Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: Guillermo Luijk on February 06, 2016, 05:17:18 am
-
Thousands of times I hear: "WHEN USING A TRIPOD SET THE STABILIZER ALWAYS OFF". An image is worth 1000 words:
(http://www.guillermoluijk.com/article/tripodandis/viento_ibisoff.jpg)
(http://www.guillermoluijk.com/article/tripodandis/viento_ibison.jpg)
(http://www.guillermoluijk.com/article/tripodandis/bodegon.jpg)
Shots done with an E-P5, 135mm, 1/1,6 s @ f/5,6, ISO200, rear wind over the tripod+camera combo was simulated with a fan.
Intuitively I always thought the IS could improve shots on a tripod when the conditions are not stable (e.g. wind or ground vibrations). When shooting on a stable environment (indoors, outdoors in absecene of wind), there is no practical difference between using the IS or not, so you just don't need to care. But when the conditions are the opposite the IS helps to improve image sharpness.
So the "WHEN USING A TRIPOD SET THE STABILIZER ALWAYS OFF" myth is false, I'd rather say "WHEN USING A TRIPOD SET THE STABILIZER ALWAYS ON", and by extension in practice: "LEAVE THE STABILIZER ALWAYS ON".
Regards
-
Depends on the stabilization software, some old implementations blur images when the camera is perfectly still while more modern versions can recognize when the camera doesn't need it and turn it off automatically.
Then again, 135mm on a m43 body is a very long lens and more susceptible to tiny vibrations from pretty much anything, so leaving IS on might benefit this camera/lens combination a lot more than having a D810 with a 14mm lens on it.
-
Depends on the stabilization software, some old implementations blur images when the camera is perfectly still while more modern versions can recognize when the camera doesn't need it and turn it off automatically.
I have heard that many times but I don't think the IS is turned off in the case of my E-P5, since I can hear it working during the whole exposure. It simply works better than older versions not introducing any noticeable blur.
I am having reports from people obtaining worse results when the IS is ON in long exposure with old gear like the Olympus E-510 or the Tamron SP 17-50 f2,8 for DSLR cameras. But with more recent IS lenses and IBIS bodies, the IS seems to have zero effect when the tripod is perfectly stable. So the old myth doesn't seem to hold true anymore.
Regards
-
Very early after acquiring my Canon 100-400 (first version) I noticed that leaving IS turned on while shooting with a tripod gave better results. The first times I used the lens I followed Canon's instructions to turn IS off and was disappointed with a number of images. Then I tried with IS turned on and have kept it that way ever since.
I wouldn't generalise and suggest to do so on every lens with every camera, but it's well worth investigating, should some tripod-mounted shots show camera shake.
-
Thank you for sharing this valuable info!
-
The tripod is obviously moving as it is too flimsy. IS is doing what it is supposed to do and stop the movement. So it is not really a test. Try it indoors and on a decent tripod.
Whether you have IS on or off is explained in the manual for the lens. Some you can and some you can't.
Leaving IS always on just drains the battery and produces an annoying noise.
-
The main purpose of this test was not to make true the "STABILIZER ALWAYS ON" mantra but prove false the "ON A TRIPOD ALWAYS SWITCH IT OFF" claimed by so many manufacturers and users.
When I know for sure that the IS will add nothing to image quality I switch it off. I also know that if I forget to do it the result will be the same in practical terms.
Regards
Enviado desde mi GT-I9195 mediante Tapatalk
-
Hi Guillermo,
Thanks for sharing your findings. Things coming from you used to be good food for thougt.
I think it varies from system to system. With my older Sony cameras it used to be an issue.
Best regards
Erik
The main purpose of this test was not to make true the "STABILIZER ALWAYS ON" mantra but prove false the "ON A TRIPOD ALWAYS SWITCH IT OFF" claimed by so many manufacturers and users.
When I know for sure that the IS will add nothing to image quality I switch it off. I also know that if I forget to do it the result will be the same in practical terms.
Regards
Enviado desde mi GT-I9195 mediante Tapatalk
-
I once had a night shoot completely ruined when I forgot to turn Canon's IS off.
