R6 Mk ll - Odd night sky image problem UPDATED 7-19-24

[justadude]

Updated, but not solved.  The update will be here at the top in italics.  Everything below italics is the original post.

I have been in contact with Canon Tech Support - first by email.  Three different technicians, a lot of back and forth questions asking pretty much the same thing that is scattered through this post.  No one had any suggestions, but just a lot of questions on what I could have been doing wrong, none that panned out though.  Finally did the phone option, and got a little further.  This tech let me send in RAW files, could actually see what the concern was.  He was also an astro-photographer, so when I explained various things, he understood what I was talking about.  He also confirmed that I was not crazy... there is zero EXIF data for IBIS, only IS through the lens, which did not apply here.  So again, we do not know what caused this, or what to do. 

I asked "Is this the type of movement one would see IF the IBIS was still on, while on a tripod?"  He replied there would be movement, but he cannot find anything saying it would be a small rotational type of movement.  So now we are both trying to do some homework to find out.

This brings up a question for anyone here...  Do any of you have any idea what type of movement that YOU get in your camera if you leave IBIS on (with a lens that does not have IS) while on a tripod?  Out of all of my cameras, this is the only one with IBIS, so I cannot test this out myself since the technician and I agreed this would need to be tested with a different camera.  I'm stuck and at a point where I have zero other resources to figure this out.

Thanks!

I have a question on some night sky images I have been taking with the R6 Mk ll.  I should mention first that I am not new to night sky photography.  I've been doing it since pre-digital days.  I've also shot a lot of night sky images with 4 different Pentax DSLRs, 3 different Canon DSLRs, and now 3 different Canon Mirrorless bodies.  The R6 Mk ll is the first camera I've experienced the following with.

If you look at the two attached images, one is nice and clear, the other seems to have a bit of a rotation to the whole image.  I've only shot night sky images a few times with this camera.  Last month I saw about 10% of my images had this weird rotation.  I wrote it off as shifting sand around the tripod.  Last night I was out taking about 100 shots.  Very sporadic, but roughly half of them look like this.  This last one was taken in the parking lot (solid ground - no shifting sand).

Few notes:
* Camera set to Manual mode
* IS turned off in camera
* The tripod is solid, tight connections on legs and ball head, and far from being underrated for this gear weight.  
* I've used two different lenses where this is happening.  Rokinon 14mm F/2.8 EF mount... Venus Optics Laowa 15mm F/2 RF mount.  Both of these lenses are full manual - no auto setting, so the camera is set to shoot without a lens.

Am I missing something in the settings that is new to this specific camera?  I've never had this happen on any other camera.  Also, I only see this in night sky shots, when on a tripod.  Day shots, on a tripod are never a problem.

[justadude]

Plenty of good information there, John.  Thank you!

It does make sense that the IBIS is trying to compensate even on non-IS lenses.  But what we are trying to find out is what type of motion will you see on a tripod when this is happening.  All of the images I had that were bad showed the exact same rotation/motion.  Is that what others are seeing under these conditions?  Or is it random types of movements?  That is what we need to determine.  

If others are seeing the same pattern of movement and rotation, then we know for sure the IBIS has been kicking on, even though the menu settings had IBIS set to off.  If that is what is happening, then I can send the camera in for warranty work.

I have a question on one of your comments.  You said "The hardware and electronics for IBIS must always be active to keep the sensor positioned."  True when it's on, but that should not be happening when in the menu you turn IBIS off, correct?

[johnrmoyer]

---

@justadude wrote:

Plenty of good information there, John.  Thank you!

It does make sense that the IBIS is trying to compensate even on non-IS lenses.  But what we are trying to find out is what type of motion will you see on a tripod when this is happening.  All of the images I had that were bad showed the exact same rotation/motion.  Is that what others are seeing under these conditions?  Or is it random types of movements?  That is what we need to determine.  

If others are seeing the same pattern of movement and rotation, then we know for sure the IBIS has been kicking on, even though the menu settings had IBIS set to off.  If that is what is happening, then I can send the camera in for warranty work.

