Soon, 40MP without the tripod: A conversation with Setsuya Kataoka from Olympus

I suppose that camera movement has a more pronounced visible effect in very high resolution sensors, in the sense that you can't get full sensor resolution benefit (eg. the images look sharp, but not 40MP sharp). This is somewhat logical, imagine what would happen if you would look at the hand movements with a smaller and smaller microscope. The movements would appear larger and larger.

yes, i think both Luke and me know that. The point is that a 40MPx image obtained from a 16MPx sensor shifting camera over a total exposure of 1/60 seconds, handheld, will be MUCH more affected by this than a 40MPx sensor shooting a single exposure of 1/600 seconds.
Also the "OMD users can use or not a tripod while others have to use it always" is just plain stupid...

Luke, I think he means that you can achieve 16mp handheld with IBIS, but not 40mp without IBIS.
So for 40mp you always need a tripod, no matter what camera you use. But for 16mp you can use a stabilized camera like OMD.
I also guess that IBIS for a 40mp sensor is currently not possible, but we'll see what Sony does with the 7R.2.

Badi, how many of these stabilized lenses can really resolve 40 or 50 mp?
I still believe his statements more than yours. :-)

@agachart the vast vast majprity of us domt wamt a higher resolution sensor. You go up in megapixels and high iso performance will suffer. 16mpx is more than enough provided your framing doesnt suck consistently

Somehow a similar multi-sampling technique has been in use for military radar imagers for many decades. Multiple samples with a slight time/space difference allow for better resolution for a number of reasons, a prominent one being diffraction.

"impossible to use.."

Wow.... the most popular mirroless camera on flicker, seems like lots of people find it "possible"

"M43 camera with controls that someone has actually TESTED FOR USABILITY!"
Huh? Have you shot with various Oly and Panny MFTs?
GH2, GH3, GH4, EM1 are totally usable.
GX7 and EM5 are very useable, although due to their smaller size you can accidentally bump a button.
Face it, there's a limit to small size without conflict in the physical layout.
If the smaller cameras were bigger, then everyone would be griping about how big they were. Except those who gripe about small cameras not looking impressive enough. Or griping about bumping buttons by mistake.
The dimensions of all cameras are clearly posted. Pick the one that works for you. It is not really so difficult.

@richard stern
the olympus em-1 has tracking autofocus up to 9 frames per second with the most recent update. The gh4 is really good at it too.

You're completely incorrect about the ISOs being "named" wrong. You can't name ISO numbers arbitrarily... in fact, the very name, "International Standards Organization", ought to suggest that fact. All major camera makers rate their sensor based on the ISO 12232:2006 standard. And while that standard actually offers five approved techniques for measuring effective ISO, all Japanese camera manufucturers are required by the CIPA DC-004 standard to use either the REI or SOS method, both improved methods introduced in the 2006 version of the ISO 12232 standard.

The sensor works great and there is no reason to change it - there is no better technology available at the moment. These refinements and improvements are the sign of a maturing system. There is not much else that they can do to make the system better except to provide more lenses and features like these.

Oh, and the discussion of ISO 100 was some of the most pertinent commentary than I've seen from an exec.

@HowaboutRAW - the point is, everything has to start somewhere. You can't just write everything off as a "gimmick".

BTW, the rate of technological progression happens much faster today than it did back in the early 1980's, three decades ago. That's what happens when computing power has been doubling every two years (Moore's Law).

And still my 4 year old laptop is faster than most, 2015 fast laptop.

The OMD EM5 II multi-shot samples (at least some of them, at some sites) show a clear, substantial improvement in image detail over the single frame 16MP images, so why should we care that Pentax says that it was not able to achieve any significant IQ gains in its efforts? I am tempted to quote Aesop "Those grapes are sour anyway."

Rich,

The EM1 uses a Panasonic sensor I believe, though yes there are plenty of Sony sensors in wide use.

However Sony does NOT build the Pentax in body sensor stabilizer.

Zvonimir- sorry but this is a weird thing to say considering the 40 MP mode clearly provides for more resolution, even without color moire artifacts, than the nominal 16 MP resolution.

I don't understand why you mention all that with clear evidence already to the contrary.

Rich- I wouldn't say Pentax "brings nothing to the table" either though. I mean, their K-3 and other cameras capable of turning the AA filter on/off (the actual results, not that it really has an AA filter) is actually pretty innovative, and Pentax had several axis IBIS long before Olympus.

Raist3d - That is not true about Pentax having multi-axis IBIS long before Oly. Prove it.

