Mirror Lock-Up & IS Lenses on Canon EOS-3

Discussion in 'Canon EOS' started by rishij, Feb 23, 2009.

  1. Hi,
    I have a 70-200mm f/4L series lens mounted on a Canon EOS-3 film body.
    When I turn mirror lock-up (MLU) on (by setting Custom Function 12 to '1'), Image Stabilization no longer engages on the lens (mode 1 nor 2). This is the case for any 'Drive' setting (single, multi, 10 sec, 2 sec).
    Interestingly, the EOS-3 manual says that when MLU is on, 'a tripod is required'. If the lens assumes a tripod is being used, perhaps it's turning its image stabilization off?
    Please tell me I'm doing something wrong. As I'd very much like to use MLU with the 200mm lens for shutter speeds between 1/4 - 1/125 sec when I happen to not have my tripod with me.
    Many comparisons have shown MLU to be useful at such shutter speeds handheld, so I can't fathom why image stabilization would just be turned off assuming that if you're using MLU, you're not handheld...
    Thanks in advance,
    Rishi
     
  2. Well if you lock up the mirror without using a tripod, you won't know what you're pointing the lens at or whether it's in focus. I suppose you might substitute a pile of rocks for a tripod but then you wouldn't need IS anyway...
     
  3. Karl, sure you do. You focus & compose before-hand. Focus doesn't change during the 2 seconds that the mirror would be locked up, and composition might change *slightly* due to your hand movement but I have still enough hands that this wouldn't be an issue.
    I just want IS & MLU to work together. Is that too much to ask for? I really must be missing something here; otherwise, this'd be a huge design flaw.
    Rishi
     
  4. you hit it on the head in your post - the camera thinks you are using a tripod. Canon has long said to not use IS when you are on a tripod so it is automatically turning off the IS. It is an older camera and is living by the rules that were set when it was new and IS was also relatively new...
    Stark-Arts
     
  5. Thanks Joseph.
    So the Canon 5D Mark II, for example, wouldn't behave like this? Well, I guess I'll go test it out at a store.
    Any film bodies that WON'T do this?
    Firmware updates? Should I contact Canon and ask for this as a feature? Not that they'd care on an old film body anyhow... :(
    This is extremely disappointing.
     
  6. I own a Canon EOS 3 and a 70-200 mm f4 L (non IS) lens and this is a very fine combo. Honestly I don't understand why you give so much importance to the possibility of using MLU handholding the camera... You loose the control of the viewfinder and I don't think this technique would improve the quality of your shots... In this case a good tripod + MLU would give you more freedom and outstanding results.
    Regards, Alberto.
     
  7. although I'm not familiar with the eos-3, have you tried assigning AF to a button other then the shutter? Like an AF-on button. I believe that IS is activated along with AF, maybe that can trick the camera although AF won't work since the mirror is up... Worth a try I guess, that is a design flaw, IS should never be off unless you asked it to be.
     
  8. Sam, have not tried that.
    But just to be clear: AF is working just fine with the shutter button, even with mirror lock up engaged. Perhaps people are misunderstanding: with MLU engaged, the mirror locks up AFTER you've already focused (by pressing the shutter butter down half-way) and THEN pressed the shutter button down COMPLETELY. Then you press the shutter again to take the image.
    But generally I find that method inane. So I activate the 2 second timer function. In this case, you press the shutter button down halfway to focus (at this point IS should engage!), then when you've composed and are ready to take the shot, you depress the button fully.
    The mirror locks up. 2 seconds later, the shutter is opened for exposure.
    Best way to do it... the 2 seconds allows the vibrations to drown out, and the lack of the need to press the damn shutter button again keeps out finger induced camera shake. You'd be surprised how steady your hands can maintain the camera with IS. I've shot 1/4 of a second at less zoom and had tack sharp images verified by loupe, light microscope, and 8000ppi scans.
    However, when zoomed in further, MLU becomes more critical. Perhaps it could be argued that IS performance starts degrading enough at those zooms that MLU doesn't matter. That may be valid, but I'd still like MLU with IS.
    I take it the newer digital cameras allow you to use MLU with IS? How about any of the other film bodies? The 1-series?
    Am I the only one that cares? Such is often the case, sadly... :)
    Rishi
     
  9. Late model film cameras like the EOS 3, 1V, Elan 7 series automatically disable IS during self-timer operation since a tripod is normally used. I've always disabled IS during MLU since I was using a tripod. However, nobody uses MLU without a tripod or other solid support so disabling IS makes sense.
     
  10. you don't have an understanding of how Image Stabilization works. it requires user-interaction, pointing, to maintain a point in space and works against accelerations about that point.
    daniel taylor
     
  11. you don't have an understanding of how Image Stabilization works. it requires user-interaction, pointing, to maintain a point in space and works against accelerations about that point.​
    So? What on earth does that have to do with a mirror locking up for 2 seconds before the shutter is actually opened for the exposure?
    I certainly have a rudimentary understanding of what is required for image stabilization, but beyond that, unless you are or are chummy with a Canon engineer, it's all guesswork as to the specifics .
    With all due respect, Daniel, your point is irrelevant regardless of the depth of my knowledge surrounding the technicalities of gyroscope-based or accelerometer-based image stabilization.
    I've repeatedly stated that I set my composition, point to a subject, depress the shutter button down halfway to focus on the subject, then fully depress to take the shot. At that time, I want the mirror to lock up for 2 seconds, then open the shutter, and do all this while image stabilization is engaged .
    Am I not speaking English here or something? This isn't very difficult to understand.
     