-
Depends on the stabilization software, some old implementations blur images when the camera is perfectly still while more modern versions can recognize when the camera doesn't need it and turn it off automatically.
Then again, 135mm on a m43 body is a very long lens and more susceptible to tiny vibrations from pretty much anything, so leaving IS on might benefit this camera/lens combination a lot more than having a D810 with a 14mm lens on it.
This is the truth. It really depends on the implementation in the lens and the actual stability of the platform itself.
It has been demonstrated many times, for example, with older Nikon VR that shutter speeds over 1/500th that the VR would degrade the image. Also, if you have a tremendous amount of weigth cantilevered in high wind, how stable is the platform?
Bottom line as in almost all things is testing with the equipment you shoot in the conditions you shoot. It is usually beneficial to experiment with new lenses and body combinations to see what works best.
-
I ran a quick tripod test with a 5DS R, 70-200mm IS II at 200 and F11 on 3 different exposure times. 1/500th, 1/1000, and 1/2000. Image stabilizer made no difference at 1/2000 or 1/1000. It reduced motion blur from mirror shake about 30% at 1/500th. At lower speeds it helps even more.
I was looking for a transition region where the IS amplified motion blur but didn't see it with this combination.
So for tripod use with a Canon 70-200 IS II it seems IS On is the way to go.
There is an issue with extremely long exposure times. the IS will gradually drift. I'm not sure where, exactly, it starts to deteriorate on a tripod but guess it is on exposures longer than 1/10th.
-
Thousands of times I hear: "WHEN USING A TRIPOD SET THE STABILIZER ALWAYS OFF". An image is worth 1000 words:
Intuitively I always thought the IS could improve shots on a tripod when the conditions are not stable (e.g. wind or ground vibrations). When shooting on a stable environment (indoors, outdoors in absence of wind), there is no practical difference between using the IS or not, so you just don't need to care. But when the conditions are the opposite the IS helps to improve image sharpness.
So the "WHEN USING A TRIPOD SET THE STABILIZER ALWAYS OFF" myth is false, I'd rather say "WHEN USING A TRIPOD SET THE STABILIZER ALWAYS ON", and by extension in practice: "LEAVE THE STABILIZER ALWAYS ON".
Regards
I tested the a7II with a 180 and a big RRS tripod, and came to a different conclusion:
http://blog.kasson.com/?p=8391
The situation is a bit more muddled with the a7II with a light tripod, and even more so with the a7R, a camera that interacts with OSS in some strange ways:
http://blog.kasson.com/?p=9743
Jim
-
Again, until you can come up with a sharp image on the tripod with IS off then the test was obviously flawed.
The tripod is poor quality tripod or nor weighted down or the head is dropping. Is the mirror locked up? If not then the camera will shake. Are you using a cable release? If not then again the camera will shake.
Bottom line. THE CAMERA IS MOVING, so of course IS will improve it.
-
Again, until you can come up with a sharp image on the tripod with IS off then the test was obviously flawed.
The tripod is poor quality tripod or nor weighted down or the head is dropping. Is the mirror locked up? If not then the camera will shake. Are you using a cable release? If not then again the camera will shake.
Bottom line. THE CAMERA IS MOVING, so of course IS will improve it.
I understand what you are saying and agree. But, It is a given that if the camera is totally steady, VR should have nothing but a negative effect. However, the real world discussion is really VR on or off in Situation X under Conditions Y. If the camera isn't totally steady in this scenario then VR likely will help and it should be turned on. Again, a little testing should always be done.
-
We agree but the exposure here if I am reading it correctly is only 1.6 secs. You should be able to take exposures of minutes without visible movement or something is wrong beyond needing IS.
-
Again, until you can come up with a sharp image on the tripod with IS off then the test was obviously flawed.
The tripod is poor quality tripod or nor weighted down or the head is dropping. Is the mirror locked up? If not then the camera will shake. Are you using a cable release? If not then again the camera will shake.
Bottom line. THE CAMERA IS MOVING, so of course IS will improve it.
Again, the purpose of the test was to prove that the "on a tripod always switch OFF the stabilizer" is a false myth. The IS can help improve image quality in some situations.
