Are you experiencing this phenomenon with IBIS on your Canon EOS R5?

Aug 27, 2019
667
1,414
Summary of the issue as I see it.
  • Occurs when using lenses without IS
  • Only the first shot is affected.(Powering off or image review appears to reset the first shot so it will occur again. )
  • Does not appear to be an issue when using a tripod.
  • Does not appear to be an issue when using lenses with IS.
  • Does not appear to be an issue if you set IBIS to be on only when taking a shot.
I have reported the repo steps to CPS Canada. Nice find @juanmaasecas.

If you are using a lens without IS be sure to check those shots for critical focus and be sure to report the issue to Canon. I do not use my RF50 much so not something I am to worried about but the fact that the issue likely affects the 28-70 and the fast primes Canon needs to address this.
 
  • Like
Reactions: 2 users
Upvote 0
You mentioned earlier that switching to 'Only for Shot' mode solved the problem. If so, I am curious as to why that is the case. I believe the the 'Only for Shot' mode is only available when using a non-IS lens, otherwise, it is blanked out for Canon's IS lenses. This is what I understand about the IBIS. When set to 'Always On', you can hear the whirring sound in the camera if you put your ear close to it. In the 'Only for shot' mode, you can only hear the whirring sound if the shutter is half pressed. This seems to me that IBIS only kicks in when the shutter is half pressed in the 'only for shot' mode, i.e. it needs some time to be activated before the shot is taken, whereas, it is already on in the 'always on' mode. If this is true, I wonder why activating it later in the 'only for shot' mode would make the twisting problem go away, as the difference in the two modes seems to be a matter of when the IBIS kicks in. Could it be that in the 'always on' mode, the IBIS is essentially 'sleeping', and hence has to figure out the amount of compensation for the first shot, which it doesn't always get right, and hence the 'twisting'? And that in the 'only for shot' mode, the activation of the IBIS when the shutter is half pressed allows the IBIS to 'wake up' sufficiently to calculate the necessary compensation and hence there is no 'twisting'? Wonder what you think.
At least in my camera, when ibis is set to only for shot, it is enabled only for shot, it is still disabled during focusing (half pressed button).

with Sony it works like that and it’s how it should work with Canon as well, engaging ibis only during af and shooting. With Canon you need to chose between having it on all the time (draining battery) or only when the picture is taken (not having a stable image during af).
 
  • Like
Reactions: 1 user
Upvote 0

f119a

My Cars Always Break Down
CR Pro
Jun 18, 2013
48
14
Delaware
www.yichenhu.com
You can see the twist after the liveview feedback is back after taking the picture. The twist seems to happen at the end of the exposure, and still ongoing when the screen is back on.
Quite a few cameras' IBIS systems are prone to overcompensating when working, creating an effect like using tilted/de-centered lenses. I don't think sony cameras are doing this but they got their own problems. I just tested on my R5 and it seems ok at the moment, but I know it did similar things before (and I don't think that's a real problem). Don't forget your second press could induce vibration as well.

Is it a bug? Maybe.
Is it still saving more pics than having no IBIS? Definitely yes based on my experience, especially on higher MP cameras like GFX100.
 
Last edited:
Upvote 0
Summary of the issue as I see it.
  • Occurs when using lenses without IS
  • Only the first shot is affected.(Powering off or image review appears to reset the first shot so it will occur again. )
  • Does not appear to be an issue when using a tripod.
  • Does not appear to be an issue when using lenses with IS.
  • Does not appear to be an issue if you set IBIS to be on only when taking a shot.
I have reported the repo steps to CPS Canada. Nice find @juanmaasecas.

If you are using a lens without IS be sure to check those shots for critical focus and be sure to report the issue to Canon. I do not use my RF50 much so not something I am to worried about but the fact that the issue likely affects the 28-70 and the fast primes Canon needs to address this.
You don’t need to even turn the camera off nor review images. It will happen again as soon as you release the shutter button and try to take a picture again. The problem is the first picture in a series without fully releasing your finger from the shutter button (for example a burst, or many pictures taken consecutively while keeping half pressed the shutter button).

it’s not happening absolutely 100% of the time, but most of the time.

if you try to shoot a burst (remember, slow shutter speeds, try for example 1/6-1/15s) you will see how the twist in the sensor that happens at the end of the first exposure is also affecting a bit the second picture, then the third one is perfect.
I am convinced there is some bad code in the ibis software…
 
  • Like
Reactions: 1 users
Upvote 0
Aug 27, 2019
667
1,414
You don’t need to even turn the camera off nor review images. It will happen again as soon as you release the shutter button and try to take a picture again. The problem is the first picture in a series without fully releasing your finger from the shutter button (for example a burst, or many pictures taken consecutively while keeping half pressed the shutter button).

it’s not happening absolutely 100% of the time, but most of the time.

if you try to shoot a burst (remember, slow shutter speeds, try for example 1/6-1/15s) you will see how the twist in the sensor that happens at the end of the first exposure is also affecting a bit the second picture, then the third one is perfect.
I am convinced there is some bad code in the ibis software…
I agree with you on the bad code.