I have a question on one of your comments.  You said "The hardware and electronics for IBIS must always be active to keep the sensor positioned."  True when it's on, but that should not be happening when in the menu you turn IBIS off, correct?

---

My guesses are based upon what little I know about IBIS. It seems to me there must be a mechanism to park the sensor on power off so that it will not be damaged by camera shaking similar to parking auto focus mechanism in a lens. I guess IBIS must be powered on to keep the sensor in place even when IBIS is not being used to stabilize an image. I guess that the parked power off position of the sensor is not suitable for making photos. I have read that some camera manufactures leave the sensor loose to flop around when powered down, but it would seem to me that parking the sensor would be better.

The trick seems to be to select the combination of camera settings that does not trigger firmware bugs that impact the photo you are making. This seems to me to vary from one firmware version to another. The firmware in these cameras is very complex and when one problem is fixed, it might cause another thing to behave differently as well.

I guess, but do not know, that a strong magnetic field could move the sensor since magnets are used to position it.

https://www.dpreview.com/news/1854379286/canon-patent-application-details-what-canon-has-in-store-fo... 

"

https://www.canonrumors.com/patent-here-is-canons-ibis-unit/ 

The Halbach array, believed to have first been discovered by John C. Mallinson in 1973, is a collection of magnets that is particularly arranged so that one side of the magnetic field is magnified while the opposite side is effectively canceled out. Halbach arrays have multiple uses ranging from something as simple as a refrigerator magnet to something as intricate as a particle accelerator (where it’s used to focus particle accelerator beams).

Canon’s implementation, however, would use Halbach arrays to ensure that when a correction is applied to one axis, it won’t negatively affect another axis. Particularly, Canon’s patent application details how it would use a Halbach array on the vertical (y-axis) stabilization unit to ensure that the horizontal correction (x-axis) isn’t skewed when applying y-axis corrections.

"

https://www.canonrumors.com/canon-patent-application-ibis-improvements/ 

"

In this patent application (2024-071858) Canon looks to improve the operation of IBIS on your cameras near and dear to your heart. While we'll never know if this patent will end up in the R1 or the R5 Mark II, all these types of patents silently make the little improvements that ultimately add up to big improvements upon release.

The general concept of this patent application is to compensate for the small vibrations from the IBIS unit from the camera's movement sensor. I could see this type of movement compounding using IBIS during the video, as it could cause undesirable effects if the control logic were not written correctly. This patent was written with the background of a shutter movement, so I suspect Canon is looking at this to improve the IBIS functionality during stills, but any improvement for stills should also carry over to video.

How Canon breaks down this patent application is as follows. This is machine-translated Japanese AND a patent application, so the wording is awkward. Canon mentions the background of this problem is with hybrid IS + IBIS camera operation as mentioned in the forward of the patent application. Interesting to note that the Camera knows the vibrations caused by the shutter and how to handle them specifically.

As imaging devices have become more sophisticated, many imaging devices and photographing lenses are equipped with a blur correction mechanism. The blur correction mechanism enables users to reduce the effect of blur on a photographed image when using an imaging device to take a photograph. Several types of blur correction mechanisms have been proposed for use in imaging devices, including a method of performing blur correction by driving a part of a lens in a photographing optical system and a method of performing blur correction by driving an image sensor. In lens-interchangeable imaging devices, the former is a method of performing blur correction by driving a part of a lens in a photographing optical system in an interchangeable lens barrel, and the latter is a method of performing blur correction by driving an image sensor in a camera body. A method of performing blur correction by combining the two and driving both a part of a lens in a photographing optical system and an image sensor is also known. In imaging devices using such a blur correction mechanism, there is a problem in that vibrations generated by driving a drive unit in the imaging device, such as a shutter are transmitted to a blur detection means provided to detect camera shake, and the blur correction mechanism operates regardless of camera shake.

In response to such problems, a technique is known in which the signal processing of the shake detection means is changed while the drive unit in the imaging device is operating, thereby reducing the effects of vibration.

But Canon goes on to say, that this type of compensation is not done for vibrations that occur because of the IBIS Unit.

