Image Stabilization: How IBIS and OIS Interact

[Astigmatism]

Image Stabilization: How IBIS and OIS Interact

I've found it difficult to get clarity on how IBIS and OIS work relative to each other in the Fuji X system, so I spent a while studying it and think I have an understanding now. Here is what I think is going on:

Some camera bodies have In Body Image Stabilization (IBIS) and some don't. Some lenses have Optical Image Stabilization (OIS) and some don't. All combinations of these are compatible.

At any given moment, you either have image stabilization, or you don't. If you're using a body with IBIS and a lens with OIS, you can't select just one or the other.

Lenses that have OIS require a switch to turn it on and off, because they have to be compatible with all bodies even if some bodies have no provision for turning stabilization on or off in their controls. In some cases the switch on the lens is the only way to turn stabilization on or off.

Bodies that have IBIS require a switch (such as a menu item) to turn it on and off, because they have to be compatible with all lenses even if some lenses have no switch.

Therefore, there have to be two places where you can turn stabilization on and off, in combinations of an OIS lens on an IBIS body. Since you could set the two switches to contradict one another, which prevails? The switch on the lens wins (and grays out the menu item in the body, at least in my X-T4).

From the X-T4 manual p146, "The Shooting Menus" > "IS MODE", "The setting selected with the lens image stabilization switch, if any, takes priority over the setting chosen with the IS mode." This is definitive about how the controls work, at least for this camera body, though opaque regarding what's going on inside.

When you use an OIS lens on an IBIS body and have stabilization turned on, how do the OIS and IBIS subsystems interact? We users don't actually have to know, all we have to know is that stabilization is either on or off at any given moment. But, still, it's interesting, isn't it? Mechanically, OIS and IBIS have some different advantages.

For longer and longer lenses, the image jumps around much more with the same amount of camera shake, which would require more and more physical movement range for the IBIS to accommodate it, and any given IBIS mechanism would have some lens length limit to what it could manage, which is not good for compatibility (especially considering future lenses not even designed yet). If you want to keep the body small and light, you don't want a huge IBIS mechanism. This all amounts to a big incentive to build OIS into longer lenses. Besides, they tend to be big and heavy and cost a lot, so there's more design opportunity to manage shake there without driving up camera body prices and sizes and weights.

But for all lenses, IBIS provides more degrees of freedom to compensate for more types of shaking motion. I think OIS has two degrees of freedom and IBIS has five (not 100% certain on these numbers). There are certainly some motions for which IBIS could be better.

A good guess is that Fuji takes advantage of both IBIS and OIS when both are available. Certainly some other manufacturers do, and it only takes software to do so. There are clues online. I have an X-T4 and an X-T30 II, and I could do an experiment to see whether OIS lenses yield better image stabilization on the X-T4 than they do on the X-T30 II, maybe I will. Fuji wouldn't have to announce to the world exactly what they're doing inside with OIS and IBIS.

Last point: you can choose to have stabilization constantly or only when you shoot (which seems to include when you have the shutter button half pressed). It's your battery, use it when you want to. This doesn't change the fact that at any given moment, you either have image stabilization, or you don't.

So, that's what I think.

Now for some miscellaneous quotes online that provide hints (I didn't write these, I found them):

 

********************************************
********************************************
********************************************

 

OIS and IBIS work together  OIS does as much as it can and IBIS adds a bit. [this was about the X-H1]

I can't find the link, but it's been reported since the announcement of the XH-1 that the IBIS and OIS work differently for different lenses.  Obviously it's IBIS only for primes (and the 16-55) without OIS.  For the shorter zooms with OIS, if I recall correctly,  certain axis rely on IBIS and others the OIS.  And for the longer zooms and the 80mm macro, the system defaults entirely to the OIS in the lens. That may be correct, but we will have to wait for a manual with confirmation to be sure. There is a certain logic in that advice.  Canon & Nikon have long argued that OIS works better than IBIS for long lenses because the corrections are larger for longer lenses and the OIS can be tailored to the precise requirements of their FL and design.  The usual counter argument is that they would say that anyway because they didn't invest in IBIS. The reported default to OIS for the longer lenses also raises a question as to whether there is a maximum FL for which the Fuji IBIS is effective.

