We need to talk about a chart that Fujifilm executives pulled up during the recent introduction of the GFX100RF, a camera that does not have sensor-based or lens-based image stabilization:
We have a lot to unpack here. The claim Fujifilm makes is that under the "normal" handheld use of the GFX100RF, image stabilization isn't necessary (blue area, lower right corner). This is based upon the assumption that people can steadily handhold 1/focal length, which in the case of the GFX100RF would be 1/28 second. Another way of deciphering the chart for the camera they just introduced is that Fujifilm recommends a tripod for anything slower than 1/28 second, handholding for 1/28 or faster shutter speeds.
This is a little disingenuous, as any photo instructor can tell you that there's a large subset of the population that can't (or won't) hold a camera steady at 1/focal length. Fujifilm seems to refer to this group as "Snap" under the section where image based image stabilization (IBIS) works. The GFX100RF doesn't have IBIS, so you "snappers" should probably avoid the camera ;~). Okay, you "snappers" are going to need higher shutter speeds to compensate for your movement. But that's where the f/4 aperture of the lens starts to come into play. The base ISO of the camera is 80, so the Sunny 16 exposure is f/4 at 1/1280. No problems there. Heavy overcast probably puts you at f/4 and 1/80, so you're still probably okay. But if you move into edge of day, night, or less well lit indoors scenes, you're at or beyond that 1/28 limit. But if you're a "snapper" and need two more stops to be "steady," Sunny 16 is 1/320, overcast is 1/20, and indoors and edge of day, well, good luck.
The good news is that the GFX100RF uses a leaf shutter, which doesn't tend to impart additional slap, so it really comes down to how steadily you hold the camera as to whether or not you'll get the full acutance goodness from those 100 megapixels. As I tried to point out, some may find the IBIS-equipped 40mp X100VI a better choice when all is said and done.
But Fujifilm's chart probably raised a few other points you want to know about. First, they show something around 16mm only allowing 1/2 second exposures with IBIS, and that shutter speed gets progressively higher as you increase the focal length. Fujifilm has a "normal" focal length lens showing as only good to about 1/6 second with IBIS engaged, and the range at which they say "IBIS Works" is only 3.5 stops.
Olympus has made claims of 1 second handheld, and we now have many makers claiming 8 stops of correction for their stabilization systems. Is Fujifilm lagging behind, or are you getting gamed?
You're getting gamed.
While there's no doubt that IBIS systems are beneficial, the marketing messages and claims about them are at best case exaggerated, at worst case deceptive.
Let's begin with one thing suggested in that Fujifilm chart: the range at which IBIS is useful shifts as focal length increases. This is absolutely true. You can't violate geometry. (Well, you could, but only by making up pixels where they aren't. See: AI.) Moreover, sensor movement has a limited range, so even if you wanted to try to 100% correct things at 200mm, you might not be able to if the "shake" is too great.
Which brings us to CIPA. CIPA is a marketing organization that all the Japanese camera makers belong to, and which, among other things, assigns standards by which the member companies need to comply. This includes how you measure the size of a product, how you weigh it, and much more. One of CIPA's standards has you strapping your camera/lens to a special platform that is intended to create a repeatable, simulated camera movement. You might have noticed that over time the numbers being reported for that test have gone up. You probably thought that was entirely due to IBIS systems getting better. Well, some of that has happened, but in practice, I'm not seeing the same level of improvement in real world testing. We didn't go from 2.5 stops to 8 stops in a short period. We went from the test rig was pretty crude to the test rig getting better at the simulation and the camera companies getting better at designing to the test rig.
Next up on our list is that Fujifilm claims that IBIS doesn't work above 250mm. This is also related to the geometry, but there's much more going on here that needs explanation. First, marketing will claim that their stabilization system is 5 Axis. (That would be pitch, yaw, roll, horizontal, and vertical.) Well, sort of. IBIS can't change the pitch or yaw of the image sensor (and I believe most can't change their rotation, either, which is roll). In-lens stabilization can't change rotation, and tends not to do horizontal or vertical because then the image circle at the focus plane would be offset.
At some point as you go higher into the telephoto range, pitch and yaw done around the optical center is the best way to correct the geometries of camera/lens movement, and IBIS becomes far less useful. Is the termination and/or crossover of usefulness at 250mm? I do not believe it is. My testing tells me that the crossover tends to occur somewhere in the 100-200mm range and that the usefulness of IBIS just tends to degrade beyond that, but not exactly terminate.
Note that Fujifilm seems to imply that Lens OIS (their form of lens-based stabilization) works all the way down 1 second with 250mm and longer lenses! Sorry guys, but that's simply not even close to true. Just as with the IBIS range Fujifilm shows in the chart, there needs to be a lens-based range, too. (And I'd argue that instead of "tripod" at the left of the lens-based IS range, it should be "gimbal".)
Our final little nugget to unpack is right up at the top of the chart: "New Exposure Program." I suspect Fujifilm is referring to Program exposure mode, that little automatic exposure-setting aid that doesn't always tell you exactly what it's doing. Nikon used to put charts for their programs in their manuals. Shame on them for omitting them these days (and having entire pages that are mostly blank but at the top say something like "The X mode sets X"; yes, you documented something, but not in any useful way). Since I don't have a GFX100RF nor the manual for it, I can't tell you how Fujifilm changed the Program exposure mode. It seems to me that with apertures only a five stop range between f/4 and f/22, to be "fully" useful said program would also need to consider Auto ISO.
Kudos to Fujifilm for trying to explain why they didn't feel they needed to complicate the GFX100RF with sensor-based image stabilization. Some demerits for not getting things clearly explained.
Now, if you think the above doesn't apply to you because you don't have and won't buy a GFX100RF, you weren't reading carefully enough. Image stabilization is one of those much ballyhooed "savior" features that doesn't get explained carefully enough by the camera companies. They all avoid the useful explanation for the simplified marketing message. The days of camera companies providing useful technical white papers about their features and performance seem to be long gone (rest in peace Chuck Westfall). Executives seem unable to explain things. Engineers would rather avoid talking to you. This is no way to run a consumer-facing business.