  12. Also, Daniel, it's not like the IS mechanism in the lens is doing image analysis to perform stabilization. Hence it doesn't care about what/where your focus point is or what you're pointing at. Heck you can have the thing entirely out of focus and it'll still stabilize just fine.
    Angular accelerations are detected and countered. Maybe translational motion also, to a certain degree anyway.
    This has nothing to do with mirror lock-up.
    Rishi
     
  13. as an engineer, I used to work on similar systems. you are missing the most salient point Rishi. Image Stabilization, regardless of the underlying precepts, requires that the point be established and 'continuously' updated via camera translation. the only thing IS does, is effectively damp the displacements around the aiming-point.
    with MLU active, without the viewfinder , you cannot visualize and aim at the target. having said that, I was under the impression that later cameras enabled IS in virtually all scenarios .. with the onus on the photographer to deal with it.
    daniel taylor
     
  14. Image Stabilization is not the same a gyro-stabilization.
    I've discussed this at great length here on this forum. I would be more than happy to answer any questions or delve into the technical aspects deeper.
    dt
     
  15. with MLU active, without the viewfinder , you cannot visualize and aim at the target.​
    As I said before, the IS mechanism in the lens doesn't care what I see through the viewfinder... if I can't visualize & aim the target or hold the darn thing still for an additional 2 seconds, that's my problem (thankfully, actually, it's not a problem for me).
    If you look at the handheld shots below, I believe the softness in the top shot (EV 0) was caused by vibrations originating from the mirror:
    [​IMG]
    Link to Full-Size Image
    Yes, this could also just be that 1/125 sec is too slow for handheld at 200mm zoom, but from my personal experience with this 70-200mm f/4L lens, I doubt that because I just think it's that good . Realize that's the best I can do, b/c there's no objective way to decouple softness due to mirror shake or due to inability of the IS mechanism to compensate for handheld shake (if you can design an experiment to decouple these two, please do tell).
    So, you could argue that MLU is unnecessary since, handheld, IS wouldn't be effective enough to get rid of the handheld shake at such a zoom to begin with. But:
    1. I doubt that, as I think IS is quite capable
    2. If the shake of the interior housing due to the mirror flipping up does not propagate to the lens , but does propagate to the film, IS will be ineffective in dampening it. Which is precisely why I want MLU with IS to begin with!
    Cheers,
    Rishi
     
  16. IS does nothing (good) if you indeed, hold the camera still. it requires a displacement and a correction. however, later versions of IS were sensitive enough to damp out mirror and shutter related vibrations, if on a tripod. remember, the tripod is now doing the aiming and applies the correction back on point. however, MLU is a different matter. the mirror is locked-up and any contributions have settled out. enabling IS on newer long-telephoto lenses could reduce pointing-errors due to environmental factors. and that is open for discussion as well.
    peace and photographic bliss,
    daniel taylor
     
  17. however, later versions of IS were sensitive enough to damp out mirror and shutter related vibrations, if on a tripod.​
    My 'latest version' of IS (this is 2007 lens), with this body, doesn't function on a tripod . Many versions of Canon IS lenses state 'tripod detection' (& concomitant IS shut-off) as a new feature . So why are you talking about 'later versions of IS... on a tripod' if all the newer lenses shut IS off when the camera's on a tripod?
    enabling IS on newer long-telephoto lenses could reduce pointing-errors due to environmental factors​
    I really have no idea why we're still talking about 'pointing'. Image stabilization in lenses counters angular & translational acceleration.
    My original fear/point still remains: serious design flaw , Canon.
    Rishi
     
  18. "I really have no idea why we're still talking about 'pointing'. Image stabilization in lenses counters angular & translational acceleration."
    well, if you don't 'point' or re-aim the camera there are no accelerations to correct.
    my comments are related to your initial query with the Canon EOS-3 . I haven't made a study of the newer firmware, or how Canon supports IS in newer cameras. I have two EOS-3's sitting here, a EOS-1v, and a 40D. I suppose I could do some experiments and better define the operation with IS, MLU, 2-second shutter timer, but I have other (and better) things to do today. perhaps others will have a better memory than I do regarding these matters.
    daniel taylor
     
  19. I just want IS & MLU to work together. Is that too much to ask for? I really must be missing something here; otherwise, this'd be a huge design flaw.​
    If you are hand holding, what's the point of mirror lock up anyways? I don't see how that small vibration from mirror slap is going to have much effect on a hand held shot, even with IS.
     
  20. For what it's worth, Pentax also turns off SR (shake reduction) when using the mirror lock-up (2-second self-timer), apparently with the same idea that you're usually using a tripod with this setting, so Canon's engineers aren't the only ones who think this is a good idea. It would be nice if they would include a custom setting for this behavior for those who'd rather remember to turn off SR for tripod use. Perhaps the thought is that if you're hand-holding, even with SR/IS, mirror slap/vibration is the lesser worry. My understanding however is that mirror slap has a significant effect mostly at speeds generally slower than what you're describing.
     
  21. My 'latest version' of IS (this is 2007 lens), with this body, doesn't function on a tripod .
    It's not how long ago your lens was manufactured, or even how long ago it was designed. The feature sets of the IS implementations on various IS lenses do not follow a chronological progression any longer.
    The IS superteles (300/2.8, 400/2.8, 500/4, 600/4) introduced a special tripod detection feature which aimed to correct for mirror slap if the lens detected it was on a tripod. I don't know if any of the newer IS primes do this or not.
    AFAIK, no IS zoom has this mode. Many recent IS zooms detect if they're used on a tripod, but only so that they can turn off the IS actuators and lock down the IS lens elements, in essence turning IS off (but the manuals suggest you do it manually because the IS motion sensors remain active and consume power, reducing battery life). My three current IS zooms (17-55/2.8, 24-105/4, 70-200/2.8) all behave this way.
    As for the EOS 3, at the time it was released, there was no such thing as either of these tripod detection systems in any IS lens, and Canon's recommendation was that IS was to be turned off when used on a tripod. It may be going too far to say that nobody uses MLU without a tripod, but certainly it's not far off; almost nobody does. And therefore, as of the time the EOS 3 was designed, the use of MLU pretty much necessitated that IS be off. So many bodies from around that time automatically disable IS when you engage MLU. Ditto for the self-timer.
     
