500F4IS: IS Mode 2 doesn't turn off if on tripod

[Joerg Rockenberger]

Hi,

it is my understanding that later generation IS systems on modern lenses like the Canon 500F4IS are expected to turn themselves off in Mode 2 if they are mounted on a tripod. For that reason I leave the IS mode set to 2 regardless if I use a tripod or not.

Now, there is no doubt that upon half-pressing the shutter button on my 500F4IS the IS motors activate (I can hear them) and the image in the viewfinder stabilizes even if the lens is mounted on the tripod and in mode 2. I've always wondered about the apparent contradiction to what's stated in the lens manual and the recommendations here and on other gear forums but didn't think too much of it.

But today, I tried shooting at very low shutter speed and decided to see if I can get better sharpness using a remote shutter release. The 500F4IS was in mode 2, mounted on a Gitzo, the Mongoose was locked down, no wind, single focus shot, central focus point, I was not touching the lens or camera and upon half-pressing the remote shutter release the IS activated. Not only that but the image in the viewfinder moved considerably (usually to the left) and then slowly reapproached the original position. Total time it took to restabilize was 1 - 2 sec. Frankly, I was and am stunned.

For a given focus target, the IS creep and back was very reproducible. On a slightly different target it tended to reproducibly go left down (think 7 o'clock) and went back up.

What can cause this? Has anyone else experienced this? And how did you resolve it?

Thanks in advance. JR

[Jay Gould]

Hi John, thanks for raising this question; I too based upon Jim Neiger's recommendations have been leaving my 300 set to #2.

[John Chardine]

There is a bit of confusion here. Mode 2 IS is for panning. In mode 2, the IS first detects the direction of panning, then operates the IS in a direction perpendicular to the panning direction. The IS shutoff function is independent of the IS mode and kicks-in when the IS system detects no motion, i.e., when the system is locked down tight. As I have mentioned before in other threads, with a super-tele such as a 500 attached, elimination of all micro-movements on a tripod is almost impossible unless you balance the gear on the head, lock the head down, set mirror lockup, use self timer, use a cable release, and there is no wind or vibration through the floor. If any of these occur, the IS continues to work- and so it should- to reduce camera movement even when the gear is mounted on a tripod. The tell-tail sign that the IS is "fighting" a rock-solid mounting is that you will see the image in the view-finder jump.

[Joerg Rockenberger]

Thanks John for taking the time to comment. But aren't the two following statements somewhat contradictory?

The IS shutoff function is independent of the IS mode and kicks-in when the IS system detects no motion, i.e., when the system is locked down tight.

vs

The tell-tail sign that the IS is "fighting" a rock-solid mounting is that you will see the image in the view-finder jump.

I definitely observed the latter (and the IS motors were working) but based on your first statement and statements by others the IS should turn itself off. No?

Also, the rig was very stable as there was no movement in the viewfinder as long as I didn't touch the camera/lens and turned the IS off completely. That also took the sharpest image.:o

Thanks, JR

[John Chardine]

Could be. Didn't mean it to be. What I should have gone on to say is that I have actually never witnessed the IS completely turn off on my 500/4, maybe because I never lock the gear down sufficiently for this to happen. When I approach perfect lock-down I have sometimes seen the IS make the image jump, which I assume happens when there is no motion at all. Then when it picks up a little micro-movement it stops the positive-feedback loop that causes the jumping and starts to work properly. I've also seen it jump when panning in IS mode 1 (forgot to change to 2). Bottom line here I think is despite the way Canon describes it, its IS system on the super-teles does not work in a black and white way, i.e., mount on (any) tripod, IS switches off, take off tripod, IS turns on.

[William Malacarne]

First let me say I have not tried this myself, only repeating what I have heard. As an experiment if you have live view on your camera. When using a long telephoto lens turn the IS off and button everything down with live view on and on 10 times zoom. Watch and see if you have any movement.

Bill

[John Chardine]

First let me say I have not tried this myself, only repeating what I have heard. As an experiment if you have live view on your camera. When using a long telephoto lens turn the IS off and button everything down with live view on and on 10 times zoom. Watch and see if you have any movement.

Bill

Bill- Results will depend on a few variables such as ground vibration and breeze. If there is neither, and you let the gear settle down from the last time you touched it (say 10s), then you will likely get a still image. However, the effect of mirror slap and shutter vibration will ruin all this under normal shooting conditions.

[William Malacarne]

Yes I agree, but he may be getting movement that he is not aware of and with a big lens it takes very, very little movement.

