 |
 | 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 5D Dark Corners - Yes and No.
I have been lurking through this debate and decided to put in my two cents, and the original topic was getting too crowded with other stuff. I am am an electro-optic engineer, and have seen odd things that occur with focal planes that use microlenses. In the particular case I looked at the increase in sensitivity did not increase as quickly as one would have expected at fast F-numbers, and we found that it was due to the nature of the focal plane (with microlenses). Someone did mention this before. If anyone is interested, I will post another message showing why this occurs. In order to get the maximum sensitivity, I think ALL, or almost all, digital camera focal planes use microlenses, and I would wager that ALL full-frame 35 mm focal planes will exibit the same characteristic. It's a microlens focal plane thing, not a Canon thing. Yes, one could use a focal plan without microlenses, and I would expect that the corner fall off would be less pronounced. But the camere sensitivity would probably go down by 1 1/2 to 2 stops. The interesting test would be for someone to take a photo of the same scene with a film camera and the 5D with the same lens (wide angle, full aperture), and compare the results. (No, not me. I don't have the resources or the time.) My guess is that the corner fall off will be worse for the digital camera. Hack --//-- >I have been lurking through this debate and decided to put in my two >cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 14 lines] >corner fall off would be less pronounced. But the camere sensitivity >would probably go down by 1 1/2 to 2 stops. They are coming. http://www.digitalcamerainfo.com/content/Cypress-Announces-New-Sensor-With-No-Mi crolenses.htm And some are here now. http://www.luminous-landscape.com/reviews/cameras/expensive-ps.shtml "Phase One backs, there are no microlenses to compensate for the acute angle that the light rays take when hitting the edges of the frame, especially when the limits of the lens' imaging circle are used, with shifts beyond 10mm." No Microlenses... ***************************************************** "Vietnam is what we had instead of happy childhoods." Tim Page in "Dispatches" by Michael Herr >>I have been lurking through this debate and decided to put in my two >>cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 29 lines] > > No Microlenses... Lens Cast As with all good things, as usual there is a dark underbelly, and in the case of wide angle lenses combined with camera movements and digital backs, that problem is lens cast. Simply put, the issue is that the distance between the rear element of the lens and the back's sensor is minute. They are almost touching, unlike on an SLR camera body with a mirror housing. And in the case of the Phase One backs, there are no microlenses to compensate for the acute angle that the light rays take when hitting the edges of the frame, especially when the limits of the lens' imaging circle are used, with shifts beyond 10mm. The solution that Phase One provides is a combination of a simple diffuser plastic card and software. You shoot calibration frames with your lens and back through the filter at each major shift distance, and then using the Phase One software save these as Lens Cast calibration files. The software then will semi-automatically apply this calibration to your raw files to compensate for green or magenta casts at the edges of the frame. Once the calibration files are created – about an hour's work –their application is simple thereafter. This is not dissimilar to creating noise profile frames when using some of the better noise reduction programs. > ...In order to get the maximum sensitivity, I think ALL, or almost all, > digital camera focal planes use microlenses, and I would wager that ALL [quoted text clipped - 9 lines] > the resources or the time.) My guess is that the corner fall off will > be worse for the digital camera. It seems like a pretty simple test side by side with the same lens. I suppose different films have different contrasts so convert the raw file to match. >I have been lurking through this debate and decided to put in my two >cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 20 lines] >the resources or the time.) My guess is that the corner fall off will >be worse for the digital camera. I just 'measured' the sensitivity to angle of incidence in a quick test with the 5D - it is negligible within the constraints of the Canon lens mount. Unless film had some magical property that made it *more* sensitive to light at higher angles of incidence, then this is a non-issue. I assume that you are aware that the quoted f/# for fast lenses is rarely accurate. This isn't just lens makers being optimistic, but the use of the wrong formula. For a lambertian absorber, such as film and silicon sensors, the amount of light seen by the surface is proportional to 1/(2 sin(a)), where a is the semi-angle of the optical cone. For fast lenses, this deviates quite dramatically from the common photographic approximation of f/d. If you didn't take account of this it could explain some of your lost sensitivity with fast lenses.  