[Marc Sabatella]

Even if it doesn't help at faster shutter speeds - and empirical tests suggest that is true - it isn't going to hurt, either. Turning it off is just wasted effort, and makes it that much more likely that you'll forget to turn it back on when you do need it.

Also, note 1/250" isn't necessarily fast enough to not benefit from SR. It would be with your 10-17 at the wide end, but not with the 70-300 at the long end, and certainly not even close with your 500.

[JHD]

However if you are shooting with a fast focusing lens, as long as the shutter speed is high enough, turning the SR off might not be a bad idea... With the DA 40 and the Sigma 50-150 I find the focus is always ready before the SR is stabilized. The resulting delay is tantamount to shutter lag... It can be quite annoying.

[Marc Sabatella]

Another situation where SR can lag is when pre-focusing, using the button on the back of the camera to enable - or temporarily cancel - AF. I might focus once then shoot a whole series. And, of course, when using manual focus. Still, these tend to be the situations where SR is going to be most valuable, so I rarely turn it off. I just try to remember to wait for it, and use the half-press as much as possible even while in MF mode to warm up the SR.

[Winder]

I think the biggest reason for blurry images with Pentax cameras has to do with the SR no being ready when the shutter is released. My Sigma 50mm f/1.4 is significantly faster at focusing than the SR system is at getting ready. It wont do Pentax any good to develop a blazing fast AF if the SR takes forever to spool up.

[geekette]

I am thinking my blurry images came from ignorance of not knowing to watch for the SR hand in the viewfinder before snapping; now that I know, I am trying to make sure I see that hand every time before I snap. Thanks for all the help / replies!

[Gimbal]

I think the biggest reason for blurry images with Pentax cameras has to do with the SR no being ready when the shutter is released.

But then this is not a specific Pentax problem but a stabilization problem in general, no matter what system is used. Even lens based systems needs time to settle before shooting.

Think about it from the cameras point of view.
Say that the user suddenly turns the camera 1 degree to the left and fires.
Now the camera has a problem, was that turn a shake that should be countered, or did the user intentionally change the composition right before he fired?

The camera has no way of knowing this unless you first aim for a while letting the camera know that this is the direction you want to shoot, any deviation from this should be countered against.

Therefore I only turn off SR if I am about to shoot things that involves fast and sudden camera movements, otherwise chances are that I trick the SR to try to counter my movements and the result would not be that great.

[Winder]

But then this is not a specific Pentax problem but a stabilization problem in general, no matter what system is used. Even lens based systems needs time to settle before shooting.

Think about it from the cameras point of view.
Say that the user suddenly turns the camera 1 degree to the left and fires.
Now the camera has a problem, was that turn a shake that should be countered, or did the user intentionally change the composition right before he fired?

The camera has no way of knowing this unless you first aim for a while letting the camera know that this is the direction you want to shoot, any deviation from this should be countered against.

Therefore I only turn off SR if I am about to shoot things that involves fast and sudden camera movements, otherwise chances are that I trick the SR to try to counter my movements and the result would not be that great.

I have used the Olympus E-3 and the Sony A900. The Olympus IS is definitely faster that the K-7. If I remember correctly the A900 turns on the SR when you raise the camera to your eye, so it activates sooner and is ready sooner. The A900 SR might not be any faster than the Pentax IS, but because it activates when you bring the camera to your eye is seems to always be ready when you are ready to release the shutter. Olympus has the best IS I have used so far. I only used the A900 for a few events, so I have limited experience.

Pentax definitely could improve the speed of the SR.

[m42man]

...Think about it from the cameras point of view.
Say that the user suddenly turns the camera 1 degree to the left and fires.
Now the camera has a problem, was that turn a shake that should be countered, or did the user intentionally change the composition right before he fired?

The camera has no way of knowing this unless you first aim for a while letting the camera know that this is the direction you want to shoot, any deviation from this should be countered against...

Yes, the reason behind the delay is somewhat puzzling. However, my take on it is as follows:

1. Finger off shutter button. The sensor-positioning control system uses sensor-position sensors to hold the platter in the "nominal" position, ready to take a shot. The vibration sensors aren't being used yet.

2. Shutter button halfway depressed. AF takes place, but I don't see why the SR mechanism has to do anything different at this stage. Are we saying that we're not allowed to recompose - especially after the hand appears?

3. Shutter button is fully depressed. I would have expected that here the camera inserts a very short delay in order to allow the SR mechanism to switch from the sensor-position sensors to the vibration sensors as its reference. Then the mirror flips and the shutter is fired, and during this period the vibration sensors tell the SR control system where to position the platter.

So why do we have to wait for the hand symbol? My thought was that in (1) above, the camera might limit the current available to the SR mechanism, to save power. Then it would need some time to maybe reposition the platter more accurately during (2). But this would imply that the camera consume more power when SR is turned off, which seems very unlikely.

So what's really going on while we wait for the hand? Anyone have any definitive knowledge?

[Gimbal]

Yes, the reason behind the delay is somewhat puzzling. However, my take on it is as follows:

1. Finger off shutter button. The sensor-positioning control system uses sensor-position sensors to hold the platter in the "nominal" position, ready to take a shot. The vibration sensors aren't being used yet.

2. Shutter button halfway depressed. AF takes place, but I don't see why the SR mechanism has to do anything different at this stage. Are we saying that we're not allowed to recompose - especially after the hand appears?