  22. If you are hand holding, what's the point of mirror lock up anyways? I don't see how that small vibration from mirror slap is going to have much effect on a hand held shot, even with IS.​
    Keith, as I've already posted above, here's proof that MLU helps with hand-held shots:
    http://www.stockholmviews.com/mup/mirror-up.html
     
  23. well, if you don't 'point' or re-aim the camera there are no accelerations to correct.​
    Really? My, then whatever is the point of image stabilization?
    Of course there are accelerations to correct -- the accelerations induced by an unsteady (non-robotic? ) arm, magnified by the magnification factor of a zoom lens!
     
  24. My understanding however is that mirror slap has a significant effect mostly at speeds generally slower than what you're describing.​
    Andrew, yes, generally that's true -- shutter speeds probably slower than 1/60... but with lots of light coming through the lens, at high zooms, even faster shutter speeds can be affected.
    Remember there's a sweet spot for the benefit of MLU... above a certain shutter speed (meaning, quicker exposure), the exposure is fast enough that mirror-induced vibration will not register, just like hand-induced vibrations won't register. However, below a certain shutter speed (say, slower than 1 second), mirror-induced vibration also will not register, since not much exposure occurs, relatively, during the time over which the mirror-induced vibration dies out (i.e. the duration of mirror-induced vibrations is considerably smaller than the duration of the entire exposure).
    Hence it can occur over a variety of shutter speeds. Which is why I'd like to rule it out as a source of image softness.
    Rishi
     
  25. As for the EOS 3, at the time it was released, there was no such thing as either of these tripod detection systems in any IS lens​
    Thanks Steve. My guess is that since there was no tripod detection system in the IS lenses back then, Canon left it up to the EOS-3 body to disengage IS when it felt a tripod was being used.
    And, clearly, it thinks a tripod is using when MLU is engaged since the manual says that for MLU 'a tripod is required'.
    Bollocks.
    -Rishi
     
  26. Really? My, then whatever is the point of image stabilization?
    Of course there are accelerations to correct -- the accelerations induced by an unsteady (non-robotic? ) arm, magnified by the magnification factor of a zoom lens!
    you are just being obtuse now Rishi. your expectations and understanding are skewed if you think aiming a camera, without viewfinder or LCD, at a point (I am going to take a wild assumption that you have a subject and not a desultory point in space) ... introduces less relative-movement than the pertubations introduced by the mirror.
    daniel taylor
     
  27. introduces less relative-movement than the pertubations introduced by the mirror.​
    No... I said the opposite, that your hand-holding introduces MORE shake than the mirror would, but that IS takes care of that to a point at which mirror-induced shake may introduce larger perturbations after IS has already done its job stabilizing handheld shake (well, technically, while IS is doing its job stabilizing handheld shake).
    Who's being obtuse?
    Are you just tryina pick a fight?
     
  28. no, I am suggesting that you use the camera and lens within the limits of its design.
     
  29. no, I am suggesting that you use the camera and lens within the limits of its design.​
    With all due respect, if I wanted to do just that, I wouldn't be here would I?
    Progress doesn't result from using stuff within the 'limits of its design', especially when designers are human & don't necessarily ultimately make the best decisions and/or conclusions.
    Case in point. Stupid design. All it'd need is a firmware hack that enables a new Custom Function that allows the user to decide when and when not to turn of IS. After all, there's a damn on/off button for IS on the lens!
    Whenever processes are automated, there should be options to revert to manual. Imagine if all SLRs now were 'P' or 'Av' or 'Tv' with no 'M'. That would be a nightmare .
    -Rishi
     
  30. Well me being me thinks Rishi is being obtuse and argumentative. Your examples, well the examples you point people to, are not taken with an IS or VR lens, you will get much better vibration compensation with the IS on and no MLU than you will with no IS and with MLU hand held. Unless you don't, in which case do the latter. Accept that you can't use both, it is not a "huge design flaw" most (all apart from you?) people are very happy with the sharpness from their EOS 3 and 70-200. Your mountain shot looks out of focus to me not movement blurred, all of my IS lenses can easily correct one stop slower than focal length/shutter speed. I'd also be surprised if 1/125 second caught mirror vibration, it is generally considered to affect longer speeds than that, like the 1/40 that is the better example you link to. I have a 300 2.8 IS and a 1V, if I can be bothered at the weekend I'll see if you can use both on that.
    It just seems so obvious and simple to me.
     
  31. you will get much better vibration compensation with the IS on and no MLU than you will with no IS and with MLU hand held.​
    What about IS on & MLU handheld? Wouldn't it be nice to add that comparison the mix? The reason I don't have any examples of MLU + IS handheld is b/c I can't do that with my darn camera , & maybe no one else can? Haven't had the time to go to the local camera shop and try my lens on a 5D or such... yet.
    Your mountain shot looks out of focus to me not movement blurred​
    The shots were taken with AEB one after the other (continuous shooting mode), so focus was not adjusted between the two shots. So, even if the shots were out of focus, you can't ignore the fact that EV -1 looks sharper than EV 0, and both shutter speeds should've sufficed for handheld shooting w/ IS. Furthermore, the larger reason you probably think they look soft is probably b/c they're unsharpened Nikon LS-4000 scans (this is before I got my Minolta and/or access to an Imacon 848... so I should rescan)... but that's a moot point, b/c using a loupe on a lightbox, EV 0 is clearly less sharp than EV -1.
    Accept that you can't use both, it is not a "huge design flaw" most (all apart from you?) people are very happy with the sharpness from their EOS 3 and 70-200.​
    I've had random shots ruined every now and then that I can't chalk up to hand-held movement or focus issues (b/c a separate AEB shot was tack sharp). Perhaps said 'happy people' just haven't noticed or put two-&-two together to figure out that lack of MLU is the problem. It wouldn't surprise me if you, Scott, fell into this camp, as you adviced me earlier to 'just ignore' banding problems on the 5D Mark II and just go out and shoot and be happy :)
    And you may be right that that philosophy brings more happiness; I'm not gonna argue that. What I *am* going to argue is that if everybody just did that, there'd be no such thing as progress .
    -Rishi
     
