Today i officially felt left behind with being a Canon shooter

[3kramd5]

IBIS has nothing to do with mirrorless, that's a red herring, it can be in any camera.

Yes it does. At least in viewfinder operation.

I must admit I’ve never thought about this before, but: why?

What has the viewfinder to do with IBIS? Does IBIS use image analysis (as opposed to accelerometers and gyros) to inform the sensor actuation? (I’m not speaking of canon’s “crop the image” style here)

 

[Kit.]

Let's not forget that mirrorless digital consumer cameras have been around for way longer than DSLRs - if you look at all digital cameras and don't fixate on those that have fixed lenses vs those that have removable lenses. Mirrorless isn't a technology that's developed over the last ten years. It's a technology that's developed over the last 25 years or so

My first camera was a fullframe mirrorless. In 1970s.

While the DSLR has taken over the entire market for SLR cameras, mirrorless cameras of all kinds have overtaken them and outsell by several orders of magnitude simply because they can now be reduced in size to the point they fit inside a cellphone.

Mirrorless cameras were always outselling SLRs.

The average person no longer needs a camera.

The average person already has a camera. And that camera is mirrorless. What the average person does not need (and never needed) is an ILC.

And now the DSLR is stuck. Without a full live reading of the digital sensor for the camera CPU to interpret the DSLR has to do its best from the limited view of the image it gets from the low resolution PDAF sensor behind the mirror. It's quite incredible how far they've pushed this technology and how well and how fast a modern DSLR such as the 5D Mark IV can focus,

The main advantage of an electronic SLR is a moving semi-transparent mirror that allows it to use three different sensors to prepare and capture the same image. For the modern camera, it's a 3d-information sensor at the bottom of mirrorbox, fast 2d color exposure and object tracking sensor above the mirrorbox, and the final image sensor behind the mirrorbox.

If you think that the two specialized fast sensors are not good enough for you because of their "low" resolution (which is still too high for the contemporary image-processing neural networks to run at full fps on embedded chips), you can turn them off by locking the mirror up, which will conveniently convert your camera into "mirrorless".

but the DSLR is hindered. It can't do proper eye autofocus, realtime histograms, focus peaking and all the other things that we take for granted with the modern generation of mirrorless cameras.

They can do that - and more. It is just not the market segment they are optimized for. When you really need fast autofocus acquisition (and not just tracking), you have neither time nor screen space for focus peaking and realtime histograms. You still may need to grasp what is in focus and what is in the dynamic range, but your brain should be pretty capable of doing it over OVF image.

When people realise their phone camera has a more intelligent focus and tracking system than their $3000 DSLR they start to wonder if they made a mistake.

...except that it doesn't.

And if it does, it burns the phone battery in no time.

Now, I'm sure many of you will start saying that your DSLR is perfectly capable of taking the photos that you want to take today - and that's true. And that's why I also still use a DSLR.

My phone is perfectly capable of taking the majority of pictures I need. My PowerShot is perfectly capable of taking the majority of the rest.

I'm not saying the DSLR is useless. I'm simply stating the economic case for why camera companies will inevitably drop them.

I am not saying the market for DSLRs won't shrink (or even disappear). I'm simply stating that there is no reason to believe that FF MILCs will do any better.

 

[neuroanatomist]

I must admit I’ve never thought about this before, but: why?

What has the viewfinder to do with IBIS? Does IBIS use image analysis (as opposed to accelerometers and gyros) to inform the sensor actuation? (I’m not speaking of canon’s “crop the image” style here)

IBIS stabilizes the image seen in the EVF, the same way lens IS stabilizes the image in an OVF or EVF. I find that stabilization to be useful with longer lenses (but there, IBIS is less effective than lens IS).

 

[3kramd5]

IBIS stabilizes the image seen in the EVF, the same way lens IS stabilizes the image in an OVF or EVF. I find that stabilization to be useful with longer lenses (but there, IBIS is less effective than lens IS).

But is there something which fundamentally prevents IBIS from stabilizing the captured image in an SLR, e.g. requiring image analysis to function?

If as I always assumed they use gyros and accelerometers, I’m out of reasons why an optical viewfinder and its mirror are a contraindication for IBIS. Although composition without it in the viewfinder could be more difficult, why couldn’t IBIS be operating throughout and thus active when the shutter is open?

[edited for boneheadedness]

 

[Don Haines]

Sorry answered to the wrong post (and there is no delete post button)...

Worst space scenario, ask the admins to delete it, but when I make a post, the option to delete shows up at the bottom....

 

[takesome1]

Worst space scenario, ask the admins to delete it

They do not have a problem deleting some of my posts. There so nice I do not even have to ask.