3. Shutter button is fully depressed. I would have expected that here the camera inserts a very short delay in order to allow the SR mechanism to switch from the sensor-position sensors to the vibration sensors as its reference. Then the mirror flips and the shutter is fired, and during this period the vibration sensors tell the SR control system where to position the platter.

So why do we have to wait for the hand symbol? My thought was that in (1) above, the camera might limit the current available to the SR mechanism, to save power. Then it would need some time to maybe reposition the platter more accurately during (2). But this would imply that the camera consume more power when SR is turned off, which seems very unlikely.

So what's really going on while we wait for the hand? Anyone have any definitive knowledge?

At point 1, the image sensor (platter) isn't positioned at all, it lays loose at the bottom of it's well free to move around if you tilt the camera. The famous "clunk" noise can be heard if you tilt the camera.

At point 2, the accelerometer sensor activates and starts measuring the background shake frequency and amplitude, also a sort of general direction builds up. When all necessery data is collected the hand lits up, you should not recompose unless you stay in the new composition for 0.7s before you fire.
The platter is still loose and no power is feed to the positioning servos.

At point 3, the mirror starts flipping and the positioning servos powers up and moves the platter from wherever it is to its calculated position and, hopefully, moves in sync with the background shake and also ready to counter any movement from the general direction built up during the 0.7( or so )seconds before the shutter was pressed. Now the shutter blades opens.

If you fire before the SR system is ready (hand is not lit) the mirror flips and the positioning servos lift the platter to its center position and holds still, exactly as is it would do if the SR system was shut off.

When people claim that the SR ruined their shot because it wasn't ready, I bet it really was because they changed their aim right before they fired and the SR tried to counter it. I call it user error.

However, the K5 have changed the SR system behavior in that it does stay active as long as the metering is active (you don't have to hold the shutter half pressed). Also notice if you now recompose when the hand is lit, the light goes out showing that the data collected is thrown away and a new 0.7s cycle is started collecting data in the new direction.

[m42man]

Aha, thanks for that, Gimbal. So, basically the sensor is free to rattle around until you actually take the shot, when it (very) rapidly takes up its correct position while the mirror's flipping. Got it!

Very interesting about the recomposing - it would seem you have to be rather careful. I must say I haven't had any problems here, but that's probably because I've been rather slow and deliberate. (I always use centre-spot AF.)

Very interesting also that the K-5 has incorporated an improvement in the hand symbol functionality.

So, the bottom line (especially if you don't have a K-5) seems to be not only to wait for the hand symbol before shooting, but to be very careful with the camera once the hand is lit.

[m42man]

Ah, it's just dawned on me as to why the camera takes its time before it can use the shake sensor information: the shake sensors are of course accelerometers, and the SR system needs to know velocities before it can do its stuff. No wonder it needs 700ms...

[conradj]

inadvertently i did some SR testing of my K-7 a few days ago when i was lens testing, comparing various zooms to the FA28-105mm Tamron rebadge.

Found that at the long end of these lenses, around 105mm, at shutter speeds 1/80-1/125s, i was getting less blurry results when i handheld the camera than when i had it mounted on a monopod. Found also that i was able to get some clearer shots when i handheld the camera at 1/60s, than at faster shutter speeds up to 1/180s either hand held or on a monopod. Got a lot of blurry shots too at 1/60s, but the clear ones were clearer than anything i got 1/80-1/180s.

Shots became clear again from 1/200s, monopod or handheld.

Wasn't doing any regimented testing specifically for SR, so just anecdotal stuff.

1/60-1/200s is a shutter speed range i end up in a lot with these short tele zooms in daylight, a bit of a hassle to not have the camera SR behavior sorted out in my head; going to have to test just for that and figure it out.

[Gimbal]

inadvertently i did some SR testing of my K-7 a few days ago when i was lens testing, comparing various zooms to the FA28-105mm Tamron rebadge.

Found that at the long end of these lenses, around 105mm, at shutter speeds 1/80-1/125s, i was getting less blurry results when i handheld the camera than when i had it mounted on a monopod. Found also that i was able to get some clearer shots when i handheld the camera at 1/60s, than at faster shutter speeds up to 1/180s either hand held or on a monopod. Got a lot of blurry shots too at 1/60s, but the clear ones were clearer than anything i got 1/80-1/180s.

Shots became clear again from 1/200s, monopod or handheld.

Wasn't doing any regimented testing specifically for SR, so just anecdotal stuff.

1/60-1/200s is a shutter speed range i end up in a lot with these short tele zooms in daylight, a bit of a hassle to not have the camera SR behavior sorted out in my head; going to have to test just for that and figure it out.

Sounds like the famous k7 shutter blur problem...

[conradj]

Sounds like the famous k7 shutter blur problem...

o yeps.

i just found it very interesting that in the affected shutter speed range, and down to at least 1/60s, using a short tele, planting the camera on a monopod made the blur problem worse than hand holding the camera.

Hand holding the camera doesn't make every shot less blurry, just when i'm careful and keep it as steady as possible.

At least a 1/20s and below, using a monopod gives me less blurry shot than hand holding.

All in all non-linear behavior that's a bit frustrating.

[Ken T]

SR is always ON for me. I don't even turn it off when on tripod simply because I tend to forget when camera is not on tripod. Instead, I use timer/remote which itself turns it off automatically

Same here .