PDA

View Full Version : IS and tripods


StephenL
4th February 2009, 09:33 AM
I know this subject has been aired a few times before, but I thought I'd throw in my real-world observations. ;)
It is the commonly held belief, supported by statements in the relevant Olympus instruction manuals, that you should turn-off IS when using the camera on a tripod, as it would try to compensate for non-movement. Well, my recent experience may throw a pigeon amongst the cats.
I was taking a few shots with my E-3 on a tripod - long timer shots at speeds in excess of 30 seconds, and I left IS on (OK, I forgot to switch it off! :o ).
But when I came to edit these shots, I found the results to be razor sharp, indicating that either:

a) the above mentioned instructions were incorrect
or
b) I have a rubbish tripod
or
c) no tripod can guarantee absolutely stable support.

I will add that the shots were taken with the 'pod on a stone floor indoors with no wind or other vibrations.

Any thoughts? *chr

Xpres
4th February 2009, 10:11 AM
I think you have your answer in b or c.
A lot can happen in 30 seconds even on a heavy tripod - which is what you need as a lightweight one will be no use at all in this sort of shot. How you use the tripod is also important for stability, but in your case you seem lucky enough to have the right combination for sharp shots so don't change it! :) Keep IS on for long exposures.
Option d of course ... the IS is broken?

photo_owl
4th February 2009, 10:32 AM
love to see a 100% crop of an area of detail from the image please

RogerMac
4th February 2009, 10:43 AM
:)A couple of points that are not mentioned in the manuals (but I am sure must be relevant) are:
1. The gyroscopes that measure the camera movement will "drift" over time and there must be a maximum length of time over which they can measure movement (or lack of it)
2. When the sensor is pushed in one direction the camera is also pushed in the opposite direction (Newton's laws of motion) and this camera movement must be allowed for in the complex calculations to detemine the sensor position and movement. If the camera becomes a lot heavier, by being being firmly attached to a heavy tripod, those calculations will be wrong. As the designers have no idea how heavy the tripod will be they play safe and just adivise that IS is switched off.

It sounds as though in your case the gyros did not drift and there was in fact no camera movement to compensate for (thank that stone floor) and you got away with it.

Incidently I have made the same mistake myself and also got away with it. I think the message is don't wobble the tripod!:)

Roger

Archphoto
4th February 2009, 10:49 AM
Roger, I think you are right with that.
If you leave the camera on the tripod for a minute or so, the "gyro's" would have been stabilized and you should be OK with IS "on" .

I will make some testshots the next few day's as I am relatively new to IS (got my E520 on december 30th) and will be doing so asignments in the next weeks: on and off the tripod, mostly interior shots.

Peter

Archphoto
4th February 2009, 12:25 PM
I made those test-shots this morning (for the UK: afternoon):

E520 with 40-150mm set at 150mm, a building about 1km away, f:16, tripod: Manfrotto 055 with 410 head, cable release by Oly
The camera was set to the ON position on the tripod and left for 1 minute, so the IS could stabilize.

1) IS ON, mirror down
2) IS ON, mirror up
3) IS OFF, mirror up
4) IS OFF, mirror down

Result: it makes more diference whether the mirror is up or down, up being a bit better.

A will do an other testrun tonight and see if there any diferences.

Peter, Goi‚nia, Central Brazil

Steve Lane
4th February 2009, 01:56 PM
I recall doing a similar comparison when Canon's 10D came to market. I used a really sturdy tripod and the camera's self timer. All shots taken with IS switched on were not as sharp as when it was switched off. I guess the difference between the camera's is that the Canon's IS was in the lens and Olympus' is in the body.

Cheers, Steve.

StephenL
4th February 2009, 02:02 PM
Some interesting comments there.
I must add that something else I'd forgotten to do (old age) is lock up the mirror. But I did use a cable release!
I think it also emphasises my view that very few tripods are absolutely rock solid.
Also, at the risk of appearing even more ignorant, can I have advice on how to publish, and what is meant by, a 100% crop?

RogerMac
4th February 2009, 02:04 PM
I made those test-shots this morning (for the UK: afternoon):

E520 with 40-150mm set at 150mm, a building about 1km away, f:16, tripod: Manfrotto 055 with 410 head, cable release by Oly
The camera was set to the ON position on the tripod and left for 1 minute, so the IS could stabilize.

1) IS ON, mirror down
2) IS ON, mirror up
3) IS OFF, mirror up
4) IS OFF, mirror down

Result: it makes more diference whether the mirror is up or down, up being a bit better.

A will do an other testrun tonight and see if there any diferences.

Peter, Goi‚nia, Central Brazil

Very interesting - thanks for posting

Roger

photo_owl
4th February 2009, 03:31 PM
juzt stick the images in your gallery here and post a link.

re the 100% bit just zoom to 100% on your photo editor and crop out a chunk of detail large enough to view the detail and post that at it's full pixel size to the gallery. This way we can open in at 100% on screen and see what you are seeing.

In my 'definitive' tests I could cleary see the impact at 100% but it was almost impossible to detect at 50%.