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) >I have been lurking through this debate and decided to put in my two >cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 20 lines] >the resources or the time.) My guess is that the corner fall off will >be worse for the digital camera. Precisely correct, except that a sensor will microlenses will show less light fall-off than one without. >Precisely correct, except that a sensor will microlenses will show >less light fall-off than one without. In practice this doesn't appear to be significant compared to even minimal levels of lens vignetting. 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) >I have been lurking through this debate and decided to put in my two > cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 20 lines] > the resources or the time.) My guess is that the corner fall off will > be worse for the digital camera. Aren't we essentially talking about brightness changes with viewing angle on a LCD computer monitor vs. a CRT, except in reverse? The LCD uses microlenses to focus the light towards the center of the individual pixel, at the expense of the corners, ergo you get much higher brightness viewing directly as opposed to viewing from the side. Since there are no microlenses on a CRT the brightness doesn't change with the viewing angle. Seems to me the same applies to light going in either direction. Toby Toby >Aren't we essentially talking about brightness changes with viewing angle on >a LCD computer monitor vs. a CRT, except in reverse? No. >The LCD uses >microlenses Which direct view LCDs use microlenses? There are some projection and microdisplays that have them, but not any full screen size panels. They use colour filters, but that is not microlenses. > to focus the light towards the center of the individual pixel, >at the expense of the corners, ergo you get much higher brightness viewing >directly as opposed to viewing from the side. Since there are no microlenses >on a CRT the brightness doesn't change with the viewing angle. The viewing angle restrictions on an LCD occur for two main reasons: * light leakage through the crossed polarisers, reducing contrast by raising the black level and * thickness of the liquid crystal layer through which the light polarisation is rotated, making each pixel appear as a small tube through which the light passes and, as you move off axis, is attenuated. This is a completely different mechanism from the sensor. 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) >>Aren't we essentially talking about brightness changes with viewing angle >>on [quoted text clipped - 23 lines] > > This is a completely different mechanism from the sensor. OK, thanks for the clarification. However, wouldn't the real "acid test", so to speak, be to shoot a frame with your 5D and with a film camera using the same lens, and then comparing the actual amount of vignetting? In that way you use the film frame as a control--compare the relative falloff recorded on the film and on the digital frame. That seems to me to be a real-world test, as you are no longer worried at all about different angles of incidence in the light reaching, and being recorded, by the two different media. Toby > However, wouldn't the real "acid test", so >to speak, be to shoot a frame with your 5D and with a film camera using the >same lens, and then comparing the actual amount of vignetting? It would be the "acid test" *if* it were not for the following issues: non-linear film response dynamic range compression in negative emulsions dynamic range expansion in positive emulsions non-linear grain visibility across the dynamic range differences between camera shutters difference between rated film and digital speeds level of vignette present from the lens alone the small magnitude of the difference you are looking for In effect, the "acid test" is conducted every day by people who have used the same lenses on their film cameras as they now use on digital bodies and their consensus is that there is no significant difference. That hasn't stopped those who have reached full frame digital from a different route or even those who do not even have a full from digital camera claiming that their is a difference. They all bleat about some unproven theory which predicts that the problem *must* be worse on digital than film. So, in fact, the only test that matters is the one that tests the magnitude of that unproven theory. As it turns out, that test has shown a small effect in *favour* of the digital full frame sensor, although that in itself would be visually small and completely swamped by the lens vignette. 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 interesting test would be for someone to take a photo of the same > scene with a film camera and the 5D with the same lens (wide angle, > full aperture), and compare the results. I don't get why they don't just do this either. Seems like a simple enough test to me and if I was using a FF camera with non-telecentric lenses, I'd want to know the answer. SignatureStacey >I don't get why they don't just do this either. Seems like a simple enough >test to me and if I was using a FF camera with non-telecentric lenses, I'd >want to know the answer. You seem to have missed all the posts from the owners of these cameras who are using them with the same lenses they had on film cameras. They have consistently said they see no difference, that the problem of light fall off on the digital camera is pretty much the same as they saw on film. So how would one more comparison of the same type be any more convincing. That is why I tried to *measure* the only parameter that the knockers cite as the *cause* of this non-existent effect. The cause turned out to have negligible contribution and what there is is in the opposite direction of that claimed - this supports the observations that the effect is negligible too. 