 

[Don Haines]

But is there something which fundamentally prevents IBIS from stabilizing the view in an OVF, e.g. requiring image analysis to function?

If as I always assumed they use gyros and accelerometers, I’m out of reasons why an optical viewfinder and its mirror are a contraindication for IBIS.

IBIS works by shifting the sensor around.... in a mirrored camera, there are two separate light paths..... one to the viewfinder, and the other (mirror up) to the sensor. Whatever is done shifting the sensor around will have no impact on the view through the OVF..... in fact, when the mirror is up, you have no view....

As to which is better, In lens or sensor image stabilization, IBIS is better for wide angles, while OS is better for long lenses..... an interesting variation comes from Panasonic, who uses both.....

 

[3kramd5]

IBIS works by shifting the sensor around.... in a mirrored camera, there are two separate light paths..... one to the viewfinder, and the other (mirror up) to the sensor. Whatever is done shifting the sensor around will have no impact on the view through the OVF..... in fact, when the mirror is up, you have no view....

As to which is better, In lens or sensor image stabilization, IBIS is better for wide angles, while OS is better for long lenses..... an interesting variation comes from Panasonic, who uses both.....

Yes, I was being dumb when I said “view in the OVF.” Really I meant “image capture,” and I edited the post accordingly.

You wouldn’t be able to see it working, but that wouldn’t prevent it from working, right?

 

[Don Haines]

Another interesting possibility.... I wonder if Canon is going to use a combo of OS and IBIS?

 

[Don Haines]

Yes, I was being dumb when I said “view in the OVF.” Really I meant “image capture,” and I edited the post accordingly.

You wouldn’t be able to see it working, but that wouldn’t prevent it from working, right?

It isn’t the same, but on the 6D2 when shooting video, the software can jiggle the used area of the sensor around to add more stability to video, but there is no comparable function for stills.....

 

[Sporgon]

But is there something which fundamentally prevents IBIS from stabilizing the captured image in an SLR, e.g. requiring image analysis to function?

If as I always assumed they use gyros and accelerometers, I’m out of reasons why an optical viewfinder and its mirror are a contraindication for IBIS. Although composition without it in the viewfinder could be more difficult, why couldn’t IBIS be operating throughout and thus active when the shutter is open?

[edited for boneheadedness]

Pentax seem to manage just fine with IBIS in their K series dslrs. Pretty effective too on the ones I've tried though I didn't like the way the sensor assembly sloped about in the body when the camera wasn't activated. Obviously it doesn't stabilise the viewfinder.

 

[fullstop]

Pentax seem to manage just fine with IBIS in their K series dslrs. Pretty effective too on the ones I've tried though I didn't like the way the sensor assembly sloped about in the body when the camera wasn't activated. Obviously it doesn't stabilise the viewfinder.

true, thx. . overlooked the Pentax implementation.

@3kramd5 - I stand corrected, you were right to question my statement.

Still wondering whether IBIS in a DSLR will not also come with some "unwanted effects". At least image captured might be slightly off from the one seen in viewfinder, if IBIS moves sensor to an extreme position right at the moment of capture ... eg. in high fps sequence? Can CPU/software/algorithms really sort all of that out? Is that (one of the) reasons why K-1 II can only do 4.4 fps max speed?

 

[3kramd5]

true, thx. . overlooked the Pentax implementation.

@3kramd5 - I stand corrected, you were right to question my statement.

Still wondering whether IBIS in a DSLR will not also come with some "unwanted effects". At least image captured might be slightly off from the one seen in viewfinder, if IBIS moves sensor to an extreme position right at the moment of capture ... eg. in high fps sequence? Can CPU/software/algorithms really sort all of that out? Is that (one of the) reasons why K-1 II can only do 4.4 fps max speed?

It’s conceivable I suppose that there could be some cropping, exacacerbated by less than full viewfinder coverage. I don’t think the sensor moves all that much. Maybe it could be mitigated by making it slightly bigger such that it would never move off the format image circle.

 

[crazyrunner33]

IBIS has nothing to do with mirrorless, that's a red herring, it can be in any camera. Short flangeback distances cause big light fall off issues at the edges of the sensor that is difficult to correct. Canon DSLR's have silent shooting if you know how to use it.

I suspect that the percentage of people using older manual focus and manual aperture lenses on a high mp camera will be few, particularly if they are aware of the poor resolution performance.

Nikon proved this with the failed DF.

Which ones and how? The only silent modes should have been called quieter mode. The photographers we used for BTS photographs used those modes including LV silent and still have the sound of the shutter popping up on the audio track.