Bill

[Joerg Rockenberger]

Bill,

I agree that there could be vibrations which are not visible in the viewfinder but which I would pick up in liveview zoom. If that's why the IS continues to operate fine. But the more troublesome issue for me is this jump/creep in the viewfinder which as I understand the latest IS generations are supposed to avoid as they should sense that they are on a tripod.

Can someone try to reproduce my experiment and report here?

Thanks, JR

[John Chardine]

John- I just tried a little experiment. 50D mounted on 500, all mounted on Mongoose 3.5 head and Manfrotto 055 tripod base. Tripod inside house on solid floor and head locked down, cable release installed. Live View on and set to highest magnification on the LCD, IS on, mode 1, AF set to off. I then waited for all visible vibration on the LCD to calm down to zero (ca. 10s) before pressing cable release half-way to activate IS. Result is that with this set up, image initially drifted (not jumped) up and to the left and then settled down to hovering back and forth about this new point. When cable release let go, image drifted back to starting point. Results were the same if AF was on and regardless of the IS mode. IS gyros were audible throughout test.

My conclusion is that IS on my 500 is not completely disabled when the camera/lens is motionless. However, although working, it's possible that the IS was not functioning as it would if the camera/lens were shaking, and that the small hovering motion that I saw would not affect IQ. I'd be interested to get Canon's "take" on this.

[Joerg Rockenberger]

Thanks John for doing this test. So, there are at least two systems out there where the image drifts after activating the IS despite the lenses being mounted on tripods under very controlled conditions. :) The behavior in my case was a little bit different than what you describe but I am sure it is related. And it definitely would (and did) affect the IQ in my case as the drift/creep was on the order of 1-2X of the size of the square of the central AF sensor in the viewfinder.

And yes, there seems to be a disconnect between Canon's marketing material and our observations. Anyone here with a good connection to some Canon folks? Just curious... :)

Best, JR

[Charles Glatzer]

I have never had the IS in either mode 1 or 2 "turn off" by itself when using any lens on a tripod regardless of how stable it was on a tripod. If you are worried by this...I would shoot similar images with the IS TURNED ON AND OFF MANUALLY and compare the results. And. please let us know your findings.

Chas

[Joerg Rockenberger]

Chas,

Thanks for your input. In the experiment I described above the sharpest image was obtained by turning IS completely off. However, these were somewhat special conditions as I was trying to shoot at very low shutter speeds.

I am doing more field shots under more common conditions at the moment to see if it makes a difference.

JR

[John Chardine]

I finally had a chance to bring up this topic with Chuck Westfall- Canon USA technical guru. He was kind enough to send me some Q&As on this topic- answers of course written by Chuck himself. Rather than edit I thought pretty well everything here is relevant to the topic so here it is (by the way just emailed him out of the blue a couple of hours ago so how's that for customer service!:

Note in particular Q8/A8
---------------------------------------------
Q1. The instruction sheets with my IS lenses say to turn the IS function OFF when using a tripod. I have heard other photographers argue the opposite. Would you discuss the pros/cons of this?
A1. The answer depends on which type of IS lens you're talking about. With the IS super-telephotos (200/2L IS, 300/2.8L IS, 400/2.8L IS, 500/4L IS, 600/4L IS and 800/5.6L IS) as well as the 400/4 DO IS, it's OK to leave IS turned on when the lens is mounted on a tripod. For older-vintage IS lenses introduced before 1999 (28-135/3.5-5.6 IS, 75-300/4-5.6 IS, 100-400/4.5-5.6L IS, and 300/4L IS), it is much better to shut off the IS function when the lens is mounted on a tripod. For other IS lenses, see A9.
Here's the explanation for the 28-135 IS, 75-300 IS, 100-400L IS, and 300/4L IS lenses:

Q2. The instruction manual warns that the IS stays on after pressing the shutter button half way, and that use of IS on a tripod may damage the lens.
A2. Here's what the instruction book says (page E-8): "Do not set the image stabilizer switch to 'I' when using the camera on a tripod. Doing so may cause the image stabilizer to act erratically. Turn the image stabilizer off '0' before using the camera on a tripod."
There is no language here or anywhere else in the instructions concerning damage to the 28-135 IS caused by using the camera on a tripod. All they're saying is that using IS while the camera is on a tripod "may cause the image stabilizer to act erratically."