  32. "I've had random shots ruined every now and then that I can't chalk up to hand-held movement or focus issues (b/c a separate AEB shot was tack sharp)". Rishi your camera will refocus between each AEB shot, it is just a sequence, try it with a slowly moving subject in one shot or predictive, most of the shots will be in focus, the only way to know focus wasn't changed is to turn AF off. This could well be your issue, not the inability to use IS and MLU hand held.
    You aren't arguing progress, you are just arguing, you often take a non issue and beat it to death, like your spectral (?) analisis of one pixel from an image that also wasn't yours from a camera that hadn't been released and so none of us had even touched. Surely you are not nieve enough to think anything you or I could say in these forums will change one tiniest aspect of manufacturers product plans, let alone produce progress. How many pros and prosumers etc etc have called out/begged for Canon to put a dedicated MLU button on their cameras for years, even now it only comes as a workaround for liveview. How many 1Ds MkIII's are plugged into printers to use the direct print button?
    Don't worry, with age you will become more philosophical and make more progress, they are not mutualy exclusive, I am a happy person :)
     
  33. Rishi your camera will refocus between each AEB shot, it is just a sequence,​
    Sorry, forgot to mention that for these tough focus shots, I focus, then switch to MF, then shoot the 3 shots using AEB. So, no, focus is not the issue.
    like your spectral (?) analisis of one pixel from an image that also wasn't yours from a camera that hadn't been released and so none of us had even touched.​
    I can see you're a fan of hyperbole :) 5D Mark II had already been released, & many people already had their hands on one, which is why sample images existed. Just b/c you're not a pixel-peeper, don't hate on our kind.
    You aren't arguing progress, you are just arguing, you often take a non issue and beat it to death​
    If that makes you feel better about ignoring said issues/topics, so be it. Just don't take it out on me.
    Surely you are not nieve enough to think anything you or I could say in these forums will change one tiniest aspect of manufacturers product plans, let alone produce progress.​
    Yup. CALL ME THAT NAIVE. Actually, instead of begging the manufacturers, I just change/make $hit myself. How do you think Erik developed the Scanhancer? Or how I developed a film holder than holds film completely flat within the Minolta Dimage Scan Elite scanners while not introducing Newton rings, nor anti-newton etch patterns, & not requiring any fluid-mounting bull$hit?
    Don't worry, with age you will become more philosophical and make more progress, they are not mutualy exclusive​
    Well to that I say:
    1. Thanks for implying I'm young. Always a compliment at 29 before I hit the big 3-0
    2. I disagree, with age will come more experience & knowledge that will aid my progress. Not philosophy that will aid ignorance .
    Cheers,
    Rishi
     
  34. Rishi, I had never seen a claim that MLU helps hand held. So I just tried IS with and without MLU on my 5D with 100-400L lens at manual exposure of 1/25 F5.6 ISO100. My 5D (Mk1) doesn't automatically lock out the IS when MLU is activated. I took 9 shots with and without MLU at 400mm. Without MLU 3 of the 9 looked sharp. With MLU only 2 looked sharp. The rest all have various levels of blur.
    My results with MLU were not any better than without MLU.

    With MLU on I frequently had difficulty keeping the center focus point on my target. My opinion is that on average the shots with MLU had more motion blur in them than the shots without MLU. The only thing I can say for sure is that I am somewhat surprised at how well IS works at 400mm and a shutter speed of 1/25. I have never tested this lens at such a slow shutter speed hand held.
    In my experience I have only had mirror induced vibration affect only 1 image and that was on my first very cheep tripod at 300mm. Since then I have purchased a Bogen tripod and have not had an issue with mirror vibration. Not, I have confirmed that with IS on and the camera on a tripod (with MLU), IS will cause motion blur. So Canon's recommendation not to use IS on a tripod is valid.
     
  35. Actually, I must interject & say something here.
    One of the reasons I come to photo.net is b/c I've found a problem, I pose a question re: the problem to try & find a solution; if I don't find a solution, I try and investigate further to find and/or make a solution (or not, if I have better things to do).
    Yet almost every time I do so, I get haters trying to tell me to ignore the problem. What possible purpose do said haters serve? What are you doing here? If you're not gonna provide a solution , then either bring up a valid counter-argument and present me with evidence that proves that my problem is not a problem (in this case, show me a side-by-side comparison that shows that IS with/without MLU @ 200mm zoom look exactly the same) or forever hold your peace . I showed you evidence that MLU helps in hand-held shots, and no one's yet showed me evidence that IS reduces mirror vibrations, so it's a viable possibility that my softness is coming from mirror vibrations .
    Your suggestion, Scott, that it was a focus problem, was a valid one, so thanks for that even though that was not the case. The rest of your comments though? Useless , and it appalls me that every time I come to photo.net I have to put up with such useless remarks.
    One of these days, I'll just learn to ignore them.
    -Rishi
     
