I built one once, using a laboratory shake table and some elastic bands. I found that 16 shots for each condition handheld under controlled conditions, analyzed for MTF50, and ensemble statistics computed, gave essentially the same results.
In my book, your anecdotal experience with R5's IBIS cannot be used to support your general claim that IBIS likely adds fine jitter even when working correctly.
FWIW, I have not noticed any issue with IBIS on R5, though I am using it only with RF lenses.
That was meant to say that an IBIS system does some negative things, that are usually ignored in the context of overall benefit, which can be harmful if the corrections really aren't needed. Obviously, if you intentionally shake up a camera while shooting at a low shutter speed, IBIS will definitely help more than it hurts in every axis, but if you are using a fast-enough shutter speed, then any time it twitches it can cause blur, like when it has to correct itself at the end of the range of motion, or if it tries to correct small motions that don't really need correction and does them imperfectly.
While this is a reasonable theoretical argument, and there's no reason not to turn IBIS off (as much as you can turn it off on your particular camera) at high shutter speeds, I have found the effects to be minimal in testing of Sony, Nikon, Hasselblad, Leica, and Fuji cameras.
Did you test for the type of situations that a sports or wildlife photographer has to deal with, where the camera and lens are bouncing all over the place while tracking active subjects or training on new ones that just popped up and may be gone in a fraction of a second?
No to both of those. But it's not been my experience in sports and wildlife photography that the camera is "bouncing all over the place". As to the last, it's been my experience that recentering takes place slowly, and is only an issue for longer exposures with the camera on a triod.
so manufacturers being dumb or lazy pay NO attention in their IBIS implementation to exposure speed and do not adjust IBIS actions for 1/8000 shot vs 1/30 shot ?
You are assuming a lot. I don't think it's an indication of being dumb or lazy, but in the cameras I've tested, the IBIS algorithm doesn't appear to change with shutter speed. I don't think that's necessarily a bad thing. If the goal is to minimize motion blur, that is a worthy goal without regard for shutter speed.
and that sounds strange indeed if they don't adjust IBIS operation at all perfectly knowing in well in advance ( processor/firmware operations wise) before the shot what will be the exposure time and if any material ill effects were detected during their testing ... just thinking - say 1/8000 once focus locked/metering done freeze IBIS compensation and restart after exposure over ... there is always enough time to do it
So you've been assuming all along that the IBIS algorithm changes with shutter speed? If you're the user, why not just turn it off when you don't need it? I am not a fan of unnecessary automation.
I was not assuming anything - in my purely subjective and non-engineering/scientific experience with IBIS with 4 cameras that I used ( from Olympus E-M1 to Canon R5 ) I never had any issues w/ non stab primes ( up to 135mm eq FOV - I am not shooting anything past that tele ) within my range of expectations ( for example I was not really expecting any "8 stops" hand held ) - my subjects have eye irises and eyelashes and so on that you want to have non blurred and for as long as the shutter speed was enough to kill target motion I was ok with the resulting rendering @ 100% pixel=peep ) ...
given that IBIS systems are not new, used by various brands and are in their like 3rd+ generation in some cases my hoi polloi assumption was/is that if any material ill effects with hand held shooting @ "1/8000"-like exposure times are common then something was done to address that by now ...
so I can't blame any blur on IBIS in my own shots when I pixel peep ( I can blame subject motion because I was trying to increase exposure time past reasonable limit or focus errors /operator or sometimes eye/face AF/ or noise because there was not enough sensor saturation or trying to use wider aperture - now those reasons I can clearly see and blame in my own shots ) and I never ever switch IBIS off ...
granted I try to use decent demosaicking software w/ decent I hope corrections ( iridient *transformers and dxo photolab ) that might correct some blur to make things that I need "sharp" sharp enough