenses are used to increase the sensitivity of the pixel by capturing all of the light incident on the pixel area and concentrating that onto a sensitive area which is smaller than the entire pixel. Microlenses effectively increase the fill factor of the pixel to almost 100%. Depending on the thickness of the microlens and whether it is placed directly on the surface of the detector or not, light from extreme angles will be focussed off-axis in each microlens, and thus some of the light *can* miss the small sensitive area of the pixel. For example, see the light paths shown in the diagrams on this page: http://www.olympus.co.uk/consumer/dslr_7045.htm This was a problem with very early (early 1980s vintage) single chip video cameras which used coloured strip lenticular lenses on the CCD window to produce the colour sensitivity. Consequently there was a significant gap between the lenses and the light sensitive surface. These days, all of the sensors in digital cameras have the microlenses deposited on the sensor surface directly as immersion lens designs. That places the optical centre of each lens a fraction of a micron above the image plane, requiring near grazing angles of incidence to shift the focused light off the sensitive area in each pixel. You will notice that even to show the effect schematically, Olympus have to exaggerate the distance between the microlens and the sensitive surface in the diagrams above. In practice, the pixel is 5-10um in size, the sensitive area around 50% of this and the optical centre of the immersion lens almost incident on the silicon itself, making it approximately 200-400nm above the absorption depth of the light. Given the the dimensions of the Canon mount and the apertures of the lenses, you can calculate the amount of image shift that such geometries produce and thus how much light fall-off due to the microlenses actually occurs - it is trivial and totally swamped by the light fall-off from the lens itself. Consequently this is quite simply *not* a sensor issue. 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) >>>>> Of course it is far worse. Film will accept light from almost any >>>>> angle a lens will supply to it, whereas digital sensors, even with [quoted text clipped - 36 lines] >see the light paths shown in the diagrams on this page: >http://www.olympus.co.uk/consumer/dslr_7045.htm I was thinking of a pixel as a clearly defined sensor spot, not one that has greater sensitivity in a minute area. " In any case, the diagram they show wit the light be "diverging after it hits the positive lens is WRONG. The light ray would "converge" or become less perpendicular to the sensor plane, unless they've figured out how to defy physics and make a positive lens behave as if it were negative. -Rich >>For example, >>see the light paths shown in the diagrams on this page: [quoted text clipped - 4 lines] >diagram they show wit the light be "diverging after it hits the >positive lens is WRONG. Eh??? I can't see *any* such diagrams on that page. The light path in all of the diagrams is from left to right, and converges in every case when it encounters a convex interface to higher index or concave interface to lower index, just as I would expect it to. >The light ray would "converge" or become less >perpendicular to the sensor plane, All of the rays in all of the diagrams converge correctly as far as I can see, and they all converge at the focal plane. Are you sure you are not just mixing up the convergence of ray bundles, in yellow, with the natural divergence between the principle rays to each field angle, drawn in red? Which particular diagram is causing you concern? 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) >For some reason, the sensor EOS 5D appears more vulnerable to light >fall-off than the EOS 1D Mk II. The 1DmkII does not have a full-frame sensor. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >> The only cure would be a sensor smaller than full frame which sort of >> defeats the object. > >And solves the problem. As does cropping the image from the full frame camera if you find this a problem in any particular situation. 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) >>> The only cure would be a sensor smaller than full frame which sort >>> of defeats the object. [quoted text clipped - 3 lines] > As does cropping the image from the full frame camera if you find > this a problem in any particular situation. Quite so, as amongst others demonstrated by e.g. the points of attention that Gretag Macbeth add to their (EyeOne Photo SG) digicam profiling module. Not shooting the target full-frame will improve target luminance uniformity (as will stopping down), but then resolution is a smaller issue when profiling based on shooting a calibration target. Bart Canon probably guessed that, since the problems existed with film cameras since the beginning (I even have a little drop off on some transparencies taken with an Exacta and a Schneider 50mm f2 lens...made in the early 50s) that most photographers would accept this as the nature of the beast. And, so far, they've been closer to right than wrong. Look at the posts on DPReview on the subject. One guy will bring up the subject (again!) another will concur, and 50 will jump in and bellow that "it's not a design fault, it is what happens when you use a full frame camera, whether film or digital!" I, personally, have no problems with it. It is something I dealt with, as mentioned, for a long time. And any cropping at all eliminates 90-100% of it. As in taking an 8x12 image down to 8x10, or up to 11x14 or 16x20. None of these are a 3:2 ratio, and, frankly I haven't even seen issues at 12x18, which is. SignatureSkip Middleton http://www.shadowcatcherimagery.com > >> I agree. [quoted text clipped - 15 lines] > > Rita SNIP > Canon have let me down by not advising consumers of the design > defect they have with the 5D. Are you suggesting that the light fall-off is more than with full-frame 35mm film bodies using the same lens??? If you do, please provide some evidence. > As long as I can get a refund, and I've spent nearly £4,000 on new > Canon products in the last 2 months, then I'll revert to the Nikon > D-200. Sure, by all means do. Just don't complain when you run into the shortcomings of that limited camera. Bart > Are you suggesting that the light fall-off is more than with > full-frame 35mm film bodies using the same lens??? If you do, please > provide some evidence. People have been "suggesting" (and demonstrating) that for years. SignatureJeremy | jeremy@exit109.com >> Are you suggesting that the light fall-off is more than with >> full-frame 35mm film bodies using the same lens??? If you do, please >> provide some evidence. > >People have been "suggesting" (and demonstrating) that for years. Do you have a URL? I can't remember seeing a detailed comparison of Canon wide-angle glass on both film and full frame DSLRs. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency >> People have been "suggesting" (and demonstrating) that for years. > > Do you have a URL? I can't remember seeing a detailed comparison of Canon > wide-angle glass on both film and full frame DSLRs. Not immediately at hand, no. It's been posted about and talked about here roughly seven hundred thousand times, though. SignatureJeremy | jeremy@exit109.com >>> People have been "suggesting" (and demonstrating) that for years. >> [quoted text clipped - 3 lines] >Not immediately at hand, no. It's been posted about and talked about here >roughly seven hundred thousand times, though. Yes, there are plenty of examples of light fall-off with digital cameras. I don't know any examples that it doesn't happen with film. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency > Yes, there are plenty of examples of light fall-off with digital cameras. > I don't know any examples that it doesn't happen with film. It does happen with film. The contention is that it doesn't happen quite as much. SignatureJeremy | jeremy@exit109.com >> Yes, there are plenty of examples of light fall-off with digital cameras. >> I don't know any examples that it doesn't happen with film. > >It does happen with film. The contention is that it doesn't happen quite >as much. But that requires detailed measurements. If print film doesn't show light falll-off it doesn't have to say much, because print film tends to compress the dynamic range. If a slide film does not show light fall-off, then it is impressive because slide film tends to increase contrast. Some people blame the lenses, other people blame digital cameras. Only with detailed measurements is it possible to figure out how much each part contributes. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency Philip Homburg skrev: > But that requires detailed measurements. If print film doesn't show light > falll-off it doesn't have to say much, because print film tends to compress > the dynamic range. If a slide film does not show light fall-off, then it is > impressive because slide film tends to increase contrast. I feel a little insecure to the meaning af the terms used: Compress DR is as I see it what happens when a /lets say/ 12 full stop range scene is pictures as /say/ 8 full stop range image. Right? Increase contrast is, as I understand, increasing the gradient = making the curve between foot and shoulder more vertical = compressing. Right? SignatureMed venlig hilsen, Ole Larsen. New Images And Design Febr. 2006 http://home.tiscali.dk/muggler http://www.supportdenmark.com >Philip Homburg skrev: > [quoted text clipped - 7 lines] >Compress DR is as I see it what happens when a /lets say/ 12 full stop >range scene is pictures as /say/ 8 full stop range image. Right? Yes. >Increase contrast is, as I understand, increasing the gradient = making >the curve between foot and shoulder more vertical = compressing. Right? No. Contrast is increased when difference between light and dark at the output of a system is larger than on the input. So if a slide film has a gamma of 2 in a certain region, a difference of one stop on the input translates into a 4 stop difference in the resulting slide. Compared to print film, most slide films have a more or less constant gamma and a relatively sudden change to saturation. Print film is much more complicated. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency >>> Yes, there are plenty of examples of light fall-off with digital cameras. >>> I don't know any examples that it doesn't happen with film. [quoted text clipped - 5 lines] >falll-off it doesn't have to say much, because print film tends to compress >the dynamic range. Not only that, but the optics used in the printing process will (almost) always have some light fall off. In the very nature of a neg/pos printing system, the two will tend to cancel out. > If a slide film does not show light fall-off, then it is >impressive because slide film tends to increase contrast. One would see the same cancellation in prints made from slides, or in a projection system with fall-off (i.e. most of them). Direct observation would of course reveal any fall off, as would sensitometry. Scanning would probably also show it up, assuming the scanner had an even response. >Some people blame the lenses, other people blame digital cameras. Only >with detailed measurements is it possible to figure out how much each part >contributes. David SignatureDavid Littlewood >>But that requires detailed measurements. If print film doesn't show light >>falll-off it doesn't have to say much, because print film tends to compress [quoted text clipped - 3 lines] >always have some light fall off. In the very nature of a neg/pos >printing system, the two will tend to cancel out. Of course there will be some light fall off. However, if there is enough light fall off to make a difference in a print, then any frame that doesn't have significant light fall off will show brighter corners. For this reason, I doubt that high quality enlargers will have significant light fall off. Of course, these days just about everybody uses scanners to reproduce film, and I am not aware of light fall off in scanners. >> If a slide film does not show light fall-off, then it is >>impressive because slide film tends to increase contrast. [quoted text clipped - 4 lines] >would probably also show it up, assuming the scanner had an even >response. This sound like a logic error. If the slide shows light fall off and the project lens has light fall off as well, then the effects will multiply and the apparent light fall off increases. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency >>>But that requires detailed measurements. If print film doesn't show light >>>falll-off it doesn't have to say much, because print film tends to compress [quoted text clipped - 10 lines] >For this reason, I doubt that high quality enlargers will have significant >light fall off. The best enlarger lenses for 35mm typically show 1/3 to 1/6 stop corner fall-off in the corners at their optimum aperture (usually f/4 or f/5.6). Of course, people who use f/16 may not see it, but they won't get very sharp prints either. Cheaper lenses could be noticeably worse than this. >Of course, these days just about everybody uses scanners to reproduce film, >and I am not aware of light fall off in scanners. Yes, I kind of implied this. >>> If a slide film does not show light fall-off, then it is >>>impressive because slide film tends to increase contrast. [quoted text clipped - 10 lines] >as well, then the effects will multiply and the apparent light fall off >increases. You are correct; my apologies, it was 2 am! David< |
Enable EMail Alerts
Start New Thread
Reply to this Message