IBIS is synchronized with the OIS in OIS lenses and they both are used in conjunction.

Some modern camera systems now allow combining OIS lenses with camera bodies that have IBIS, and the combined stabilization effect is increased beyond what just one or the other would be capable of. If you have an OIS + IBIS capable system then you have the best of both worlds and you don’t need to worry about which one is better.

If your lens has a OIS button, that will switch off both OIS and IBIS. All OIS/IBIS switches act on both for that matter: you can't have OIS without IBIS, or IBIS without OIS.

I've had a reply from Fujifilm UK on this now, and it confirms the last post. They state "The OIS switch on any lens 'overrides' the IBIS setting in the camera.  If the camera is set to continuous and then the OIS switch on the lens is set to 'off', the IBIS switches off as well."

Like on many other cameras, IBIS and OIS seem to be coupled: they are either both on or off. For 18-55, Fuji lists 4 stops of stabilization on cameras without IBIS and 6.5 stops when using it on X-T4.

Fujirumors has published a slide (2/26) listing the measured number of stabilization stops for each lens when used with X-T4 (and X-H1). 18-55 is listed to have 6.5 stops of stabilization when used with T4's IBIS vs 4 stops when using OIS alone. As expected, IBIS seems to help.

[Note the significance of these last two quotes; the 18-55 zoom has OIS, and it gives 4 stops by itself but IBIS contributes a further 2.5 stops. However this doesn't necessarily mean the OIS is still operating. If we compared this OIS zoom with a non-OIS lens on this body with IBIS on, and saw less than 6.5 stops of stabilization, I think that'd prove it, for the X-T4 anyway.]

IBIS and OIS work together for Fujifilm lenses - you can't use IBIS without OIS (for an OIS lens).

I asked Fuji service Europe about IBIS/OIS: Has OIS priority with longer lenses because it is more effective for longer focal lengths while IBIS has priority with shorter focal lengths? They answered: >> in practise you don't have the option to choose between OIS and IBIS. That means that both options work always together. The fact is when you use lenses with longer focal length the OIS has priority as of the reason you mentioned. This means the stabilisation measured in EV steps will not be so effective as the use of lenses with wider focal length.<<

With non-stabilised lenses you have the option to turn IBIS off from the same menu, but when you use a lens with OIS built-in, the option to turn off IBIS disappears. In this instance, image stabilisation is controlled from the OIS switch on the lens.

Can you use both in-lens stabilization and IBIS? In short, yes. While it is dependent on the camera system you’re using (for example, Panasonic has a list of compatible lenses), but you should be able to use them together. With Sony systems, activating both systems will delegate 3-axis stabilization to the IBIS, and leave the pitch/yaw adjustments for the Optical Steady Shot (O.S.S) in-lens stabilization. Fuji systems, at least the Fujifilm X-H1, work in a similar fashion; delegating specific axis’ to different systems to achieve the standard 5-axis stabilization.
 

[Astigmatism]

I got a bit more clear, through experiments and the manuals.

If there's OIS with the lens hardware switch ON, or IBIS, or both, the menu just controls whether stabilization is CONTINUOUS or SHOOTING ONLY, and also maybe has a shutter speed limiting option for automatic exposures. For a body with IBIS plus a lens without OIS, the menu additionally has an OFF. If there's an OIS lens with its switch off, on any body, the menu is grayed out.

Looking through both manuals, I see no mention of this dependence system. They read read as though all menu items are always available. However, there not; it's designed so that there's no way you can set controls to malfunction.

The X-T4 has IBIS. My X-T4 manual doesn't make this clear, but: With one exception, there's just a single thing the "IS MODE" menu item controls: does the image stabilization run the whole time the camera is turned on (CONTINUOUS), or does it only run while the shutter button is at least halfway down (SHOOTING ONLY)? Nothing about OIS or IBIS specifically. The one exception is that when shooting with a non-OIS lens, which has no OIS switch, an OFF option appears for turning off IBIS.