The E-P5 has no mirror so it's impossible to lock it up, I used remote shutter and Anti-shock mode to minimise shutter vibration since the camera has no electronic first curtain. The speed was 1/1,6s not 1,6s. The focal length, travel tripod and fan were chosen to make more clearly visible any loss of sharpness. It was a desired situation in order to compare IS vs no IS better.
(http://www.guillermoluijk.com/article/tripodandis/simuladorviento.jpg)
Regards
-
I ran a quick tripod test with a 5DS R, 70-200mm IS II at 200 and F11 on 3 different exposure times. 1/500th, 1/1000, and 1/2000. Image stabilizer made no difference at 1/2000 or 1/1000. It reduced motion blur from mirror shake about 30% at 1/500th. At lower speeds it helps even more.
I was looking for a transition region where the IS amplified motion blur but didn't see it with this combination.
So for tripod use with a Canon 70-200 IS II it seems IS On is the way to go.
There is an issue with extremely long exposure times. the IS will gradually drift. I'm not sure where, exactly, it starts to deteriorate on a tripod but guess it is on exposures longer than 1/10th.
Could you show 100% crops Doug? would be interesting to see.
Regards
-
I once had a night shoot completely ruined when I forgot to turn Canon's IS off.
Could be that the sensor/movement of IS is optimized for some sensible time/movement range, and that if you use IS in 1 hour exposures, some accumulated error will start do do nasty things? Won't an accelerometer have to be integrated twice in order to provide "position", and is not such processing prone to "DC-errors"?
The IS cannot point the camera in a totally different direction. If the sensor says that the camera has moved from one extreme to the other, the sensible thing to do for an IS might be to do a "wrap around", move to the other extreme and start reducing movement over there. Or to throw its hands in the air, go to the center position (where optics are presumably best) and stay there.
I don't know how extensive the camera-stabilizer interaction is. I would assume that IBIS cameras share whatever info they deem necessary (exposure on/off timings, planned exposure time), while lens IS may or may not talk as verbosely with the camera.
-h
-
Could you show 100% crops Doug? would be interesting to see.
Regards
Hi,
I don't actually use target images with sharp lines or edges but rather a specialized image with superimposed sine waves below Nyquist that increase in spacial frequency and I typically photograph them so that I get response magnitudes of 10, 20,30, ... 70 lpmm. I take a raw picture then process it using dcraw -4 -D -T to produce a linear gray scale image w/o CFA interpolation. The individual RGGB plane pixels are rescaled to the image white and black points then the data is analyzed using FFTs.
My main use of this was an industrial QC setting where I could use the responses at the frame corners to determine the precise perpendicularity of the sensor plane to sub micron precision. I've adapted it to look at motion blur since it produces nice, quantifiable numbers and identifies a principal motion axis as well.
I'll dig out a center sample of the target and a couple shots when I get a chance.
-
That is not a tripod. It is a selfie stick with legs. And to make it worse you have it on a timber floor.
-
That is not a tripod. It is a selfie stick with legs. And to make it worse you have it on a timber floor.
I have to agree about that extended center column being a truly bad idea. Turns a stable tried into kind of a monopod - especially difficult outdoors with even the slightest breeze. I removed my (Gitzo) center column and mounted a stable plate into which my 3-way (Manfrotto) gear head is securely fastened.
-
Could you show 100% crops Doug? would be interesting to see.
Regards
Note that these pictures are from a target w/o sharp edges. The spacial frequencies are below Nyquist which eliminates artifacts and variations from the pixel sampling positions when processing the data through the FFT transforms:
All at 1/500th.
208a5338x.tiff: Live view mode 1;
208a5332x.tiff: Standard, No IS
208a5335x.tiff: Standard, w IS
If you compare the images you can see a slight blurring in the vertical on the Standard, no IS, a bit less on the standard w IS compared to the Live view which has no motion blur.
The image is from the circular superposition of 7 sine waves which are spaced at 10, 20, 30, ... 70 lines per mm. From this MTF can be measured and the additional effects of motion blur can be seen.
Corrected. Mirror lockup misstated
-
Isn't it a rather simple conclusion that if your camere is unstable, being handheld or on a flimsy tripod or in a gail storm, then image stabilization will likely help, whereas if your camera is well supported and completely stable it will not.