  36. Steven F: THANK YOU!
    See, now there was an exemplary experiment/post/contribution.
    Glad to hear that the 5D doesn't lock out MLU. Even though it seems one wouldn't really use it w/out a tripod from your tests.
    Yes IS is amazing. I've had 1/6 of a second hand-held shots turn out tack sharp under a microscope. If mirror vibrations really aren't the problem, then I guess I'll have to chalk it up to momentary hand-held shake/jitter that IS wasn't able to entirely compensate for. More an issue with film than digital, since you can't tell until after you've developed the film :)
    I hear you on the tripod thing. So many of my shots were ruined on a cheap Velbon tripod. Hardly any shots ruined after I got my Gitzo. Sad I had to pay $700ish for such a tripod/head, but definitely worth it in the end.
    Thank you again Steven for that very valuable contribution .
    Cheers,
    Rishi
     
  37. With MLU on I frequently had difficulty keeping the center focus point on my target.​
    Steven I'm confused by this. Can you turn on MLU on the 5D such that after you depress the shutter fully, the mirror will lock up and the shutter will open automatically after 2 seconds (rather than you have to press the shutter button again)?
    Oh, perhaps you're saying that even within those 2 seconds, your target drifted off center...? That makes sense.
     
  38. Whatever, it seems dimentia has set in before the big 30 though. Your analysis was of Vincent Laforets pictures posted before the camera was released, nobody had one, when they did you came up with the banding issue and then jumped on the black spots next to highlights issue. I don't see why there shouldn't be a balanced reaction to the fairly pointless, from a photo quality point of view, pixel peeping you engage in. It does not help photography in general and even misinforms and confuses many. As a technical interest, if you have nothing better to do, then fine, but to try to push this stuff as "major flaws", hyperbole to be sure, then I would have to surmise "Are you just tryina pick a fight?" . you are.....
    Rishi, go spend a month with the HRH the Dali Lama, then tell me philosophy will aid ignorance, you are just being youthful, and arrogant.
    You are welcome, Scott.
     
  39. Rishi, go spend a month with the HRH the Dali Lama, then tell me philosophy will aid ignorance, you are just being youthful, and arrogant.​
    Since I posted my initial question, I've been called:
    • Obtuse
    • Argumentative
    • Arrogant
    • Youthful
    • Doesn't understand the basic concept of image stabilization (equivalent to: idiot)
    • Having dementia
    And you tell me *I'm* being youthful & arrogant?
    Seems like the only one(s) who need a month with some Buddhist philosophy would be those name-calling in the first place. Practice what you preach, dude.
    I, for one, actually love Buddhist philosophy, was lucky enough to see the Dalai Lama up close, visit Buddhist monasteries & monks & nuns near Tibet, and read lots of Buddhist texts in my time. If I had to pick a religion, it'd be Buddhism, so interesting you bring it up.
    But that's not the philosophy I was referring to when I spoke of philosphy aiding ignorance , but good work finding a way to take my use of the word philosphy out of context. I was speaking of your "quit worrying about technical details and go out and shoot " philosophy.
    And what of my comments about black dots on Vincent Laforet's videos/picture? Clearly Canon didn't catch this issue until the public (which I'm a part of) pointed it out; then, some of us decided to have a technical discussion about it b/c believe it or not, there are curious folk out there who like to engage in scientific discussion. Some even commented on what the ultimate cause might be and what a fix might be.
    What did all this talk/speculation do? It got Canon to provide a firmware update to fix the issue.
    So if everyone were like you, we'd still be living with the problem. If everyone were like me, the problem would probably have never existed as it wouldn't have passed by Canon QC.
    So what's your point, Scott? Ignorance is bliss? Here I got an idea for you: go write a book on it & title it 'Regress: My Vision for the Future of Progress '.
    -Rishi
     
  40. "Progress doesn't result from using stuff within the 'limits of its design', especially when designers are human & don't necessarily ultimately make the best decisions and/or conclusions."
    Thank you for sharing this piece of information, and sorry I don't have answer for your
    "question".
     
  41. "Progress doesn't result from using stuff within the 'limits of its design', especially when designers are human & don't necessarily ultimately make the best decisions and/or conclusions."
    Thank you for sharing this piece of information, and sorry I don't have answer for your
    "question".
     
  42. Rishi I could answer each and every point you raise but why bother? Stephen F answered your original specific issue with tests, it is no advantage to use MLU and IS.
    You didn't comment on the black dots until actual users did, you added nothing to the issue, you have never used a 5D MkII, your analysis prior to its release was regarding noise, not black dots, they were such a quality issue you, and everybody else, missed them for weeks! That is why I don't worry about your pixel peeping too much. If everybody was like you we would all be shooting with film and scanning with your wonderful film holder.
    Peace Brother :) (You sure can type faster than me too.........)
     
  43. That is why I don't worry about your pixel peeping too much.​
    Yet you worry about putting me down for the sake of... making yourself feel better? Hopefully by the time I'm your age I won't be so jaded after having to interact with folk such as yourself to have regressed to being as youthful & arrogant as you.
    I'm sure future thread perusers will appreciate having to sift thru your caca to eventually get to the heart of my question & that of Steven's answer. Just like they'll have to sit & read your dumba$$ comments about the Emperor & his new clothes in the Canon 5D Mark II black dots thread (where, actually, I certainly contributed to the discussion by speculating on cause & potential fix, as well as layering a couple of images for comparison of before/after firmware update, adding to the discussion of what the firmware actually did, etc.). Yup, all without an actual 5D Mark II in hand b/c, guess what, this ain't my profession & I don't have $3k to blow on a whim. If you're gonna speak, try and have a point.
    If everybody was like you we would all be shooting with film and scanning with your wonderful film holder.​
    That's really mature. Comb photo.net for the multitude of posts re: how to hold film flat. Better yet, call Julio & his fluid-mounting company ScanScience aimed at holding film flat a sham & a waste of time. Even better yet: call up Fuji & Kodak R&D and call them demented lunatics for still developing new films . My, you get more & more ignorant by the minute! Is that something I have to look forward to too as I age?
    Finally, don't wish peace when you wage war . It's extremely two-faced; haven't you learned that by now in your infinite age-ushered wisdom?
    -Rishi
     
