Any evidence for turning of IS on current gen lenses when "not needed"

[fjrabon]

edit: ugh, nothing worse than a typo in a thread title; obviously I meant "turning off IS"

so, I often see the "turn off your IS if you don't need it" trope bandied about here. When on a tripod, when shooting faster than shutter speed 1/200, etc.

My question, is there any actual evidence for this on current gen IS/VR lenses? I know that Canon's manual for the 70-200 f/2.8 IS II L lens specifically says that you should leave it on, even on tripods and high shutter speeds, as the lens's autofocus is designed to work with IS engaged and that they claim there is no benefit to turning it off anyway.

Now, I know that even 5 years ago, this probably wasn't the case, you could absolutely see the difference when IS was engaged with a tripod mounted lens. But I've run some tests myself and with both Nikon and Canon's current gen IS, I just can't see any difference, even when pixel peeping, with IS turned on while using a tripod. And in my experience, it absolutely does help Canon's lenses focus both faster and more accurately in low light action situations. When I first started shooting sports with the 70-200 IS II L lens, I turned IS off, based on advice I received here, and at times I was mildly surprised by it not being the focusing beast I had always heard about. Then I read the manual and read that bit about leaving IS turned on for focusing speed and accuracy and it was a noticeable difference, especially in servo and continuous modes.

I asked a Canon rep about this once and he said he thought they wouldn't have even included a switch for turning IS off on their current gen lenses, except for the fact that he figured people might look for it when buying, and might not buy a lens that wouldn't allow you to turn IS off.

Anyway, I'm wondering if anybody has done tests on current gen IS systems to see if there actually is a benefit to turning them off when tripod mounted or shooting high shutter speeds.

 

[Braineack]

view this at 100% and you'll see why you should turn it off:

DSC 7219-10 Flickr - Photo Sharing

 

[fjrabon]

view this at 100% and you'll see why you should turn it off:

DSC 7219-10 Flickr - Photo Sharing

the eyes? To me the ring, shirt and hands look sharp, to me that looks as much like not having much DoF (f/4 on a full frame at 60mm at close range on a moderately deep subject) as it does VR issues.

 

[Braineack]

if VC was off (or shutter speed slower), that shot would look sharp enough to cut your face.

I can find other examples, but I've noticed that with higher than 1/500 shutter speeds and if I leave VC on, the image is never as sharp as without VC on--It has a weird motion blur look to the image.

 

[fjrabon]

if VC was off (or shutter speed slower), that shot would look sharp enough to cut your face.

there are parts that are sharp, there are parts that are out of focus? VR issues would kill sharpness picture wide, not just in certain areas, right?

 

[fjrabon]

if VC was off (or shutter speed slower), that shot would look sharp enough to cut your face.

I can find other examples, but I've noticed that with higher than 1/500 shutter speeds and if I leave VC on, the image is never as sharp as without VC on--It has a weird motion blur look to the image.

yeah, not necessarily doubting you, just hard to make heads or tails off one image with no direct comparison.

 

[fjrabon]

also, what lens was that? Nikon's 24-70 doesn't have VR, right? Is that the Tammy?

 

[Braineack]

yeah, see my sig. I like the Tammy stuff.

I can do some bench tests to show you. I will definitely notice it on shots where I leave VC on and shoot fast shutter speeds.

 

[fjrabon]

yeah, see my sig. I like the Tammy stuff.

I can do some bench tests to show you. I will definitely notice it on shots where I leave VC on and shoot fast shutter speeds.

ah, yeah, that makes sense. I love the Tammy stuff too (my favorite APS-C lens of all time is a Tammy), but I think their VC may be a bit behind Canon and Nikon.

 

[KmH]

There is a limit to how fast in-the-lens image stabilization can move.
That's why IS need not be turned on above some shutter speed threshold.

While Nikon specific this may give some insight into in-the-lens image stabilization regardless the lens maker:
Nikon VR explained

 

[Braineack]

I prefer it over VRI and VRII.

 

[fjrabon]

There is a limit to how fast in-the-lens image stabilization can move.
That's why IS need not be turned on above some shutter speed threshold.

While Nikon specific this may give some insight into in-the-lens image stabilization regardless the lens maker:
Nikon VR explained

Sure, I can maybe get behind the idea that VR can be bad at faster than 1/500, but I think people here go too far with that. Advocating turning off VR at shutter speeds like 1/250 or when on a tripod. and we also have precious little actual evidence of it with current gen systems, hence the genesis of this thread.

Do we have any evidence of this for Canon and Nikon's current gen? Other than Thom Hogan's hypothesizing and what photographers have done out of habit due to the originally fairly bad IS/VR.

 

[table1349]

Besides what Braineack mentioned those things don't power themselves. IS/VR gets it power from your camera battery. When I am shooting two back to back games I don't need to waste battery life for something that is not doing any good.

 

[Derrel]

if VC was off (or shutter speed slower), that shot would look sharp enough to cut your face.

I can find other examples, but I've noticed that with higher than 1/500 shutter speeds and if I leave VC on, the image is never as sharp as without VC on--It has a weird motion blur look to the image.

Exactly--a weird motion blur type of look to the image! That's what I see in this shot. The girl's hair looks stopped. But you know, I think there's something else happening there that mivght not be the VC kicking in. It looks to me like what you're seeing is the fill-flash in high speed synch mode, and the flash's oscillation is recording maybe, oh I dunno. 10,15,20 microbursts of the flash, and that is why their faces are blurred. I've had the same phenomenon happen to me shooting Nikon SB 800 + 200VR in bright August sun in the 1/2500 second range: what looks like motion blur, due to the way FP Synch flash is not one, single pop! but actually an ultra high-speed series of microbursts, so closely spaced that the human eye cannot see them. The daylight is bright, and the wide f/stop of f/4 is gathering a lot of light, so I think the woman's face especially is moving, and that might be 5-10 stacked images of her face in motion, each one just milliseconds apart from the prior and following microburst.

I do not have any brand-new VR-II lenses that I know of, but one summer I ruined many portraits with the VR-1 70-200, leaving it on in bright weather conditions...I botched a lot of stuff with that lens.

 

[fjrabon]

if VC was off (or shutter speed slower), that shot would look sharp enough to cut your face.

I can find other examples, but I've noticed that with higher than 1/500 shutter speeds and if I leave VC on, the image is never as sharp as without VC on--It has a weird motion blur look to the image.

Exactly--a weird motion blur type of look to the image! That's what I see in this shot. The girl's hair looks stopped. But you know, I think there's something else happening there. It looks to me like what you're seeing is the fill-flash in high speed synch mode, and the flash's oscillation is recording maybe, oh I dunno. 10,15,20 microbursts of the flash, and that is why their faces are blurred. I've had the same phenomenon happen to me shooting Nikon SB 800 + 20VR is bright August sun in the 1/2500 second range: what looks like motion blur, due to the way FP Synch flash is not one, single pop! but acutally an ultra high-speed series of microbursts.

I do not have any brand-new VR-II lenses that I know of, but one summer I ruined many portraits with the VR-1 70-200, leaving it on in bright weather conditions...I botched a lot of stuff with that lens.

yeah, that might explain why you see the effect in the bright areas (face, eyes and hair), but not as much in the darker areas (hands, lower portions of the shirt).

Also, yeah, I was never a fan of VR-I. Things always looked weird to me, even in cases where VR was useful, it still looked weird.