Accentuated by no (at least little) help at shorter focallenghts and shutterspeeds, and help at longer focallengths and shutterspeeds.
Regards
Johannes Elkjaer Madsen
-
Isn't it a rather simple conclusion that if your camere is unstable, being handheld or on a flimsy tripod or in a gail storm, then image stabilization will likely help, whereas if your camera is well supported and completely stable it will not.
Accentuated by no (at least little) help at shorter focallenghts and shutterspeeds, and help at longer focallengths and shutterspeeds.
Regards
Johannes Elkjaer Madsen
Sure, but it's a matter of degree. Flimsy tripods are generally better than handheld. Even good tripods can have significant vibrations exceeding many pixels until dying down and wind, of course, is another variable.
I have an extremely stable surveyors tripod. They are designed to be pretty tolerant of wind and are quite stiff and stable. They still can suffer from some wind induced motion. Surveyors often need to measure angles to better than 1 or 2 seconds of arc so they make good photo tripods. They are also cheaper. Downside: they are quite heavy. They are ideal for long exposure timed shots where there is too much wind for a regular tripod.
-
Isn't it a rather simple conclusion that if your camere is unstable, being handheld or on a flimsy tripod or in a gail storm, then image stabilization will likely help, whereas if your camera is well supported and completely stable it will not.
IS tries (and usually succeeds) in improving camera shake.
But does it ever make images worse, as suggested by some to be the case when your camera is (more or less) well supported?
-h
-
IS tries (and usually succeeds) in improving camera shake.
But does it ever make images worse, as suggested by some to be the case when your camera is (more or less) well supported?
-h
Image Stabilization OFF when mounted on a tripod is the BOOK answer. It is a generalization that is based on having a stable platform and was issued for the 1st generation of Image Stabilization.
The question how truly stable is the platform with respect to both external and internal forces. There are lots of things that will affect you in a given situation. The type and quality of the stabilization implementation, for example. A 4th generation stabilization solution is likely to be more effective in these types of situation while it is likely a 1st generation would be detrimental. Processing power obviously will contribute positively.
I don't think there is a single answer to your question as there are too many variables. Testing is useful.
-
Image Stabilization OFF when mounted on a tripod is the BOOK answer. It is a generalization that is based on having a stable platform and was issued for the 1st generation of Image Stabilization.
The question how truly stable is the platform with respect to both external and internal forces. There are lots of things that will affect you in a given situation. The type and quality of the stabilization implementation, for example. A 4th generation stabilization solution is likely to be more effective in these types of situation while it is likely a 1st generation would be detrimental. Processing power obviously will contribute positively.
I don't think there is a single answer to your question as there are too many variables. Testing is useful.
I should have been more explicit in my post.
1. I think that everyone agrees that whenever there is motion, IS tends to generally make things better (often a lot).
2. I believe that what is contested is the "BOOK answer" that IS might reduce image quality if there is no shake.
I think that the latter is the question that it makes sense to try to answer, and indeed what I expected when I opened this thread. Resting the camera on a concrete slab with and without IS might (?) provide a decent answer to that question.
-h
-
I should have been more explicit in my post.
1. I think that everyone agrees that whenever there is motion, IS tends to generally make things better (often a lot).
2. I believe that what is contested is the "BOOK answer" that IS might reduce image quality if there is no shake.
I think that the latter is the question that it makes sense to try to answer, and indeed what I expected when I opened this thread. Resting the camera on a concrete slab with and without IS might (?) provide a decent answer to that question.
-h
I agree, we make generalizations because they are useful. We can all agree that without any motion at all, IS can not do better than No IS. What I cannot answer is will it be detrimental? My suspicions is older IS implementations [biased by Nikon experience] will degrade the image while newer ones might not. I also suspect that having more processing power thrown at the solution, meaning new camera bodies, probably lessens the effect. However, this would be dependent on the design of the processing algorithm scheme which I am not privy too.
I just think there are too many different lens in play with different IS and too many different camera bodies to come up with a useful generalization. 10 years from now, IS might be an always on feature! I know with VR I on a Nikon body of the time, it was generally accepted that VR should be turned off with shutter speeds over 1/400th - 1/500th because it would blur the image. That is no longer the case for newer VR implementations. I used higher shutter speeds with the 200-500mm Nikkor and VR Sport ON with the rig mounted on a monopod last night.