  44. Wow such anger! I probably use more film than you do though, just got 100 rolls of Ilford 120.
     
  45. This will be my last post regarding Rishi's question.
    I said:
    With MLU on I frequently had difficulty keeping the center focus point on my target.​
    You Asked:
    Steven I'm confused by this. Can you turn on MLU on the 5D such that after you depress the shutter fully, the mirror will lock up and the shutter will open automatically after 2 seconds (rather than you have to press the shutter button again)?
    Oh, perhaps you're saying that even within those 2 seconds, your target drifted off center...? That makes sense.​
    At 400mm at a distance of about 10 feet my target was only about an inch in size. During the 2 second mirror lock up I couldn't see the target and my aim would drift by about 2 inches.
    As to how this post has been going, Rishi posted the following about the same time I posted my first responce.
    One of the reasons I come to photo.net is b/c I've found a problem, I pose a question re: the problem to try & find a solution; if I don't find a solution, I try and investigate further to find and/or make a solution (or not, if I have better things to do).
    Yet almost every time I do so, I get haters......​
    And then after you saw my responce you stated
    Since I posted my initial question, I've been called:
    • Obtuse
    • Argumentative
    • Arrogant
    • Youthful
    • Doesn't understand the basic concept of image stabilization (equivalent to: idiot)
    • Having dementia
    And you tell me *I'm* being youthful & arrogant?​
    Rishi, if this happens every time you post it should tell you something.
    It takes two to have an argument.
     
  46. It does take two to have an argument. One, a pixel-peeper , the other a pixel-peeper hater . All this doesn't tell me much more than that, since all my arguments have been with one person (Les Sarile), and now we can add Scott to the list.
    The reason I used the plural of hater is b/c time & again on these fora pixel-peepers get caca thrown at them. Not necessarily arguments, just stuff you'd expect from emotional fanboys, not critical thinkers or objective analysts.
    -Rishi
     
  47. This is a pretty absurd thread. Also, any test regarding handheld MSU shots is going to have a very very difficult time convincing me of anything. Unless you can have a robotic arm which will replicate the exact movements twice, once with MLU, and once without, there are far too many variables to take into account.
    Also, the shots above used to show the loss of sharpness due to lack of MLU with an IS lens also don't show much. Even IS lenses aren't 100% infallible in terms of motion blur. Focus could also be a problem here. I would also say that some of the additional sharpness in the second image comes from the increased contrast as well.
     
  48. Even IS lenses aren't 100% infallible in terms of motion blur.​
    Very true, and the only way to really test what's causing the blur would probably be to shoot the same image a number of times at the same shutter speed, then try this at a number of shutter speeds, then assay for sharpness (all at constant focus, of course, & preferably small aperture to weed out DOF issues).
    Then, since we know that mirror vibration is pretty consistent from shot to shot, any softness variation between shots of the same shutter speed would result from inability of IS to counter hand-induced shake. Not from mirror vibration. I guess this would weed out mirror vibration as the fault for softer handheld images when using IS.
    Anyway, after being bashed over the head for caring a # of times in this thread, I don't really think I care anymore...
    Rishi
     
  49. Rishi, I don't hate you.
    I just don't think the "issues" you stretch to put forward as "major issues" are. If that is the state of the enquiring mind, banal, pointless and argumentative "cleverness" wrapped up as helpful insight where you were studying then I fear for "progress".
    I don't post opposite you to make you feel small, or me big, I believe you think yourself to be far my superior, I do it to try to give enquiring folk happening by your threads a reality check. You don't raise real world image quality issues, I just try to point out the academic nature of your comments, people don't need to worry about buying a 5D MkII (despite your three times rumor mongering) or an EOS 3 because it can't do MLU and IS at the same time. Where is the harm, or hate, in that?
     
  50. Keith, as I've already posted above, here's proof that MLU helps with hand-held shots.​
    Rick, I'm sorry, I didn't read all of the replies before posting, so I missed that link.

    I took time to read the article. Good information. I had no idea that it made such a difference for hand held shots; looks like you have taught me something. I'll have to go home and do a test myself with my D200 (same camera they used). One thing to note, they didn't use a VR (same as IS) lens for the test. It would be interesting to see the results with MLU (or anti-shake) and a VR lens. I'll see what I can put together, allbeit with Nikon equipment.
     
  51. Unless you can have a robotic arm which will replicate the exact movements twice, once with MLU, and once without, there are far too many variables to take into account.​
    Although... if enough different people did the test, and you compared the results, you could make reasonable inferences about the effectiveness of MLU for handheld shots. So really, a robotic arm isn't necessary, you just need a larger sample than 1 test by 1 group.
    Which is why I am more than happy to do a test of my own tonight when I get home, and post the results here. Although it might be a couple of weeks before I do a test with VR, because I don't own (nor feel I need to own) any VR lenses.
    This is a pretty absurd thread.​
    It isn't near as absurd as the 1 millionth film vs digital thread that has popped up on this site. Of course I have skipped past most of the name calling....
    I am actually quite happy I stumbled upon this thread, disagreed with Rishi, and had him point me to some information that may be helpful to me in the future.
    Thanks Rishi.
     