StephenL
4th February 2009, 03:58 PM
OK, here goes. A couple of things to note - the exif says light source was flash. I didn't use the flash! Also, it's unsharpened, straight from the raw.

http://e-group.uk.net/gallery/data/500/e-3-40277crop.jpg

Archphoto
4th February 2009, 04:08 PM
Nothing wrong with the shapness, looks good.

Gwyver
4th February 2009, 04:10 PM
I'm most likely wrong but I think this talk of gyros in the IS system is somewhat misleading - at least in the sense in which most folk think of gyroscopes.

As I understand it the in-body IS mechanism comprises a suspended mass with adjacent sensors (either capacitive or piezoelectric etc) which sense change of motion in the suspended mass at the time of full shutter press. This sensed change is then used to briefly nudge the Image Sensor assembly in the opposite direction to cancel out the detected motion.

Hence, because the IS isn't being continuously corrected, if Stephen used a long exposure as stated the effects of the IS would have been transient and imperceptible - hence his resulting capture is not blurred.

If the IS mechanism was a continuous process it is most likely that battery life would be drastically curtailed whenever the feature was switched-on - and this is evidently not the case.

StephenL
4th February 2009, 04:20 PM
So you're saying that if I'd used a shorter long-exposure (eg 1/4 sec), the effect of the IS intervention would have been (more) noticeable than a long (30 sec) long-exposure?



Hence, because the IS isn't being continuously corrected, if Stephen used a long exposure as stated the effects of the IS would have been transient and imperceptible - hence his resulting capture is not blurred.

Archphoto
4th February 2009, 04:22 PM
Thanks for the idea Gwyver, I will incorporate that in my testshots tonight.

Peter

Gwyver
4th February 2009, 04:58 PM
So you're saying that if I'd used a shorter long-exposure (eg 1/4 sec), the effect of the IS intervention would have been (more) noticeable than a long (30 sec) long-exposure?

Stephen,
Yes. It's analogous to the ghosting which arises when objects move through a long exposure. The longer the exposure the fainter the ghost (assuming the object moves at a constant rate).

The intention of IS is only to enable acceptable hand-held photos at approx 2 or 3x slower shutter speeds than 'normal'. It has never been claimed to yield (significant) improvements when using shutter speeds of greater than approx 1/10sec.

mike_j
4th February 2009, 06:26 PM
So you're saying that if I'd used a shorter long-exposure (eg 1/4 sec), the effect of the IS intervention would have been (more) noticeable than a long (30 sec) long-exposure?

I think IS operates as long as the shutter is operating. Have you looked at the effect of IS by pressing IS when using live view? It is quite impressive.

Also - re tripod ridigity - use live view in x10 and gently tap the lens barrel, it's quite a shock.

RogerMac
4th February 2009, 10:55 PM
I'm most likely wrong but I think this talk of gyros in the IS system is somewhat misleading - at least in the sense in which most folk think of gyroscopes.



I am not sure what definition is being used for "gyro" but here is an extract from the E30 press release:

"This enables the photo-enhancing effects of image stabilisation to be enjoyed irrespective of the attached lens. In Olympus E-System cameras with built-in image stabilisation, a gyro sensor detects the precise direction of the shake. Using the on-board SWD, the shake is then compensated for by shifting the image sensor according to the movement data. For quick response at long focal lengths, cameras are equipped with two dedicated microprocessors: one to control the two supersonic motors that shift the unit vertically and horizontally, and one to provide high-precision control. "

roger

photo_owl
4th February 2009, 11:17 PM
Stephen,
Yes. It's analogous to the ghosting which arises when objects move through a long exposure. The longer the exposure the fainter the ghost (assuming the object moves at a constant rate).

The intention of IS is only to enable acceptable hand-held photos at approx 2 or 3x slower shutter speeds than 'normal'. It has never been claimed to yield (significant) improvements when using shutter speeds of greater than approx 1/10sec.

This is hand held - about 1.5 seconds exposure time (with IS and me leaning against a lampost)

http://farm3.static.flickr.com/2401/2105750427_aa3644129a_o.jpg

Nothing is ever proof but it suggests to me that IS can deliver on over 1/10th sec shutter speeds.

On the other hand this comparison illustrates the effect of IS image degredation on an image taken on a solid tripod, mlu, remote trigger - 2 identical images (AWB gave slightly different colour but I left it uncorrected)one after the other with only IS changed.
Apologies for the wide file but at less than 100% you can't really see the degredation. I have tried similar tests with fast shutter speeds and cannot get any evidence of degredation - although it's recomended to turn it off for those in the manual as well.

http://farm4.static.flickr.com/3147/2825008239_f42302ac10_o.jpg

Archphoto
5th February 2009, 12:40 AM
I did as promissed my testshots tonight.
The results I got are more or less the same as photo-owl got, although less pronounced. (30 second exposure at f:8 )

I get a bit the impression that the IS circuit gets into trouble with noise at longer exposures, otherwise I have no other explanation for it.

So: on a tripod: YES, IS OFF

And if you forgot to swich the IS off and get a great shot anyway: feel lucky.

Peter