If I had to guess I think there is a call being made to the lens IS system that does not get a response because there is no IS system. Those lines of code are then skipped on the second call and the system reacts like it should when relying on IBIS alone. In simple terms there is a missing if statement.

I see this all the time in or FPGA's on our Cameras.(work for Avigilon) The code is reused on multiple platforms but not all features are present in each Camera so we have to be sure to flash the correct firmware or the thing will get hung up trying to activate an IR array that does not exist.
 
  • Like
Reactions: 4 users
Upvote 0
Apr 1, 2016
348
321
Summary of the issue as I see it.
  • Occurs when using lenses without IS
  • Only the first shot is affected.(Powering off or image review appears to reset the first shot so it will occur again. )
  • Does not appear to be an issue when using a tripod.
  • Does not appear to be an issue when using lenses with IS.
  • Does not appear to be an issue if you set IBIS to be on only when taking a shot.
I have reported the repo steps to CPS Canada. Nice find @juanmaasecas.

If you are using a lens without IS be sure to check those shots for critical focus and be sure to report the issue to Canon. I do not use my RF50 much so not something I am to worried about but the fact that the issue likely affects the 28-70 and the fast primes Canon needs to address this.
Yep, can confirm both my R5 bodies show the issues under the conditions you mention. Will report to my local CPS department.
 
Upvote 0

rbr

Sep 11, 2010
129
64
Summary of the issue as I see it.
  • Occurs when using lenses without IS
  • Only the first shot is affected.(Powering off or image review appears to reset the first shot so it will occur again. )
  • Does not appear to be an issue when using a tripod.
  • Does not appear to be an issue when using lenses with IS.
  • Does not appear to be an issue if you set IBIS to be on only when taking a shot.
I have reported the repo steps to CPS Canada. Nice find @juanmaasecas.

If you are using a lens without IS be sure to check those shots for critical focus and be sure to report the issue to Canon. I do not use my RF50 much so not something I am to worried about but the fact that the issue likely affects the 28-70 and the fast primes Canon needs to address this.
I have seen the problem occur when the IBIS is set to "only for shot". I don't think that makes a difference. I do think the shutter speed matters. I don't see it at shutter speeds faster than 1/100 sec. or so.
 
  • Like
Reactions: 1 user
Upvote 0
Yes, definitely noticed the twist on my R5, with IS always on, for non IS lenses, since day 1 on 1.1 firmware even up to current 1.3.1 firmware.

I've noticed another annoying issue as well: when IS on but for "shot only", it often blurs my first image when it shouldn't.

Also noticed, as another person has mentioned, that the corners of my pics get progressively blurrier away from the center in some shots. I think I've noticed it happen with and without IS on. I use the same lenses on my 5D3 and it doesn't have the same issue. (Not a resolution issue because R5 shots can look significantly blurrier even when resized to match 5D3 resolution.
 
  • Like
Reactions: 1 user
Upvote 0

Toglife_Anthony

Hit the G.A.S. & pump the brakes at the same time!
Apr 2, 2020
64
80
Use a lens without OIs. NO optical stabilisation.
Set IBIS to Always ON
Go to shutter priority mode (for example)
set 1/10s shutter speed
Use af servo (also happens in one shot, but it will be easier for you to discard other issues)
single shot mode (you can try also burst, it also happens, but then people will say that the problem is pressing the button in the first shot...)
Focus at something far to better see details

As you are in servo, focus in the center point (for example) and keep the camera focusing during a couple of seconds
After that, very carefully fully press the shutter button to take the picture and, WITHOUT RELEASING THE SHUTTER, go back to keep focusing in the same spot.
Then, very carefully take another picture, and after back go back to half pressing
Take another picture, etc

You will see how the first picture is blurred 90% of the time (look at the edges and corners, the center is usually sharp of course), and the successive pictures are sharp. You can see how when the live view feedback in the LCD screen is restored, there is a visible twist that does not happen in the rest of the pictures.

edit: Use either EFCS or electronic shutter.
Following these instructions I was able to reproduce the problem. Side note, I used mechanical shutter as I forgot to switch it, and I still had the issue.
 
  • Like
Reactions: 1 user
Upvote 0
Following these instructions I was able to reproduce the problem. Side note, I used mechanical shutter as I forgot to switch it, and I still had the issue.
It also happens in mechanical indeed, but I was trying to narrow the chances of people saying it is user error (shutter shock in this case).
 