Q3. Why does tripod use cause a problem with the older IS lenses?
A3. Short explanation: The IS mechanism operates by correcting shake. With the older IS lenses (the 28-135 IS, 75-300 IS, 100-400L IS, and 300/4L IS), when there is no shake, or when the level of shake is below the threshold of the system's detection capability, use of the IS feature may actually *add* unwanted blur to the photograph, therefore you should shut it off in this situation.
Longer explanation: Remember that the IS lens group is normally locked into place. When the IS function is active, the IS lens group is unlocked so it can be moved by the electromagnetic coil surrounding the elements. With the older IS lenses (the 28-135 IS, 75-300 IS, 100-400L IS, and 300/4L IS), when there's not enough motion for the IS system to detect, the result can sometimes be a sort of electronic "feedback loop," somewhat analogous to the ringing noise of an audio feedback loop we're all familiar with. As a result, the IS lens group might move while the lens is on a tripod, unless the IS function is switched off and the IS lens group is locked into place.

Q4. Is IS best kept off, then turned on for each shot as needed, then turned off again, much as one would use the exposure compensation dial?
A4. For most hand-held work, leave the switch on; for tripod work, shut it off, as noted in the instruction book.
The recommendations above do *not* apply for the 200/2L IS, 300/2.8L IS, 400/2.8L IS, 400/4 DO IS, 500/4L IS, 600/4L IS and 800/5.6L IS. Here is some information for them:

Q5. [with 200/2L IS, 300/2.8L IS, 400/2.8L IS, 400/4 DO IS, 500/4L IS , 600/4L IS and 800/5.6L IS] Does IS operate when the lens is on a tripod or not??
A5. I can positively confirm that it does, and quite well as a matter of proven fact.

Q6. The Owner's Manual (page 12) says it does not.
A6. I know. But testing indicates otherwise.

Q7. Canon Inc.'s EF Lens Tech Report (Vol. 13) implies that it disengages but then says it stabilizes when using a tripod.
A7. I see no such implication. My copy of this report says: "With previous IS lenses, it was necessary to turn off the Image Stabilizer when using a tripod because it might operate incorrectly. But the Image Stabilizer installed on these IS lenses has an additional function that prevents improper operation when using a tripod. It also corrects camera vibration caused by slow shutter photography when using a tripod: a common problem with Super Telephoto lenses. About one second after the shutter is pressed halfway, the unit automatically determines that a tripod is being used from the output of a vibration gyro." These statements are 100% accurate based on my testing. There are no other references to the use of IS on a tripod anywhere in this report, as far as I can tell.

Q8. My own guess is this. When the tripod/head is rigid and there is little or no vibration, IS does disengage. When you loosen the head -- as for following birds in flight -- the gyros sense some vibration and the IS begins to operate again.
A8. This is not quite right. Based on my own personal testing and observation, as well as a personal consultation with one of Canon Inc.'s chief executives in the Lens Division, these lenses effectively have 2 IS modes *within Mode 1.* The decision on which mode will be used is made by the equipment, not the photographer, based on an evaluation of the degree of detected movement, as follows:
When the lens is mounted on a steady tripod and the shutter button is pressed halfway, the IS begins to operate immediately and the image in the viewfinder goes through a very slow vertical shift for about 1 second. After that, if the shutter button continues to be pressed halfway, the IS mechanism automatically goes into a special mode which is designed to detect and correct for mirror slap and shutter movement at slow shutter speeds. Based on my personal testing, it does this job spectacularly well. Close examination of results from shooting tests definitively proves that there is no advantage to be gained by shutting off IS, or even by locking the mirror prior to exposure.
When the lens is mounted on a monopod or hand-held *and* the gyro sensors detect movement exceeding a certain rate, the IS mechanism switches into its conventional correction mode which is designed to compensate for relatively gross and/or erratic lens movement.
So, what you're seeing when you loosen the ball head and start moving the lens is essentially an automatic switchover from "tripod IS" to "conventional IS."