Prove it? Pentax had axis tilt. So you have up/down and axis tilt (look it up). That's multi axial IBIS. You can even shift the sensor and tilt when setting the camera on a tripod to adjust the composition a little bit - as an option- so you don't have to move the camera on the tripod! The K-5 came long before the OMD EM5, and this capability existed in Pentax SDLR's before the K-5.

This is funny:
A. Pentax tried and couldn't come up with anything to bring to the market.
B. Olympus tried and produced a new feature.

Pentax failed, Olympus succeeded.

It seems typical of Pentax users to think their company came ahead whenever it failed. That's why they think Pentax is so far ahead of everyone else. :)

caver3d- you can look at the Pentax 100D - the first one back in December 2006. Look at the 2nd page of features. That's X, Y and Z axis. Nobody else had something like this until Olympus further developed with the OMD EM-5 pretty much. And sure, the Olympus covers more axes, but doesn't take away the Pentax anti-shake is multi-axis, including tilt.

http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCAQFjAA&url=http%3A%2F%2Fwww.dpreview.com%2Freviews%2Fpentaxk100d&ei=aRrgVPmxKcmbgwS_yYIw&usg=AFQjCNG7ZsDpyMCaUhdaPLQdGb363OzjiA&sig2=aTiVo88cJASrfEmI78lfPw&bvm=bv.85970519,d.eXY

Pentax had it first, and it had tilt. The E-3 body doesn't have tilt. Neither does the E-5 body. The OMD EM-5 is the first to have the 5-axis IBIS.

'Megapickles' my new favourite typo.

I have the new em5 mk2. And the high res is indeed better. This is not marketing but the real deal.

Some test shots with my new OMD EM5 Mk2

https://plus.google.com/photos/101191223694267273293/albums/6115716250200490145?authkey=CIPmnb2irMy_VA

Shot 1: Handheld slow shutter speed

Shot 2: ditto, converted from RAW

Shot 3-10, just some general shots. First one out of camera jpeg, second one from RAW

Shot 11-12, focus on Gundam model, Jpeg, converted from Raw, respectively

Shot 13, high res mode, focus on Gundam

Shot 14-15, focus on kid, Jpeg, converted from Raw

Shot 16-17, high res mode, focus on kid, jpeg

Lassoni - fate!
Howabout- the question was about my claim that Pentax had multi-axis IBIS long before Olympus. So yes, my answer proved the point.

Moreover, the Pentax mechanism other than gyros for measuring pitch & yaw is all that is necessary for the sensor to correct for those (see the Sony 5 axis ibis discussion elsewhere). The Olympus sensor also doe snot rotate along x & y even though Olympus 3d rendering animations show it doing so.

Zvonimir- that is fine, but the problem is that Olympus already proved that they were able to get notable gains in resolution data. The fact that the smaller sensor Olympus is using out resolves in this mode the Pentax K-3 is quite remarkable.

Nobody is saying what Pentax should do or not in order to increase resolution.You quoted the claim from Pentax and the claim as far as Olympus is concerned *has been proven wrong* already because Olympus *did achieve* better resolution.

Of course Olympus's solution has limitations, but it seems they are working to expand the operating window of this solution. So basically it doesn't matter what Pentax said regarding getting more data- Olympus *proved* they are getting more data with this approach. It's not theory.

Precisely. The high res mode is truly remarkable feature / idea at the moment, and will only get even better as time goes on. Kinda turns the E-M5 into a meril since those too get best results on a tripod?

Every company *claims* to provide. Olympus 5-axis is the first which was actually quantified by the tests.

Otherwise, you can't replace OIS in lenses on a whim. But you can replace an old camera with new one, and get an immediate IBIS upgrade an *all* lenses at once.

No matter what Panasonic, Nikon, Canon say, you can't joggle around lenses inside lenses to provide image stabilization and not expect any optical penalty. Take any multi-element lens and shift one of the elements, or tilt it even a small amount and watch what happens to the image.

You get an IQ penalty with tilting/shifting the sensor, as well, no matter what you say.

That statement simply says that it is more important to have image stabilization than to have more pixels. And I agree. Without image stabilization I would get blurry images - more pixels wouldn't improve anything.

As for other companies offering IS, only Pentax offers it in-camera. The lens based one is great if you like it, but even then those systems that rely on it don't provide it in all their lenses.

@JASC, You get an IQ penalty with tilting/shifting the IS lens element, as well, no matter what you say.

BTW, you are welcome to find any blur in the E-M5 + 75mm shots:

http://robinwong.blogspot.com/2012/07/olympus-mzuiko-75mm-f18-review-street.html

You want me to invite you to find blur in my 70-200/4 IS or 100/2.8 IS shots as well? I will post them for you in a tiny web resolution, with the EXIF stripped, just in case. Oh, and I will post a few 100% high speed crops, as well.

