Full Version: Testins IBIS
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2
I've bought my first camera with IBIS (a6600) a year ago, but until now I didn't pay attention to the new stability tool (IBIS) since I've used the camera only with OSS lenses (the only exception being the Tamron 20mm f/2.8, but the focal is so short that motion blur is easily avoided by shutter times).

Looking at the first few shoots with the Sigma 100mm f/2.8 macro it seems that IBIS is not working: blurred shots around critical shutter times. All of a sudden I'm worried about IBIS not working in my copy (and unfortunately it would be too late for the warranty...). I need to test it in a reliable way. My idea is to take a batch of landscape shots at various shutter speeds and create a statistics out of results - it's the standard way of doing that, if I'm not wrong. How many samples do you think I need? Is it appropriate to use continuous shooting mode or is it better to take single shots? Other suggestions?

image stabilisation isn't 3D, it only corrects in the vertical and horizontal axis, the movement closer and farther from the subject isn't corrected, usually for normal photography such movement is irrelevant, however for macro work it's crucial .
Is the problem only at close focus distance or at all focus distances ?
If it's only at close focus IBIS is just fine.
I'd suggest you start shooting at approximately 50x effective focal length for distance to start off with, of a static object, with and without IBIS switched on.
The problem with IBIS is that it depends a lot on the person whether it will work properly. Some have steadier hands than others, which is why you need to compare shots with and without IBIS, at the same apertures and shutter speeds.

In addition, IBIS doesn't work too well with macro due to the additional magnification you get at macro distances. In that case it is again a matter of comparing what you can do handheld without IBIS, vs what is possible with IBIS, so each time at the same focusing distance with the same aperture and shutter speed, with and without IBIS.

In itself shooting landscapes is fine for a test like that, just that the objects in a landscape often are fairly far away, and make it therefore difficult to see whether something is sharp or unsharp due to OOF objects, haze, a too slow shutter speed, or just shake, which is why I suggested 50x effective focal distance.  At least objects are more clearly visible that way. You could also shoot a brick wall with fine detail, from, say, 5 meters away, with and without IBIS switched on, all other parameters staying the same. It is much more convenient to do it that way than shooting landscapes. You could also try a test chart, for that matter.

HTH, kind regards, WIm
Thanks for the tips. Yes, the problem is not with macro (I don't expect a lot of with macro shots for the reasons you wrote), but with a few landscape shots I tried. I'll follow Wim's advice for the distance of the test subject.
Do let us hear how you fare, please.

Very interested here to see the results!

Kind regards, Wim
First round of tests done. Procedure: exposure: using zebra; drive mode: single shot, roughly one shot per second; ten shots per each set; pre-focused only once at the beginning.

Subject at roughly 5 m (IS0 100 fixed, changing aperture from ƒ/2.8-9-18 since it's in the shadow), percentages of sharp shots:

1/800: 100%/100% (IBIS on/off)
1/80: 80%/40%
1/20: 70%/0%

Subject at roughly 1000 m (ISO 100 and 50, changing aperture from ƒ/4-11-22 since it's in the sunlight):

1/1250: 100%/100% (IBIS on/off)
1/80: 70%/50%
1/20: 40%/20%

Of course in evaluating sharpness I focused on motion blur, since absolute sharpness varied because of diffraction when stopping down a lot. I'll try to redo the test in more homogeneous conditions, but I think it's enough to prove my IBIS isn't broken. At this point it probably makes sense to refine the test to have a rough estimate of the actual stops gained.
FWIW, I'm usually getting ~1/2 of the claimed IS gain based on what I consider to be critically sharp - on a fairly reliable basis that is. Of course, this doesn't mean that I don't get lucky shots beyond. I'm always baffled by the claims of some people out there but it's always the question of whether they've got the same baseline regarding critical sharpness.

Sony's IBIS is certainly not the best to start with. Oly's IS is magic in comparison for instance.

What do you think of the Sigma lens? Did the AF disappoint or is it alright?
Well, 40-70% of sharp shots at 1/20 is good for me (macro apart).

For what concerns the lens, I'm still collecting data. I can understand why some people blamed its AF: if you're used to superfast and completely silent AF, this is not the case. But noise can only harm those who do audio recording. For the speed, at the moment I can say I've lost some shots because of it, but I am not yet acquainted with the ergonomics. I mean: the focus limiter is certainly important and my finger doesn't find it automatically yet, furthermore I'm not able yet to visualize where the limit between the two zones falls. It's likely that those shots have been lost because of the time I spent searching for the button, and in some cases it wasn't properly set. So I won't draw a conclusion before a certain number of sessions.
Interesting, and thank you.

It obviously varies from person to person, some people are more stable / don't move as much as others, obviously.
Personally I managed to get very sharp shots with a Canon 100-400 L IS Mk I at 400 mm and 1/40 s, mounted on APS-C, no problem, repeatable.

However, I have to agree with Klaus. When it comes to IBIS, I think Oly is still the one to beat currently, despite its in-built magnification factor. Half a second is generally not a problem for me with a stabilized lens and body.

Kind regards, Wim

P.S.: Before I forget, in most instances the cut-off point for effective IBIS is considered the slowest shutter speed where you have at least 50% of images sharp.  So, 70% at 1/20 at 5 m is actually quite good. It means more than 3 stops effective IS considering you are using it on an APS-C camera. For landscapes that appears to be around 1/40s based on what you report, so about 2 stops gained. The 100 mm is in comparison to FF effectively a 150 mm lens, so you'd need at least 1/150s for a sharp shot based on the general rule of thumb for sharp pictures. This is the basis of my calculation here.
Re your PS: yes, I agree, it's the rough estimate I'm doing too. I'll redo the test with slower shutters, in order to see where it's still possible to get a sharp shot without tripod.

“It obviously varies from person to person, some people are more stable / don't move as much as others, obviously.”

And unfortunately I'm noting I'm slightly getting more and more shaky...

Or is it “shakier and shakier”?
Pages: 1 2