5 axis IBIS coming to next Canon EOS R series camera [CR2]

[4fun]

All major camera manufactures have discovered that it's significantly cheaper to consolidate a product and have a slightly higher cost than the significant cost that comes from designing two products. Just the tooling alone is outrageous.

IF Canon can profitably make, distribute and sell something like 10 different models of Rebel mirrorslappers ... in about 50 different SKUs ...
THEN having 3 or 4 "pure stills" cameras is no argument. It will not matter. At all.

 

[YuengLinger]

Can't wait for the "fully featured pro RF Body"

Yes, officer, you caught me running a grammerly red light.
Meantime, I'm passing on the marginally featured EOS R, thank you.
I'm sure when the IBIS model comes out, the remaining trustworthy review sites (both of them!) will let us know whether IBIS delivers.

 

[nchoh]

For everyone harping on about rigidly mounted sensors... Canon's sensors have shaken themselves around for ages to clean themselves off. In a very quiet setting, you can even hear the sensor self cleaning on the EOS R. So, I dunno, I think pining away for a "rigidly mounted sensor" is kind of a misnomer. Canon's sensors have not exactly been that rigidly mounted for a long time. I mean, are we really going to get into a discussion about how little tiny movements for self cleaning are ok, but big movements for IBIS are not? I think we're heading down a deep rabbit hole here fretting about sensor misalignment due to IBIS.

I think there is a huge difference attaching an ultrasonic motor to vibrate the sensor, versus, building a 5-axis controlled sensor mount. The former can very robustly and very simply attached to a vibrating plate, the latter requires quite a lot of moving parts. Throw in the fact the size of the stabilization mechanism and I think there is going to be a lot of trade off versus robustness.

 

[Kit Lens Jockey]

I think there is a huge difference attaching an ultrasonic motor to vibrate the sensor, versus, building a 5-axis controlled sensor mount.

Is there though? I mean, in either case, you have to have a way to ensure that where ever the sensor rests after it's done moving, it's in the proper position. I'm really not going to presume one way or another. I don't engineer cameras. But, I trust the people that do. If they say they can successfully put IBIS in a camera, make it work, and not throw the sensor out of alignment, then I trust that they could do that. I think it's a little obsessive to get all frantic over some kind of theoretical failure mode where the IBIS fails or breaks and leaves your sensor hopelessly out of alignment. I mean, does this actually happen with any of the cameras that already have it, or is this just panicking for the sake of panicking?

 

[YuengLinger]

I think there is a huge difference attaching an ultrasonic motor to vibrate the sensor, versus, building a 5-axis controlled sensor mount. The former can very robustly and very simply attached to a vibrating plate, the latter requires quite a lot of moving parts. Throw in the fact the size of the stabilization mechanism and I think there is going to be a lot of trade off versus robustness.

I do hear you. Right now I'm looking at buying a Honda CR-V, which offers a choice in engines--either the tried and true 2.4L VTEC, one of the most reliable ever built, or a new 1.5L turbo that already has reports of problems with gasoline seeping into the engine oil. And no official fix has been announced. I'm glad I didn't buy when the turbo was introduced
in 2017! I'm going with the 2.4L for sure. But this is not as easy a choice as you might think, because the 2.4L is available only in the lowest trim level, meaning I have to give up quite a few features my wife was expecting! But saving the money and having peace of mind, in this case, is worthwhile.

Are you very worried that the IBIS camera will not last as long as current dSLR's? Would 3-5 years of heavy use be reasonable before paying for a repair? Five to seven years? Just wondering what your threshold might be.

If concerned, perhaps waiting a year to jump in would be enough time to see if a new FF with IBIS is failing at a higher rate than current dSLR's.

 

[dak723]

I do hear you. Right now I'm looking at buying a Honda CR-V, which offers a choice in engines--either the tried and true 2.4L VTEC, one of the most reliable ever built, or a new 1.5L turbo that already has reports of problems with gasoline seeping into the engine oil. And no official fix has been announced. I'm glad I didn't buy when the turbo was introduced
in 2017! I'm going with the 2.4L for sure. But this is not as easy a choice as you might think, because the 2.4L is available only in the lowest trim level, meaning I have to give up quite a few features my wife was expecting! But saving the money and having peace of mind, in this case, is worthwhile.

Are you very worried that the IBIS camera will not last as long as current dSLR's? Would 3-5 years of heavy use be reasonable before paying for a repair? Five to seven years? Just wondering what your threshold might be.

If concerned, perhaps waiting a year to jump in would be enough time to see if a new FF with IBIS is failing at a higher rate than current dSLR's.