-
As you said, there are many implementations out there and it is hard to come up with a simple generalization. Following what the manual for the particular equipment says is a good start, but as with all things it's up to you to test and push your gear to achieve what you want.
-
BTW I did not show the 100% crops from a "no-wind" situation:
(http://www.guillermoluijk.com/article/tripodandis/sinviento_ibisoff.jpg)
(http://www.guillermoluijk.com/article/tripodandis/sinviento_ibison.jpg)
There is no practical difference in this case.
http://www.guillermoluijk.com/article/tripodandis/index.htm
Regards
-
I tested the a7II with a 180 and a big RRS tripod, and came to a different conclusion:
http://blog.kasson.com/?p=8391 (http://blog.kasson.com/?p=8391)
The situation is a bit more muddled with the a7II with a light tripod, and even more so with the a7R, a camera that interacts with OSS in some strange ways:
http://blog.kasson.com/?p=9743 (http://blog.kasson.com/?p=9743)
Jim
It would be nice to have a standard test of IS, including on and off good and bad tripod setups. Is there any such thing? I assume DPReview and other similar sites follow a standard approach. It would also be interesting to have Jim's meticulous test applied to Guillermo's camera and lens combination (edit: or vice versa). I suspect the differences Jim detects might be hard to see on actual images. My minimal conclusion so far is that forgetting to turn IS off on my suboptimal tripod might not always or usually end in tears.
-
It would be nice to have a standard test of IS, including on and off good and bad tripod setups. Is there any such thing? I assume DPReview and other similar sites follow a standard approach. It would also be interesting to have Jim's meticulous test applied to Guillermo's camera and lens combination (edit: or vice versa). I suspect the differences Jim detects might be hard to see on actual images. My minimal conclusion so far is that forgetting to turn IS off on my suboptimal tripod might not always or usually end in tears.
Because of the difficulty many of us have in translating MTF curves to what actual images would look like, I once made a set of images which were carefully degraded by precise amounts of motion blur, and posted them together with the MTF corresponding to that amount of motion blur.
http://blog.kasson.com/?p=9928
http://blog.kasson.com/?p=9946
http://blog.kasson.com/?p=9960
By the way, in order to evaluate the Kolari a7II mod, which changes the weight of the sensor assembly and thus potentially how well IBIS works, I came up with a controlled vibration technique based on a shaker table. It ultimately proved sufficiently non-repeatable to compare cameras, lenses, or techniques across different testing sessions.
Jim
-
That is not a tripod. It is a selfie stick with legs. And to make it worse you have it on a timber floor.
;D. Yes, Guillermo's test has shown that IS on can be useful with a less than solid setup. If you are using a monster tripod on granite with a sandbag on the camera, results could be different.
-
Because of the difficulty many of us have in translating MTF curves to what actual images would look like, I once made a set of images which were carefully degraded by precise amounts of motion blur, and posted them together with the MTF corresponding to that amount of motion blur.
http://blog.kasson.com/?p=9928
http://blog.kasson.com/?p=9946
http://blog.kasson.com/?p=9960
By the way, in order to evaluate the Kolari a7II mod, which changes the weight of the sensor assembly and thus potentially how well IBIS works, I came up with a controlled vibration technique based on a shaker table. It ultimately proved sufficiently non-repeatable to compare cameras, lenses, or techniques across different testing sessions.
Jim
Nice work.
-
;D. Yes, Guillermo's test has shown that IS on can be useful with a less than solid setup. If you are using a monster tripod on granite with a sandbag on the camera, results could be different.
I realize this is sarcastic, but it highlights the real issue here. That is, Where is the crossover point? Knowing that no rig is ever perfectly stable and there are always internal and external forces exerted on it, do I use IS or not?
Personally, I think we will eventually arrive at an IS ALWAYS ON state sometime in the future where the mechanicals, algorithms and processing power will be such that the camera will decide based on it's data collection and analysis.
Until that point, it is up to everyone to test their own equipment, make their own judgments and build their own experience bank.