Knock on the old lenses as much as you want, but there's still a lot of people who enjoy playing around with them. Or would like to buy a specialty lens that won't be used too often, but don't want to pay thousands of dollars for a modern one or hassle with renting. Video shooters benefit the most from the old school lenses where the video is still 8 megapixels on the newer camera crops.

 

[dtaylor]

And now the DSLR is stuck. Without a full live reading of the digital sensor for the camera CPU to interpret the DSLR has to do its best from the limited view of the image it gets from the low resolution PDAF sensor behind the mirror.

But that dedicated off-mirror PDAF sensor is better than mirrorless on-sensor PDAF elements at acquisition (i.e. predicting how far out of focus the lens is and driving the lens to the predicted point) and tracking speed. This has to do with limits placed on the design of on-sensor elements (i.e. you don't want to see stripes in your photos) and is not easily overcome in mirrorless.

It's quite incredible how far they've pushed this technology and how well and how fast a modern DSLR such as the 5D Mark IV can focus, but the DSLR is hindered. It can't do proper eye autofocus, realtime histograms, focus peaking and all the other things that we take for granted with the modern generation of mirrorless cameras.

Proper eye AF only requires a metering sensor with sufficient resolution. Note that past a certain point more resolution is not necessarily superior for AI tracking of objects like eyes. Focus peaking is huge with manually focused lenses, but then I don't think anyone would argue the superiority of mirrorless for adapted manual focus lenses.

When people realise their phone camera has a more intelligent focus and tracking system than their $3000 DSLR they start to wonder if they made a mistake.

There isn't a phone made that can touch the AF on my Canon DSLRs within their area of expertise.

I'm simply stating the economic case for why camera companies will inevitably drop them.

Mirrorless ILCs are not new. They have been around for a decade and are mature. They do out sell DSLRs in Japan. In Asia DSLRs out sell MILC. In Europe DSLRs are ahead 2:1. In the Americas, 3:1.

Camera companies will make what people buy. I'm not convinced people who are interested in ILC bodies are all of a sudden going to rush to EVFs simply because there are a few more niche (FF is a niche) product offerings available.

 

[neuroanatomist]

Yes, I was being dumb when I said “view in the OVF.” Really I meant “image capture,” and I edited the post accordingly.

You wouldn’t be able to see it working, but that wouldn’t prevent it from working, right?

Certainly IBIS can work in a DSLR.

But...AvTvM stated that IBIS offers a MILC-specific advantage in terms of viewfinder (EVF) operation. You asked why. I replied that it stabilizes the image in the viewfinder.

Note that you asked to fundamentally different questions, one: does IBIS offer a unique advantage on a MILC?, two: is IBIS contraindicated on a DSLR? The answer to the first is yes, the answer to the second is no.

 

[sdsr]

IBIS has nothing to do with mirrorless, that's a red herring, it can be in any camera. Short flangeback distances cause big light fall off issues at the edges of the sensor that is difficult to correct. Canon DSLR's have silent shooting if you know how to use it.

I suspect that the percentage of people using older manual focus and manual aperture lenses on a high mp camera will be few, particularly if they are aware of the poor resolution performance.

Nikon proved this with the failed DF.

You're right re IBIS, and you're probably right that the proportion of people wanting to use manual lenses, old or new, is tiny, but I would note that (a) many older manual lenses don't have poor resolution performance on high mp cameras; (b) some of the highest resolution lenses made today are manual focus/aperture (e.g. Rokinon 135mm f/2); and (c) some older lenses/lens designs are appealing despite the fact that (or even because) their resolution is poor (one of my favorite lenses to use wide open is the Helios 85 1.5 which photozone dismisses as producing "color soup" away from the center; they seem to assume that everyone will agree that's a bad thing, but if you like that lens's distinctive soup...).

 

[tron]

Worst space scenario, ask the admins to delete it, but when I make a post, the option to delete shows up at the bottom....

I found it too. But it didn't exist before! It also asks to enter the reason for the deletion.
Thanks for letting me know anyway...

 

[3kramd5]

does IBIS offer a unique advantage on a MILC?

I’m not sure I asked that. If I phrased something accordingly I misspoke. I took Spokane’s post as the subject and responded to the repudiation of it, and the “viewfinder operation” part of AVTVMs post didn’t really register (my bad)
Either way, fair enough. I had my suspicion answered.

 

[neuroanatomist]

I’m not sure I asked that. If I phrased something accordingly I misspoke.
Either way, fair enough. I had my suspicion answered.

I was referring to this exchange. But perhaps I misunderstood.