"Comment edited 3 times, last edit 3 minutes after posting"

How unnerving it is, to make a bogus comment, being upvoted by the uninformed, but try to vindicate it anyway, because you do not want to be associated with the uninformed?

Modern IS tech is long beyond the infancy point where the micro-blur due to the floating sensor or the floating lens element was an issue.

Uninformed so much? Or just reading comprehension fails you?

"the floating sensor" - is how IBIS is implemented.

"the floating lens element" - is how OIS is implemented.

For example, first hit on YT:
https://www.youtube.com/watch?v=bBX4G9J28nM

Good. Then you agree with my first post here.

"On the other hand, OMD user will ALWAYS need tripod and totally stationary subject as the whole process takes ONE SECOND."

So is it still going to be "ALWAYS", even when the process takes far less time than "ONE SECOND"? LOL. Did you even read what he said about where the technology is headed? Future OMD users will be able to do this process in fractions of a second.

He is doing blatant marketing but I think he has a point: going far above 25MP on a camera implies that, if you really whant to get all the detais, assuming you are using excellent lenses, you have to keep that camera on a tripod or just get average detail, while shoting at 16MP handheld you are not loosing anything (you paid for). What he doesnt whant to say (and should) is that you can get excellent hi res pictures with an OM-5DII at 1/3 of the cost of a 36MP camera, and excellent m4/3 lenses will cost a fraction of FF ones. You won't strictly compete but let's talk after a full day climbing a mountain which is the best camera. ; )

"40MP FF camera paired with 100mm lens requires the same amount of shutter speed 10MP m43 camera paired with 100mm lens requires to be sharp at pixel level."

On MFT you would use a 50mm lens for the same composition. So the same shutter speed would give you a better guarantee of a stable image than on the FF camera.

"On the other hand, OMD user will ALWAYS need tripod and totally stationary subject as the whole process takes ONE SECOND."

When he said you don't need a tripod he meant using the camera in regular operating mode where you're just taking a 16MP image. He wasn't talking about the high resolution mode.

I agree though that this second statement could be criticized - a FF user could always shoot in cropped mode and use a different lens. But then you're basically throwing the FF aspect away.

"How many FF cameras have IBIS?"

How many 100mm FF lenses do not have IS?

Shooting 40MP FF camera is nothing special - it is affected by hand shake EXACTLY the same as 10MP m43 camera with 1/2 FOV lens (or roughly twice 'effective' FOV). So if you can handheld 10MP 100mm 1/125sec you will be able to handheld 40MP 50mm 1/125sec just fine.

"As long as I have a usable ISO 12,800, I'm good."

Wow.

Cameras with usable ISO 12800 have appeared no more than 2 years ago.

How were you shooting before that??

Another comment from someone that doesn't understand that equivalent results are obtained on small formats at *lower ISO*, and thus small formats don't need high ISO, they just need faster lenses.

Where small formats cannot match large ones is at low ISO - that is why the discussion of ISO 100 is more relevant to the MFT system than high ISO performance.

Not seeing it:

http://www.dpreview.com/reviews/nikond3s/30

Fine details are washed away by NR.

High ISO is relevant to larger formats, where you need to keep the aperture closed to gain DOF and that forces you to push the ISO. On a format like MFT, ISO 100 is far more relevant. This is not BS, it's a frank discussion that unfortunately goes above the heard of ignorant users.

There is no way ISO X from a small format will have the same performance as the same ISO on a larger format, within the same technological timeframe. It's not possible. Talking about it just demonstrates ignorance and wishful thinking.

@HowAboutRAW: Don't mix up topics. If you want thinner DOF, High ISO won't give you that. For a given DOF, the way you get that shot on a small format is with a faster lens *at lower ISO*. That is why low ISO is much more important than high ISO for MFT.

That *was* my point: I was just describing the use on a larger format: that the need for high ISO is derived from a use of higher F-stops. On a smaller format, lower ISO and smaller F-stops will provide the same DOF and exposure time as on the larger format, so there is no need for high ISO, just a need for faster apertures and lower ISO.

Olympus knows this - that is why they talk about low ISO and about producing even faster lenses.

But not often do you hear that a camera is on the way that fixes the major limitations of another camera that hasn't been released yet.
They are kind of shooting themselves in the foot, but I'm sure Olympus loyalists will buy it regardless.

Oh c'mon. It is a niche feature anyway. Few people need the high res mode. Especially with the existing limitation, which are not going to be solved completely in the next model either.

Better IBIS, better video, fully articulated LCD - it is a nice refresh of the original E-M5.

...which can be said about practically digital every camera ever. Lol.