However, [it] does not mention that vibrations generated by the movement of the shake correction mechanism may act on the shake detection means.

On another note, isn't translated Japanese fun?

We normally don't see and appreciate these efforts, but it makes using the cameras simpler and less error-prone. Canon's research is never fully completed and is an ongoing effort. So while none of this may make it into an R1 or an R5 Mark II, any improvements no matter how insignificant, the better we are for it.

Source: Japan Patent Application 2024-071858

"

https://www.canon-europe.com/pro/stories/8-stops-image-stabilization/ 

A firmware update for the EOS R5, EOS R6 and EOS R3 also brings improved Movie Digital IS performance when using wide-angle lenses. "IBIS allows the sensor to pivot in order to correct vibrations, so the corners of the image move backwards and forwards," explains Mike. "In some situations, the edges can appear more wobbly than the centre, but the digital stabilisation firmware update will correct that movement and give a more stable image."

[Waddizzle]

I now find this post very confusing. Your well intentioned “update” has broken the continuity of the discussion. 

[justadude]

My apologies. I was going to add an update way at the end as a comment, but figured no one is going to read 4 pages of comments to see an update from the author.  Should I delete the post?  

[johnrmoyer]

Assuming for he moment that some of my guesses might be correct, IBIS without tracking could produce circular star trails if the camera cannot make a photo while the sensor is parked.

https://en.wikipedia.org/wiki/Image_stabilization
"

This is limited by the rotational movement of the surface of the Earth, that fools the accelerometers of the camera. Therefore, depending on the angle of view, the maximum exposure time should not exceed 1⁄3 second for long telephoto shots (with a 35 mm equivalent focal length of 800 millimeters) and a little more than ten seconds for wide angle shots (with a 35 mm equivalent focal length of 24 millimeters), if the movement of the Earth is not taken into consideration by the image stabilization process.

"

https://snapshot.canon-asia.com/article/eng/what-you-need-to-know-about-in-body-is

[justadude]

So if I am reading this correctly, even with IBIS turned off, I should not use more than a 10 second exposure on my 15mm lens?  Even though I should (and have with many cameras - even with this camera) made 30 second exposures without any rotation of the sensor... which is totally separate (and a lot different) than star trails?  This also makes me wonder why with 4 second exposures I'm getting this sensor rotation - which again, isn't even close to what a star trail rotation would be since I was pointing to the wrong part of the sky.  

[johnrmoyer]

---

@justadude wrote:

So if I am reading this correctly, even with IBIS turned off, I should not use more than a 10 second exposure on my 15mm lens?  Even though I should (and have with many cameras - even with this camera) made 30 second exposures without any rotation of the sensor... which is totally separate (and a lot different) than star trails?  This also makes me wonder why with 4 second exposures I'm getting this sensor rotation - which again, isn't even close to what a star trail rotation would be since I was pointing to the wrong part of the sky.  

---

What I was trying to point out is that it seems to me that there is a firmware bug when using this camera with a lens that has no electronics and with IBIS disabled.  I am suggesting that the IBIS might try to adjust for the rotation of the earth instead keeping the sensor stationary because of a firmware bug.

If your camera slept between shots, it has been my experience that with some firmware versions on the EOS R5 the IBIS functioned differently after wake from sleep than after power on. Cycling the power switch fixed that problem for me, or disabling auto power off fixed it until the fixed firmware was released. I guess you might have similar problem because first image was good and second image was bad. Also, on the EOS R5 the Fv mode and nothing auto seems to me to behave differently than M mode. If using M mode, one might try Fv mode with none of the settings auto.

I suggest that to work around the firmware bug, enable IBIS, maybe just for shot. Enable tracking while an auto-focus lens is attached and leave it enabled and before mounting the lens with no electronics. Enable the focus peaking assist features even if not planning to use them. Set the lens focal length for IBIS in the camera menu.

I am making these guesses because the IBIS uses the image from the sensor to position the sensor. If tracking is enabled, it might try to keep the same star on the same sensor pixel.

If you try this and it does not work, then my guesses were wrong and it must be a different kind of firmware bug.