  52. that's hardly conclusive science.
    in fact, as I read the article I chuckled out loud. just suppose now .. that the mirror-mechanics didn't contribute to the movement we see in the examples, but what actually occurred is that the photographer reacted to the sound of the mirror and flinched.
    based on my supposition, wearing ear-plugs would be just as effective as MLU perhaps!
    and if you think about it, it works in concert with the hypothesis as stated, that if we can't hear and we can't see our subject through the camera, our photographs improve inversely-proportional to the number of senses being used. a clothes-pin on your nose and you are ready to photograph a la Bresson!
    too funny .. at least to me!
    daniel taylor
     
  53. I'm just kidding Rishi .. you have made some good points vis-a-vis your vision and style of photography, and I respect that.
    daniel taylor
     
  54. Keith, thanks for the offer :) Let us know what you find. Additionally, two weeks from now, after an important deadline, I'll try IS with & without MLU on a Canon Rebel XTi and/or 5D that I should have access to, and then report back.
    Daniel, that's pretty funny. A reaction to the mirror hitting the housing when it flips up. Interesting... you know, one can't entirely discard that hypothesis! So perhaps when I do my tests, I'll wear ear-plugs :)
    -Rishi
     
  55. This is a compilation of 3 images at 100% (600W x 300H). The originals were 3904W x 2616H.
    D200, 50mm f/1.8 @ f/2.8 1/50, ISO 800.
    1. Handheld with Mirror Lockup On
    2. Handheld single shot
    3. Sharpest photo from a burst of 10 @ 5 fps (for comparison purposes)
    Please forgive me if I have not done a good job with the test. I haven't really done any type of pixel-peeping tests before.
    Note: I did this test 3 times, all with very similar results.
    00SZyC-111755684.jpg
     
  56. Just for kicks and giggles, I also wanted to compare using MLU vs Not on a Tripod.
    Same camera settings as above. Doesn't seem to make much of a difference here.... I would have expected a bigger difference on the tripod.
    00SZyH-111755784.jpg
     
  57. Thanks Keith!
    I would have expected a bigger difference on the tripod.​
    You might have a very good tripod that dampens vibrations.
    Sharpest photo from a burst of 10 @ 5 fps (for comparison purposes)​
    In continuous burst mode, does the mirror just remain up, or does it flip back down between each shot? I would suspect the former, in which case mirror vibrations do cause blur that is not just associated with human reaction to the mirror flipping up, haha ;)
    Now, we'll have to perform the same test above with/without IS.
    For your first pixel peeping test, I think you did a fine job Keith!
    -Rishi
     
  58. In continuous the mirror went up and down with each exposure. I threw in the burst shot just for comparison purposes.
    I don't think I can burst w/o the mirror coming down at the end each shot, even with a programmable remote, the mirror would come down after the first exposure.
    The tripod is a Gitzo 1227 with a Gitzo 1276 off-center head. I got a great deal on it used (if you want to call buying it and sticking it your closet for a decade used) and couldn't be happier with it. Gitzo tripods are definately worth the money, although I would say the head is a bit stiff to operate.
    I look forward to seeing your results with the 5D and Rebel XTi.
     
  59. Doh! Your mirror flips down each time? Then I think Daniel Taylor's theory gains credence -- user reaction to the sound of the mirror causing a flinch that causes shake... but a sound that you get 'used to' in burst mode... haha!
    Looks I'll definitely be wearing those earplugs during my test...
    OK, now this thread is getting a bit ludicrous... :)
    -Rishi
    P.S. I also have a Gitzo 2-series with a Manfrotto ball head (very easy to operate, you may want to look into it) which has solved many a problem.
     
  60. Another possible explanation for blur on a non-MLU shot but no blur on non-MLU with continuous (burst) mode is the following:
    If your finger, after depressing the shutter button, lifts off the shutter button prior to the exposure being finished (e.g. faster than 1/50 of a second in your case), or the muscle motions preparing you to lift your finger off the shutter button occur prior to the exposure being finished, then that could induce the shake. Since your finger stays depressed (give it some Vicodin?) in continuous mode, perhaps shots don't suffer as much from the shake save for the last one since they're not subject to any motion induced by lifting your finger off the shutter button...
    Seems a little far-fetched but, then again, millisecond human responses are not uncommon. At any rate, seems just as plausible as the 'reaction to mirror' theory Daniel has set forth :)
    -Rishi
     
  61. Actually, I don't think that is too far fetched.

    I am sure that I make many little movements due to habit... taking my finger of the shutter in single shot mode before the exposure is done is probably one of those. I will keep it in mind in the future.
     
  62. I know I am the devil of the thread, but, on another pragmatic note I learnt about motordrive/sequence/continuous slow shutter speed handheld shooting back in the 70's, this is all very well discussed and understood. Finger/hand movments have been known to affect image sharpness for the longest time. Any good technique book or teacher will point this out, sure I have seen it in more than one magazine article too. I am sure it is pointed out in some owner manuals as well, sure I just read it in my P&S manual.
     
  63. Scott,
    True, I've heard many times that in order to get a steady shot at slow shutter speeds/high zoom, one should take a burst of shots (not really practical with film, which is why I don't typically do it). I just never figured that the actual mechanism by which MLU with 2-second shutter delay gives you sharp pictures actually lies in the fact that your finger is no longer being lifted off the shutter button during the exposure.
    In fact, one might conclude then that with IS handheld, one should just use 2-second shutter delay. But even then my damn camera shuts of IS! This is what I hate about automation. Let me decide what I want and don't want to do with IS dammit!
    Anyway, the other, less desirable, alternative is to switch to single shooting mode and just keep my finger depressed well past the exposure. Which is still lame because my index finger constanly pushing on a button induces more jitter than my entire two hands holding up the system and letting the delayed shutter do the actual work.
    Anyone here with a 1V or 1N? Does IS work with the 2sec timer?
    Cheers,
    Rishi
     