It'a not "better" but it is an alternative wich brings several undeniable advantages: way lighter well built bodies and lenses, good full sensor resolution at 16MP, and keeping the res down, with a BSI implementation, you'll get high MP imagages with larger area per sensor dot, meaning improving light sensitivity at a fraction of the cosy. Full frame will keep selling to pro and pro wannabies, but competition will be a though one for APS-C, because most of cropped 1.5/6 sensors have crappy dedicated glasses (Ni/Ca/Pe/Sony), because FF lenses are large, expensive and get weird focals on them, and because m4/3 will suddently get more attractive for everyone. I'm going to upgrade from Canon to this OM-5D II "because" of this 40mp tecnology (and weight, and small good lens/body, and better IS, and in body perspective correction, and 100% wiewfinder coverage, and dedicated common focal lenses) and I think this camera will apeal to many other as well, but Setsuya Kataoka here is speaking of future, better cameras. Other will try this sensor shift technology, but larger sensor are penalized and Sony got much worse res on a FF axis stabilization than Olympus with their one, whic allows NOW, a 40MP image on a m4/3 sensor. Remember that this pro sealed OM-D5 II will cost about 1100€ versus 3K+ of every other Hig Res FF weighting 1/3!

In my experience, the number one complaint of the average snapshooter is poor low-light performance; so I take issue with your assertion that high-ISO settings are not often used. In the past, they were not often used because the results were too noisy, but this is definitely starting to change. I predict that better low-light image quality will be wildly popular with snapshooters the world over.

The average snapshooter doesn't need over ISO6400 unless they are shooting their superzoom at f6.3 and expect to capture "the moment" at a nighttime, candlelight wedding.

Bhima78:

That's way beyond ISO 6400, proving that those who make claims about high ISO digital shooting oft don't know the subject real well.

This quote coming from a company that has minimum ISO of 200 on a lot of their cameras. Haha.. What BS.

The statement was about snapshooters... casual users that want better images than their smartphone. They are rarely in a situation that requires anything over ISO6400. Heck I rarely find myself in need over ISO6400 and I take my camera with me everywhere. This isn't to say there aren't actual photographers that do require much better performance with even higher ISOs, but I don't think they fit the description of "snapshooter".

it has nothing to do with data acquisition from the sensor that's just movie mode. It's the IS motors actuation that needs to be super quick and precise to move the sensor to the correct spot before the next data acquisition takes place.
Oly needed to make the IS that much better to achive the hires mode. It sounds like they think they can do better. Let's hope it makes it to the em1.2.

otto, I was asking is it possible smultaneously to engage OIS and 40MP mode? The sensor shift would have to add together the movements required for both functions. So at the same time as executing its 8-point circuit to achieve the 40MP image, the sensor would need to overlay the compensatory movements required for OIS. It might be possible but it is far from self-evident, and the issue is not addressed in the article.

Agree. Funny, Canon is trying to convince us that FF now is a good alternative to MF. Olympus is telling us that m43 is even better than FF. Therefore, m43 is better than MF!

Also, don't forget that Canon makes the best sensors in the industry (according to Canon of course)

:)

In order to take maximum advantage of high resolution systems of any kind; firstly, high end primes are called for. Secondly, very solid tripods, ... weighted.
If you are hand-holding any camera systen today, .. Yes! you can get acceptably very sharp images.

A 36, 40mp is no better than the 20, 24mp when hand held.

I think what isn't being made clear here is that if you shoot mainly for uncroped images to be printed at reasonable size or for screen view mainly, then we don't need tripods.
However, be it Olympus or Nikon FF, to get the maximum results from our heafty investment, .. we need a stable platform.

I shoot both M43 and with a Nikon D800. While the D800 results in better all-around photos, the window to attain those nice photos is smaller. With my OMD, I can shoot in ridiculous low light scenarios, stay quite portable and carry a full kit for travel. M43 won't replace high res DSLRs, but I find the entire system to be far more flexible.

You don't think shutter shock is a limitation?

Canon PR guy, that DP is afraid of asking him reason important questions.

Ahh, I see. Yeah but these two are not from the same universe.
Chuck is a through & through salesman mouthpiece from Canon USA, he could just as well be selling you a used subaru. Setsuya is probably a physicist that has been designing Olympus cameras for 20 years.
I am sure the mother company Canon has these kind of guys locked up in some basement somewhere in Japan being frustrated. These guys could explain exactly how the AA filter works.

Yeah 1/60th of a second is a LOT of time for an Oly camera with the 5 axis IBIS. That multi-shot 40MP shooting mode could be used for a much larger variety of shooting situations if they can truly nail that.

It took me years to perfect my ability to brace a camera on a tiny leica-style tripod against a tree or wall, are you saying I should stop using it just when I've got the technique down? Bummer.