Q9. With regards to the EF70-200mm f/4L IS USM lens, notwithstanding the battery power considerations mentioned in the manual, does this lens have the ability to detect when it's on a tripod?
A9. The EF70-200mm f/4L IS USM (along with most other IS lenses released after 1999) has the ability to detect when it is mounted on a tripod. If you keep the shutter button half-pressed, or better yet, use a remote switch to simulate a half-press, the stabilizer mechanism will drift the image downwards for the first second or so, then the mechanism will stop moving. (It's for this reason as well as battery power conservation issues that the instruction booklet recommends shutting off the IS system while the lens is mounted on a tripod.) You can see the effect if you look through the viewfinder while half-pressing the shutter button. However, it is important to understand that this form of disabling is different than shutting off the IS function with the mode switch on the lens. In the latter case, the IS mechanism is centered and locked into place, whereas in the former case, the IS mechanism shifts the image downward slightly for a second or so, then stops moving. It's not moving, but it's not centered or locked, either. It's effectively on standby, so that it can resume its corrective capabilities instantly if movement is detected. Again, you can see this for yourself by looking through the viewfinder while pressing the shutter button halfway for at least several seconds, assuming the lens is mounted on a tripod and the IS switch is on. To my way of thinking, this is not the optimum way to use the equipment. In my opinion, if you use the EF70-200mm f/4L IS USM on a tripod, it would be best to turn off the IS mechanism via the switch on the lens, rather than depending on the tripod detection capabilities of the IS mechanism.
Here is a list of IS lenses released as of August 10, 2009 that are similar to the EF70-200mm f/4L IS USM in terms of their behavior when used on a tripod:
EF-S 17-55mm f/2.8 IS USM
EF-S 17-85mm f/4-5.6 IS USM
EF-S 18-55mm f/3.5-5.6 IS
EF-S 18-200mm f/3.5-5.6 IS
EF 24-105mm f/4L IS USM
EF 28-300mm f/3.5-5.6L IS USM
EF-S 55-250mm f/4-5.6 IS
EF 70-200mm f/4L IS USM
EF 70-200mm f/2.8L IS USM
EF 70-300mm f/4-5.6 IS USM
EF 70-300mm f/4.5-5.6 DO IS USM

Q10. [In terms of Canon lenses,] should I use IS Mode 2 all the time? If not, why?
A10. The answer to these questions largely depends on whether you're shooting hand held or from a tripod. For hand-held photography, I recommend Mode 1 for stationary subjects and Mode 2 for panning, with any IS lens that has both modes. When using a tripod, the first thing you need to think about is whether to use IS at all. Unless you are using one of our IS super-telephoto lenses (i.e., 200/2L IS, 300/2.8L IS, 400/2.8L IS, 400/4 DO IS, 500/4L IS, 600/4L IS, 800/5.6L IS), I recommend shutting off the IS for best results. If you are using one of the aforementioned IS super-telephoto lenses on a tripod, the choice between Mode 1 and Mode 2 is largely a matter of personal taste. If there is any possibility of panning, Mode 2 is best. If not, it doesn't make any difference.
(For the benefit of readers who might be unfamiliar with Canon’s nomenclature, Mode 1 on an IS lens attempts to compensate for vertical and horizontal camera movement simultaneously. Mode 2 shuts off IS in the panning direction when panning is detected.)

Q11. Could you please clarify the operation of Mode 2 (panning) IS on the Canon EF70-200mm f/2.8L IS USM lens? Does it work correctly only when panning across the long axis of the frame (horizontally) or is it smart enough to detect a smooth pan in any orientation/direction and correct for vibrations perpendicular to that detected axis (even diagonally, for instance)?
A11. Every Canon Image Stabilizer lens has two "gyro sensors" oriented at right angles to each other. If the IS lens has a Mode 2 setting, it compensates for panning by shutting off image stabilization in the direction that is parallel to the panning movement. With this system, it's entirely possible to compensate effectively for panning in any direction, even diagonally. It also makes no difference whether the camera is held horizontally, vertically, or diagonally with respect to the horizon line of the image.

[Joerg Rockenberger]

Wow, John! Thanks for reaching out to Chuck Westfall. Lots of info. But there are a few things which still don't quite make sense to me and contradict my own observation.

Q7. Canon Inc.'s EF Lens Tech Report (Vol. 13) implies that it disengages but then says it stabilizes when using a tripod.
A7. I see no such implication. My copy of this report says: "With previous IS lenses, it was necessary to turn off the Image Stabilizer when using a tripod because it might operate incorrectly. But the Image Stabilizer installed on these IS lenses has an additional function that prevents improper operation when using a tripod. It also corrects camera vibration caused by slow shutter photography when using a tripod: a common problem with Super Telephoto lenses. About one second after the shutter is pressed halfway, the unit automatically determines that a tripod is being used from the output of a vibration gyro." These statements are 100% accurate based on my testing. There are no other references to the use of IS on a tripod anywhere in this report, as far as I can tell.

Does that mean one should wait for a few seconds before taking a shot if mounted on a tripod and used in very steady conditions (e.g. remote cable release)? On the other hand, the vibrations induced by holding the camera/lens are probably enough to engage the IS.