I don't think anyone is VERY worried (or panicking, as another member has phrased it). Nor has anyone thought that the failure would occur within one year, so waiting a year is not probably going to give you any data. It was just a question that seems quite reasonable - you are adding more moving parts controlling the sensor - so the possibility of sensor alignment failure will be greater over the long run. Is it likely? Probably not, but only time will tell.

 

[YuengLinger]

I don't think anyone is VERY worried (or panicking, as another member has phrased it). Nor has anyone thought that the failure would occur within one year, so waiting a year is not probably going to give you any data. It was just a question that seems quite reasonable - you are adding more moving parts controlling the sensor - so the possibility of sensor alignment failure will be greater over the long run. Is it likely? Probably not, but only time will tell.

A year could give significant data, depending on how many bodies are being used and how often they are being used. The oil problems with the new Honda turbo that I mentioned showed up within six months. (But it took longer for Honda to accept responsibility.)

I'm not predicting problems, merely responding to a post where concerns were raised. Many people wait for new technology to be used for a year or so before buying in. That's prudence, not panic. Personally, if the body has the features I want (including IBIS), and performance is reviewed well, I'd go for it within six months or so. I like firmware to get sorted out. Not sure about being an early adopter for Canon's first FF mirrorless with IBIS. As eager as I am to get one--and the new 50mm 1.2L--the likely price of over $3000 USD would motivate me to be a little prudent.

 

[4fun]

relax guys. It is not like "innovative Canon" is about to launch the first ever camera with in-body stabilizer. It was invented 15 effing years ago. By more innovative companies. Tested and time-tried. It works. Reliably.

• Anti-Shake (AS) - Minolta and Konica Minolta (Minolta introduced the first sensor-based 2-axis image stabilizer with the DiMAGE A1 in 2003)

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

Announced in March 2007 to succeed the E-500, it represents the first use of the new Panasonic MOS sensors instead of the Kodak CCD sensors that Olympus had used previously. It also is the first Olympus DSLR to include in-body image stabilization;

https://en.wikipedia.org/wiki/Olympus_E-510

Lol. "innovative Canon". Damn laggards.

 

[mk0x55]

There is one more use for IBIS than stabilization for that traditional purposes we used lens IS for.
Namely, IBIS can help the camera do continuous exposure bursts (immediately after one another, so that it is hardly noticeable), which can be used for stacking images even if handheld. Why that? For instance for HDR, pixel shift, improving image tonality, etc.
Personally, I'd really appreciate that in the upcoming high MP EOS R-like camera, and I care about these things way more than the difference in image stabilization it can make compared to lens IS.

 

[Don Haines]

relax guys. It is not like "innovative Canon" is about to launch the first ever camera with in-body stabilizer. It was invented 15 effing years ago. By more innovative companies. Tested and time-tried. It works. Reliably.


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


https://en.wikipedia.org/wiki/Olympus_E-510

Lol. "innovative Canon". Damn laggards.

And this is the crux of it.... it is old technology (I have an E-510) but for Canon it is new technology. It remains to be seen how well the Canon implementation performs

 

[4fun]

And this is the crux of it.... it is old technology (I have an E-510) but for Canon it is new technology. It remains to be seen how well the Canon implementation performs

in innovative Canon we trust. lol

 

[Ozarker]

relax guys. It is not like "innovative Canon" is about to launch the first ever camera with in-body stabilizer. It was invented 15 effing years ago. By more innovative companies. Tested and time-tried. It works. Reliably.


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


https://en.wikipedia.org/wiki/Olympus_E-510

Lol. "innovative Canon". Damn laggards.

To be fair: Just because Canon decided to go with in lens IS doesn't mean Canon is less innovative than those who decided to go with sensor IS. The arguement can be turned around either way. Canon just chose a different route that still works better, by the way, at longer focal lengths. There are trade-offs either way. One is not the end all to be all. Stating facts one way or the other is one thing, being constantly obnoxious about things is quite another.

 

[4fun]

Canon chose a route that works for big white teles. They were not innovative enough to even think about anything else. that's all.

From a customer perspective, it would be best to have as much functionality in camera body and as little as possible duplicated in each lens. Most of us have 1 (or max. 2) cameras but multiple lenses.

 

[nchoh]

Is there though? I mean, in either case, you have to have a way to ensure that where ever the sensor rests after it's done moving, it's in the proper position. I'm really not going to presume one way or another. I don't engineer cameras. But, I trust the people that do. If they say they can successfully put IBIS in a camera, make it work, and not throw the sensor out of alignment, then I trust that they could do that. I think it's a little obsessive to get all frantic over some kind of theoretical failure mode where the IBIS fails or breaks and leaves your sensor hopelessly out of alignment. I mean, does this actually happen with any of the cameras that already have it, or is this just panicking for the sake of panicking?