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) > They > have consistently said they see no difference, that the problem of light > fall off on the digital camera is pretty much the same as they saw on > film. "pretty much the same" isn't the same as "No difference". SignatureStacey >> They >> have consistently said they see no difference, that the problem of light >> fall off on the digital camera is pretty much the same as they saw on >> film. > > "pretty much the same" isn't the same as "No difference". I think that "pretty much the same" is a reasonable response, given that most of us don't have the instruments available to give an spot on measurement of it, and, at least in my case, means "no discernable difference." You may be able to measure it, but you won't be able to see it. SignatureSkip Middleton http://www.shadowcatcherimagery.com > I think that "pretty much the same" is a reasonable response, given that > most of us don't have the instruments available to give an spot on > measurement of it, Have you stood in the same place with a 5D and a film camera, used the same WA lens in back to back shots and compared the results? What's so complex about doing that to see if there is a difference rather than "pretty much the same" from recollection. SignatureStacey >> I think that "pretty much the same" is a reasonable response, given that >> most of us don't have the instruments available to give an spot on [quoted text clipped - 5 lines] > about doing that to see if there is a difference rather than "pretty much > the same" from recollection. What's so complex is finding the time to do something that, on the one hand is simple, and on the other, is so unnecessary. The lenses perform up to my expectations, and if I do a test, and show that they do, will that convince you to sell all of your Oly stuff and buy a 5D and the attendant lenses? Of course it won't. So why do it? SignatureSkip Middleton http://www.shadowcatcherimagery.com > The lenses perform up to > my expectations, and if I do a test, and show that they do, will that > convince > you to sell all of your Oly stuff and buy a 5D and the attendant lenses? ??? Why would you even care to convince people to buy what you did? I really could care less about trying to talk people into buying what I own. > Of > course it won't. So why do it? So =you'll= know if maybe you should be looking for better optics etc. Or if there is something to the digital sensor vs film thing yourself. I tested my OM lenses vs the ZD ones to see if I was missing something. I can clearly see lenses that worked great on film aren't that great on the OM dSLR's. Maybe the canon sensor isn't like this? I'd think you'd be curious, maybe you don't want to know? SignatureStacey >> So =you'll= know if maybe you should be looking for better optics >> etc. Or if there is something to the digital sensor vs film thing [quoted text clipped - 8 lines] >> >> Stacey The curious thing here is that we all "Photoshop" our images to some extent. What is the problem with using a plugin like PTLens to fix the problem. It is present with all lenses but not in all images. I use it for effect in many images but pics of a heliocopter or an areoplane where there is nothing but blue sky, show the effect more than a bridal portrait. My widest (5D) lens is a Sigma 24 -70 f/2.8 EX DG, Macro. This lens has the least vignetting of all my Canon mount lenses.Oddly enough the Canon brand lenses produce more vignetting on a 5D than many after market lenses. Even so, PTLens or Adobe Camera RAW converter can fix the problem on the fly. It costs just $5 so cost is not the reason for not using it. Signaturewww.photosbydouglas.com www.weprint2canvas.com If you really must write,use my name at an above domain. >>> So =you'll= know if maybe you should be looking for better optics >>> etc. Or if there is something to the digital sensor vs film thing [quoted text clipped - 12 lines] > extent. What is the problem with using a plugin like PTLens to fix the > problem. Because it creates it's own problems.. > This lens has the least vignetting of all my Canon mount lenses.Oddly > enough the Canon brand lenses produce more vignetting on a 5D than many > after market lenses. Even so, PTLens or Adobe Camera RAW converter can fix > the problem on the fly. But you can't replace missing information without a loss. SignatureStacey >> The lenses perform up to >> my expectations, and if I do a test, and show that they do, will that [quoted text clipped - 4 lines] > really > could care less about trying to talk people into buying what I own. The point of the two sentences below, where you interjected this unecessart aside. Why do you keep asking for "proof," then? >> Of >> course it won't. So why do it? > > So =you'll= know if maybe you should be looking for better optics etc. Or > if > there is something to the digital sensor vs film thing yourself. I don't need to prove it to myself, I already know what the results will be. I've shot enough with both film and the 5D in _similar_ conditions to have a feel for the fall off issue. > I tested my OM lenses vs the ZD ones to see if I was missing something. I > can clearly see lenses that worked great on film aren't that great on the > OM dSLR's. Maybe the canon sensor isn't like this? I'd think you'd be > curious, maybe you don't want to know? I found that several of the lenses for my older Canon bodies did not give as good results on the 20D and, later, the 5D as they did on film, particularly the 50mm f1.8 and 28-135 IS. Again, I've shot enough with my lenses with all 3 formats (film, 1.6x crop, full frame sensor) to have a good feel for it. When I thought that the older lenses weren't up to snuff, then I tested them, I do feel that my current lens lineup is up to snuff, and the light fall off issue is withing the parameters of use with film, so there's no reason for me to check them. The reason others have wanted me, and other 5D owners to check them is to prove or disprove that light fall off is a severe problem with the 5D. Most of us don't feel it is, we already have the cameras, so what's to prove? SignatureSkip Middleton http://www.shadowcatcherimagery.com > I found that several of the lenses for my older Canon bodies did not give > as good results on the 20D and, later, the 5D as they did on film, > particularly > the 50mm f1.8 and 28-135 IS. So why do people keep saying a sensor is no different than film? SignatureStacey >> I found that several of the lenses for my older Canon bodies did not give >> as good results on the 20D and, later, the 5D as they did on film, >> particularly >> the 50mm f1.8 and 28-135 IS. > > So why do people keep saying a sensor is no different than film? No one has. All anyone has said is that the light fall off on a 35mm sensor is not enough different, if at all, from that on a frame of 35mm film to get excited about. The difference in performance of some lenses on digital has been exhaustively discussed on this and many other forums, which you well know. This discussion has not expanded to those areas in which digital sensors seem to expose the limitations of some lenses designed for film, which you also well know. SignatureSkip Middleton http://www.shadowcatcherimagery.com >> I found that several of the lenses for my older Canon bodies did not give >> as good results on the 20D and, later, the 5D as they did on film, >> particularly >> the 50mm f1.8 and 28-135 IS. > >So why do people keep saying a sensor is no different than film? He didn't say it was worse vignetting. The resolution of the sensors in the 20D and 5D are better than film in many ways, so whilst aberrations may have been acceptable, or even invisible on film those limitations would have been more noticeable on the newer cameras, particularly the smaller framed camera which would have enlarged the aberrations 1.6x higher for the same sized output - even before considering taking advantage of the improved resolution of the sensor to print larger still. 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) >> They >> have consistently said they see no difference, that the problem of light >> fall off on the digital camera is pretty much the same as they saw on >> film. > >"pretty much the same" isn't the same as "No difference". In the context of a light fall-off of a stop or more, they *are* the same. 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) >I have been lurking through this debate and decided to put in my two > cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 13 lines] > use a focal plan without microlenses, and I would expect that the > corner fall off would be less pronounced. AS microlenses help capture light from angles other than straight on i find this an odd conclusion. I suppose if you upped the sensitivity at the corners in firmware and lowered the centre ones you could counter vignetting that way. I think olympus have something like that. Still despite having electronics compensate for optical deficiencies and thus make the manufacture cheaper the oly lenses seem to be extortionate especially as the covering circle they need to produce is so small. All the written explanations of the 4/3rd format makes for cheaper optics has still resulted in very expensive lenses. >>I have been lurking through this debate and decided to put in my two >> cents, and the original topic was getting too crowded with other stuff. [quoted text clipped - 20 lines] > vignetting > that way. I think olympus have something like that. Optionally and it's only for the wide lenses which have to adhere to the laws of physics. > Still despite having > electronics compensate for optical deficiencies and thus make the > manufacture cheaper the oly lenses seem to be extortionate especially as > the > covering circle they need to produce is so small. Yea 1.5mm is a HUGE difference isn't it.. SignatureStacey |
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
|
©2009 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.