Q8. My own guess is this. When the tripod/head is rigid and there is little or no vibration, IS does disengage. When you loosen the head -- as for following birds in flight -- the gyros sense some vibration and the IS begins to operate again.
A8. This is not quite right. Based on my own personal testing and observation, as well as a personal consultation with one of Canon Inc.'s chief executives in the Lens Division, these lenses effectively have 2 IS modes *within Mode 1.* The decision on which mode will be used is made by the equipment, not the photographer, based on an evaluation of the degree of detected movement, as follows:
When the lens is mounted on a steady tripod and the shutter button is pressed halfway, the IS begins to operate immediately and the image in the viewfinder goes through a very slow vertical shift for about 1 second. After that, if the shutter button continues to be pressed halfway, the IS mechanism automatically goes into a special mode which is designed to detect and correct for mirror slap and shutter movement at slow shutter speeds. Based on my personal testing, it does this job spectacularly well. Close examination of results from shooting tests definitively proves that there is no advantage to be gained by shutting off IS, or even by locking the mirror prior to exposure.
When the lens is mounted on a monopod or hand-held *and* the gyro sensors detect movement exceeding a certain rate, the IS mechanism switches into its conventional correction mode which is designed to compensate for relatively gross and/or erratic lens movement.
So, what you're seeing when you loosen the ball head and start moving the lens is essentially an automatic switchover from "tripod IS" to "conventional IS."

Okay, so what I observed is expected - a drift after half-pressing the shutter on the remote release cable. Although in my case it was a shift and then back to the original position. And it was very reproducible but depended on what it was focused on and certainly not straight vertical.
How Chuck can say that it's not better to turn IS off in such situations is a bit surprising to me. Imagine you monitor your subject through the viewfinder and you halfpress the shutter (on the remote release cable) when you see the perfect head angle - and then the image drifts...

Q9. With regards to the EF70-200mm f/4L IS USM lens, notwithstanding the battery power considerations mentioned in the manual, does this lens have the ability to detect when it's on a tripod?
A9. The EF70-200mm f/4L IS USM (along with most other IS lenses released after 1999) has the ability to detect when it is mounted on a tripod. If you keep the shutter button half-pressed, or better yet, use a remote switch to simulate a half-press, the stabilizer mechanism will drift the image downwards for the first second or so, then the mechanism will stop moving. (It's for this reason as well as battery power conservation issues that the instruction booklet recommends shutting off the IS system while the lens is mounted on a tripod.) You can see the effect if you look through the viewfinder while half-pressing the shutter button. However, it is important to understand that this form of disabling is different than shutting off the IS function with the mode switch on the lens. In the latter case, the IS mechanism is centered and locked into place, whereas in the former case, the IS mechanism shifts the image downward slightly for a second or so, then stops moving. It's not moving, but it's not centered or locked, either. It's effectively on standby, so that it can resume its corrective capabilities instantly if movement is detected. Again, you can see this for yourself by looking through the viewfinder while pressing the shutter button halfway for at least several seconds, assuming the lens is mounted on a tripod and the IS switch is on. To my way of thinking, this is not the optimum way to use the equipment. In my opinion, if you use the EF70-200mm f/4L IS USM on a tripod, it would be best to turn off the IS mechanism via the switch on the lens, rather than depending on the tripod detection capabilities of the IS mechanism.
Here is a list of IS lenses released as of August 10, 2009 that are similar to the EF70-200mm f/4L IS USM in terms of their behavior when used on a tripod:
EF-S 17-55mm f/2.8 IS USM
EF-S 17-85mm f/4-5.6 IS USM
EF-S 18-55mm f/3.5-5.6 IS
EF-S 18-200mm f/3.5-5.6 IS
EF 24-105mm f/4L IS USM
EF 28-300mm f/3.5-5.6L IS USM
EF-S 55-250mm f/4-5.6 IS
EF 70-200mm f/4L IS USM
EF 70-200mm f/2.8L IS USM
EF 70-300mm f/4-5.6 IS USM
EF 70-300mm f/4.5-5.6 DO IS USM

Okay, now the drift is downwards. I guess we can assume that it is random. And in my case it didn't stop moving but went back to the original position.
But more importantly, is the recommendation to turn IS off for the 70-200F4IS not in direct contradiction to A8? Or are the zoom lenses with IS in a yet another category from the super teles primes and the pre 1999 models?

This definitely has been helpful. But it's still a bit confusing.

Thanks again. Also to Chuck for providing this compilation. JR