Upvote 0
I did some testing with the RF 50mm f/1.2 at 1/10th of a second with IBIS "Always On".


Honestly I couldn't tell any difference in the corners between the first shot or the second or the last. They all look good to me. I changed it to "shot only" and the images look darn near identical to me.


I did do some testing on and off of the tripod. I tested using the shutter button itself as well as a bluetooth shutter.


One thing I did manage to do once on the tripod was after every shot I could visibly see the image on the display twisting clockwise. It wasn't just the first shot - but every one. I tried to record it with my phone but it's subtle enough that you see it if you're looking for it but it's hard to record and demonstrate.

I couldn't see anything wrong in the corners of the images when zoomed in as far as I could go. I was shooting at f/9 for most of my testing.

I went into the settings and changed from 1st Curtain Electronic shutter to full electronic and the issue disappeared. I changed it back to 1st Curtain Electronic and the issue did not recur.


I do think that there may be some tripod detection logic going on but I have no way of proving it. I turned the camera off and turned it back on and could not immediately reproduce the issue.


Either way - I wasn't able to find anything that I couldn't blame on user error [like me hand-holding the camera] and I did follow the directions I found here as well as a couple of other places.

My testing is by no means exhaustive but if I tried this hard to do it on purpose and couldn't I have to wonder how much if at all it'll affect me in the real world. It's something I'll try to keep an eye out for.

As a programmer this sounds like bad logic to me. It sounds like either a wrong branch is being taken or perhaps a situation was overlooked.


Overall thankfully I'm still really happy with my camera and on the upside most of the newer lenses seem to come with IS built in and the IBIS seems to be working great with Lens IS.
 
  • Like
Reactions: 1 user
Upvote 0
I did some testing with the RF 50mm f/1.2 at 1/10th of a second with IBIS "Always On".


Honestly I couldn't tell any difference in the corners between the first shot or the second or the last. They all look good to me. I changed it to "shot only" and the images look darn near identical to me.


I did do some testing on and off of the tripod. I tested using the shutter button itself as well as a bluetooth shutter.


One thing I did manage to do once on the tripod was after every shot I could visibly see the image on the display twisting clockwise. It wasn't just the first shot - but every one. I tried to record it with my phone but it's subtle enough that you see it if you're looking for it but it's hard to record and demonstrate.

I couldn't see anything wrong in the corners of the images when zoomed in as far as I could go. I was shooting at f/9 for most of my testing.

I went into the settings and changed from 1st Curtain Electronic shutter to full electronic and the issue disappeared. I changed it back to 1st Curtain Electronic and the issue did not recur.


I do think that there may be some tripod detection logic going on but I have no way of proving it. I turned the camera off and turned it back on and could not immediately reproduce the issue.


Either way - I wasn't able to find anything that I couldn't blame on user error [like me hand-holding the camera] and I did follow the directions I found here as well as a couple of other places.

My testing is by no means exhaustive but if I tried this hard to do it on purpose and couldn't I have to wonder how much if at all it'll affect me in the real world. It's something I'll try to keep an eye out for.

As a programmer this sounds like bad logic to me. It sounds like either a wrong branch is being taken or perhaps a situation was overlooked.


Overall thankfully I'm still really happy with my camera and on the upside most of the newer lenses seem to come with IS built in and the IBIS seems to be working great with Lens IS.
It has indeed some tripod detection logic that some times is still on a few seconds after it is off the tripod and a starts a few seconds after being on the tripod. So i would try to do it again with a freshly started camera no using tripod at all.
you are the 4th person I read that carefully try to reproduce the issue and it’s not there, but I really hope it’s not a matter of bad batches but “user error” in the positive way. Just because I really hope it’s just software…
 
Upvote 0
It has indeed some tripod detection logic that some times is still on a few seconds after it is off the tripod and a starts a few seconds after being on the tripod. So i would try to do it again with a freshly started camera no using tripod at all.
you are the 4th person I read that carefully try to reproduce the issue and it’s not there, but I really hope it’s not a matter of bad batches but “user error” in the positive way. Just because I really hope it’s just software…
I did try without the tripod - as well as doing it immediately after turning the camera / moving the tripod around before setting it down and immediately taking pictures.

The issue is that I am not so steady and it was darkish in the room I was testing in so hand-holding at 1/10th wasn't really consistent for me even with IS working 'normal'. Even with clear pictures - there was no way I couldn't be sure any twist / movement of the camera wasn't me moving.

I'll need to mess around with it some more when I have more light to work with. I was stopped down so I could make sure the edges would be reasonably sharp making it easier for me to detect issues.

I do also have an EF 24-70mm f2.8L - I can test with that one too. I also have several EF lenses. Would you think it better to test wide open or zoomed in? I can certainly do both but would like to focus on the most likely opportunity to catch issues.
 
Last edited:
Upvote 0