There was a recent rumor that showed the IBIS mechanism. Even though it was drawn in schematics, the IBIS mechanism showed quite a few moving parts.

Contributors are not saying that the engineers are going to mess it up, rather, they pointed out that reliability is an issue that Canon takes seriously and that Canon would delay release until a reliable mechanism is developed,... versus some of the competition who would just deliver less robust systems.

Professionals expect Canon cameras to take a significant beating. Professionals don't expect Sony Cameras to take a beating. That's the difference.

 

[nchoh]

Canon chose a route that works for big white teles. They were not innovative enough to even think about anything else. that's all.

From a design perspective, IL stabilization is better than IB stabilization;
The camera is most stable at the 3 points of contact; left hand, right hand and face. The camera has the most shake out further away from the face. This is exacerbated when the longer the lens becomes. Furthermore, stabilizing tilt and panning in the body means that the sensors to detect the movement need to be orders of magnitude greater than if you were detecting those movements at the lens.

From the above, it makes a whole lot of sense why Canon chose the ILIS rather than IBIS.

 

[4fun]

From a design perspective, IL stabilization is better than IB stabilization;

only for tele lenses on mirrorslappers to get a stabilized viewfinder image. On mirrorfree IBIS works just as well (in principle, not necessarily every real life implementation). In 2018 lens IS is just a duplication of a costly element in each lens, that could be put into 1 camera body and work for all lenses.

Ah yes, a few cheap, tiny acceleration sensors [from any smartphone] could still be placed near front end of lens. But not the moving around of glass elements. But ... innovative Canon.

 

[Ozarker]

only for tele lenses on mirrorslappers to get a stabilized viewfinder image. On mirrorfree IBIS works just as well (in principle, not necessarily every real life implementation). In 2018 lens IS is just a duplication of a costly element in each lens, that could be put into 1 camera body and work for all lenses.

Ah yes, a few cheap, tiny acceleration sensors [from any smartphone] could still be placed near front end of lens. But not the moving around of glass elements. But ... innovative Canon.

*sigh* IBIS does not work as well as IS on tele lenses. It just doesn't, especially on FF. So like I said, each system has its trade-offs. Honestly, I had much rather have to deal with no IS on a short lens than IBIS on a 400mm. Please.

 

[3kramd5]

FWIW, IS uses gyroscopes, not accelerometers; they have to sense rotation.

 

[nchoh]

Ah yes, a few cheap, tiny acceleration sensors [from any smartphone] could still be placed near front end of lens. But not the moving around of glass elements. But ... innovative Canon.

Firstly, IS is currently implemented using direct feedback. That way the lens handles the vibration reductions itself and That is why it works smoothly. Your suggestion means that each lens must communicate back to the camera to tell it to compensate for the movement.

Several problems. Each lens sensor needs to be properly calibrated. The camera needs to have a feedback profile - when lens moves x sensor must move y... So for 5 axes the profile would be huge. Also if there is problem with IS, where would the fault lie? Camera or lens.

Perhaps that is why nobody has. Implemented your innovative idea?

 

[Talys]

only for tele lenses on mirrorslappers to get a stabilized viewfinder image. On mirrorfree IBIS works just as well (in principle, not necessarily every real life implementation). In 2018 lens IS is just a duplication of a costly element in each lens, that could be put into 1 camera body and work for all lenses.

There's just so much in that paragraph that is untrue

Take a Sony A7R3 and Sony GMaster 24-105/4, 70-200/2.8 and 100-400 4.5-5.6 all with OSS.

Take a whole bunch of real-life, hand-held photos with the IS switch "On" and "Off", while leaving IBIS on. Now try it with some adapted EF lenses that have In-lens stabilization. You will most certainly notice that IBIS does not work as well as ILIS, even at portraiture focal lengths.

Let me put it another way. If Sony IBIS worked as well as ILIS, Sony would stop putting image stabilization into their non telephoto lenses. Perhaps there is a way to achieve 4 stops of stabilization in-camera, but it doesn't exist today, and not because camera vendors are trying to squeeze us.

The whole telephoto thing... well, anyways, please just try to shoot some stuff handheld with a telephoto lens with and without in-lens IS and with and without IBIS. Since small movements will take the whole subject out of the image, just imagine trying to compensate for shake. IBIS vs. In-lens for telephoto lenses really has nothing to do with whether or not the camera has a mirror; just IBIS by itself really doesn't do a whole lot at telephoto focal lengths.