  64. I've got two 1VHS's, but not with me, when I get home on Saturday I'll try. But I learn't the technique long before digital, film is cheap, just not as cheap as digital, and you don't have to reload every 36 shots :)
     
  65. Thanks Scott.
    I guess the other sensible option is to set the AEB to record exposures as EV -1 EV 0 EV +1, i.e. have the underexposure first, since that's probably when most shake occurs and is the least susceptible to blur due to the higher shutter speed. That, at least, you can set with a Custom Function.
    Rishi
     
  66. IS is also disabled on the EOS-1V with MLU. I think Rishi wishes for a steady-cam action which is not how the IS system was designed. as I said before, the idea is that the user attempts to hold a point, and the accelerations to/from this point are detected, and within a servo-loop worked against. it should be rather obvious that to blindly point the camera, is not an effective way to use Image Stabilization. in fact, do absolutely nothing, or move the camera such that the system is not accelerated, and you can watch the servo-loop drift when there are no corrections provided by the user.
    then again, it isn't up to me, or to Canon to force the user to work within the specifications. however, I would be remiss not to state my fervent feelings that this level of scrutiny and obsession is antipodal to the spirit of photography I know and feel passionately about.
    daniel taylor
     
  67. as I said before, the idea is that the user attempts to hold a point, and the accelerations to/from this point are detected, and within a servo-loop worked against.​
    Daniel, clearly I just don't get it. This method that you're describing seems to indicate some form of image analysis going on by the IS system.
    But is that the way IS works in these lenses? There's no image sensor in the lens. I thought it was all accelerometer based...?
    I may very well be wrong.
    And, fine, perhaps I could put up with IS not working with MLU. But how about at least with the 2 second timer? Seems reasonable that if I'm using IS to steady the shot, I may not want my finger depressing the shutter button to ruin the shot (unless the idea of IS is that it should also get rid of any motion induced by depressing my finger?).
    Thanks,
    Rishi
     
  68. "This method that you're describing seems to indicate some form of image analysis going on by the IS system. But is that the way IS works in these lenses? There's no image sensor in the lens. I thought it was all accelerometer based...? "
    no, it doesn't work that way. there is not image recognition. it is a very simple correction servo-loop that essentially works against your best intentions to hold a point steady. there most certainly could be some benefit here, though not nearly as effective, by holding a camera in the blind. I just packed up my lenses and should have verified the action on my 40D. I am pretty sure Canon removed this restriction at some point along the EOS evolution. most everyone here should be able to take ten seconds, try it, and report back.
    I live on a sailboat and have to start dinner, otherwise I would give it a go for you.
    daniel taylor
     
  69. Daniel I live on a sailboat and just finished dinner!
     
  70. so... then... does being a photographer = living on a sailboat?
    time for me to switch careers! :)
     
  71. Maybe that adds to my not worrying about too much Rishi :)
     
  72. no, it doesn't work that way. there is not image recognition. it is a very simple correction servo-loop that essentially works against your best intentions to hold a point steady.​
    Exactly. I feel like that's what I've been trying to say all along. That IS doesn't care about what you're pointing at.
    These systems must be designed to correct for angular rotation and translational motion about a particular point. Which point? The one that makes most sense would be the central axis of the lens. If you treat some point along that axis as the 'center of mass' of the entire imaging system, then you can measure angular rotation in 3 axes, as well as translational motion along 3 axes, about that center of mass. Then apply the inverse of these to the rear imaging element.
    Of course, that's a gross over-simplification, and I'm not even sure of the actual devices used to measure angular & translational acceleration. But the principle remains the same: for varied acceleration sources, you measure the accelerations about a certain, fixed, point. Not a point of your (the photog's) own choosing, but a point within the lens considered its 'center of mass' about which rotation and translational motion occurs. You then correct for this.
    So you can very well blindly point at something, and allow IS to do its work. It really won't care.
    -Rishi
     
  73. it wasn't designed that way, and it doesn't work that way Rishi.
    dt
     
  74. Daniel,

    You must know more about this than I do, given that you've worked on such systems? You say that:
    1. There's no image analysis going on per se
    2. Stabilization is designed to hold a particular 'point' steady
    3. Said 'point' is not some 'point' representing the center of mass of the lens, nor is it some point within the imaging system around which rotational and translational velocity is measured.
    Can all three statements above be true?
    Here's a diagram of Canon's implementation of IS, taken from the Canon website:
    [​IMG]
    Here's a diagram of the gyro sensors that detect horizontal and vertical movement:
    [​IMG]
    Gyro sensors, if I understand them correctly, measure velocity (gyro sensors that measure rotation, for example, return 'degrees per second' or angular velocity). From a number of velocity measurements you can calculate acceleration, but actually now that I think about it, velocity is probably all you need to know. If you know the position X1 at some initial time, and the velocity V , you can calculate the new position X2 after a given interval of time, s . You can then extrapolate how to move the stabilizing element such that after time s , the entering light path will behave as if the lens were at the initial position X1 , not the new position X2 . If you do this continuously in both X & Y axes, you can continuously attempt to simulate the lens being at an initial position of (X1, Y1 ).
    Oh snap ! (X1, Y1) is the 'point' you've been talking about all along, isn't it...?
    But then there must be some image or, more likely, temporal analysis that dictates to the microprocessor in real-time what should be considered to be (X1, Y1) ... correct? I just tried fooling around with the 70-200 f/4L IS, and it seems that, when I constantly move the lens around then allow it to rest, the lens tries to snap the image back to a point around where I allowed the lens to 'rest' (well... before I started panning again). So perhaps it measures a nadir in a velocity vs. time map, sets that point as (X1, Y1) , and then attempts to stabilize about this point?
    -Rishi
     

Share This Page