The X-T30 II has no IBIS. My X-T30 II manual is similar to the X-T4 but with an added complication: there are "+ MOTION" options that shorten shutter speed when motion is detected in some auto exposure modes (I'm unfamiliar with them). The manual shows an OFF option and I don't know why, unless there are OIS lenses without switches. Anybody know why? With my several lenses on this camera, either no OIS or OIS with a switch, I played around some and never got an available OFF option. The menu depends on whether there's an OIS lens and how its switch is set, and also whether the whole camera is set to AUTO.

----- Another useful item I found online ---------

From FujiFilm Tags at exiftool.org:
Tag: ID0x1422    
Tag Name: ImageStabilization
Type: int16u[3]    
[Value 0]
0 = None 
1 = Optical 
2 = Sensor-shift          
3 = OIS Lens 
258 = IBIS/OIS + DIS 
512 = Digital
[Value 1] 
0 = Off 
1 = On (mode 1, continuous) 
2 = On (mode 2, shooting only)

[jjijack]

On 1/17/2023 at 4:53 AM, Astigmatism said:

The X-T30 II has no IBIS. My X-T30 II manual is similar to the X-T4 but with an added complication: there are "+ MOTION" options that shorten shutter speed when motion is detected in some auto exposure modes (I'm unfamiliar with them). The manual shows an OFF option and I don't know why, unless there are OIS lenses without switches

The “+ MOTION” option exists only when auto ISO is enabled and shutter speed is not fixed. For OIS lenses I’ve only used 16-80  and 70-300, both of which have no switches so I can set IBIS and OIS off in the manual.

On 1/7/2023 at 11:21 PM, Astigmatism said:

If we compared this OIS zoom with a non-OIS lens on this body with IBIS on, and saw less than 6.5 stops of stabilization, I think that'd prove it, for the X-T4 anyway.

Besides, have you ever tested it? I’m pretty curious about the result. I own XS-10 and 16-80 for the time being. The former possesses great IBIS ability, whereas the latter possesses great OIS ability. If OIS does little help on an IBIS body, I’d rather change the lens into something like tamron 17-70.

[Astigmatism]

On 3/7/2023 at 7:23 AM, jjijack said:

Besides, have you ever tested it?

I thought for a while about how to do it, but it's complicated. If there's a huge difference it would become apparent, but if at all subtle no. I thought about how to reproduce a level of camera shake, and how to measure it in an image. Any thoughts? Do you have any lenses already that would let you test this? It sounds like you wouldn't, in any case, have that Tamron and the Fuji 16-80 simultaneously, so you'd have to try doing tests on whichever you have first that will turn out to pick up the difference when you do the same tests on whichever you have second.

[Mr. Howdy]

Thanks for taking the time to compile that explanation of Fuji's IBIS/OIS situation. Pardon the following lengthy backstory here, but it serves a relevant point.

As a hybrid shooter, I came to Fuji via the X-H1 (Fuji's first body with IBIS), from the Panasonic G9, because I much preferred Fuji's colors and dynamic range, on which Panasonic's M43 cameras were weak. What they were strong on, however, is/was IBIS, and the Lumix Dual IS (IBIS/OIS conjunction) I found wonderful, from its Steady Shot locked-on static shot to its ability to pan without jarring and stuttering. With the X-H1, I was, by comparison, disappointed with its IS situation. When using non-OIS lenses in shooting video, the IBIS was nice for hand-holding, and could sometimes handle panning and motion, such as following a car on a street. With OIS lenses, however, I had to forget about panning and tilting video shots. It was dismaying that I couldn't select "either/or/and" for OIS and IBIS. For stills shooting, however, it was great to be able to hand-hold at slow speeds. Since the X-H1 was also the first Fuji body to feather the quiet, near-shockless electro-magnetic shutter, which also encouraged my transition from Panasonic (whose cameras since the GX85 have had a similar shutter mechanism), there was less for the IBIS to have to stabilize than with the "older" Fuji shutter.

I got the X-T4 because of its switch-segregated Still and Movie modes (almost/not-quite implemented on the X-H1), its continued use of the electro-magnetic shutter, the larger-capacity battery, its being the first Fuji to write video files redundantly to both SD cards, and the body style of the X-T2, which I also owned and which I preferred. Adding the SmallRig rubber grip approximated the grip of the X-H1, when I'd want it. I resented having to have the "flippy" screen – because I don't shoot video of myself, I shoot video no differently than I shoot stills, and I've seen too many LCD screens on Canon cameras broken-off at the swivel-joint – but for everything else about the X-T4, I considered it a necessary nuisance.

Along with all this about the X-T4 I expected to come improved IBIS, as Fuji would surely have ironed-out the glaring kink in it. But no, it was about the same; maybe-maybe improved, but still an impediment to non-static video shots. I also wondered why Fuji wouldn't enable the user to de-couple and selected between OIS and IBIS, which would seem possible within firmware. I also began to notice that with telephoto zoom lenses with OIS, the IBIS would struggle against it, resulting in mis-framed shots, both in stills and in video. My solution to this became to shoot still images with a telephoto zoom lens (most recently, the 70-300) on an X-T3, which doubles for the X-T4 in my hands and serves as back-up if the X-T4 goes down.

Another thing about the OIS/IBIS combo is that, when I was shooting with the X-T4 with the 16-80mm zoom, from a moving ferry boat, fixated on an island in the middle distance on the water, the rapidly-moving water in the lower third of the frame caused the IS to jump and jitter, as it was confusing the IS. If there were some selectable panning mode to the IS, as on Panasonic cameras, this might not have occurred.

In video, the "+DIS" IS mode on the X-T4 does help with static shots, with a negligible crop penalty, especially in windy conditions. However, I've found that usually I can shoot hand-held static shots without it added, and stabilize them in DaVinci Resolve, resulting in shots that look as though I'd used a tripod. I just have to factor-in a bit of a crop, but I tend to shoot too wide anyway (with the 16-80mm on the X-T4 80% of the time).

Fuji's OIS itself is great for stills shooting, but the IBIS is far greater for video. I wish that Fuji would consider the issues such as I've described (and I'm hardly the only person who has mentioned them online) and remedy them with some firmware updates (going back to the X-H1, if possible), although at this point, they would probably reserve that development for future bodies as an inducement to buy the new models. Fuji should realize that "video" doesn't mean "vlogging," and that not only do video images move within the frame, but that the frame itself also moves.

[Astigmatism]

3 minutes ago, Mr. Howdy said:

the rapidly-moving water in the lower third of the frame caused the IS to jump and jitter, as it was confusing the IS

Pardon the lengthy backstory? Nay, I appreciate it all!

About the rapidly moving water confusing the IS, I had the strong impression that IS is based entirely on sensing accelerations around 3 axes. I didn't think it used any image information at all. This would be why, when using a non-Fuji lens that doesn't communicate its focal length, you have to explicitly tell the camera what the focal length is, to set the scaling between angular acceleration and counteracting motion in the IS mechanism.

How sure are you that the camera pays attention to motion in the image when doing stabilization? Anybody?

[Mr. Howdy]

On 4/5/2023 at 6:19 AM, Astigmatism said:

About the rapidly moving water confusing the IS, I had the strong impression that IS is based entirely on sensing accelerations around 3 axes. I didn't think it used any image information at all. This would be why, when using a non-Fuji lens that doesn't communicate its focal length, you have to explicitly tell the camera what the focal length is, to set the scaling between angular acceleration and counteracting motion in the IS mechanism.

How sure are you that the camera pays attention to motion in the image when doing stabilization? Anybody?

As I'd said, I was using the XF 16-80 when I'd had the problem caused by the water in the lower third of the frame. What IS of either sort does is read motion, in order to stabilize the image. It was obviously the water rushing by in the lower third of the frame that had caused the OIS/IBIS combo to freak-out. I've read reports by others that this occurs when there's too much motion in the frame.

The setting for focal length of non-Fuji (prime, manual-focus, usually) is not only for the sake of the IBIS but also for that of the focus peaking and such, which is why non-IBIS bodies have it. However, the IBIS definitely is influenced by the setting. I've even seen it advised to set a telephoto lens to more than its focal length in order to better stabilize it, though I've never tried this.

[jerryy]

Would you post an image from the session showing the effect you describe?