Olympus working on 8K video for Micro Four Thirds

I thought 4/3 sensor would be too small for 33MP (which may have to be more given the sensor is square and video is 16:9.

The thing is about Hi res is that it shifts the sensor and takes several photos and combines them into one which takes time and doesn't work well for motion. For 8K you can't do that as Olympus is talking video resolution, not photo and to do that for video you'd need to have all the images at once and at least 24/25/30fps so the current way hi res mode works, wouldn't work for video.

Mar 14, 2017*permalink

AshMills

Not sure of the logic that tech shows up in smaller sensors first because the chip over heats less? surely a small chip of equal pixel count sends the same info, and rather than spreading that heat over a large area does so in a smaller space? I assumed the HiRes tech and IS could be achieved faster because of the smaller chip to move, but heat wasn't really a factor? Handheld HiRes will come for sure, but its not a software issue, its getting the chip to move fast enough in such a tiny amount.

Mar 15, 2017permalink

Peiasdf

Fix E-M1 II's video first. It is worse than camera half its price.

Mar 14, 2017permalink

Akira Watanabe, manager of Olympus Imaging's SLR planning department, said in 2009: "Twelve megapixels is, I think, enough for covering most applications most customers need"

Mar 13, 2017permalink

Albert Valentino

Eight years later and I still agree with this statement 😉

Mar 13, 2017*permalink

MarshallG

He was probably right, although it didn't keep the market from wanting more.
To be honest, I have a 30 million pixel camera now, and all the questions people ask are about dynamic range, not resolution. Canon has 50 mp cameras, and people aren't dumping their Nikons to get them. And a 32mp, 8K monitor can be over two meters tall and still be perfectly crisp from 12".

Mar 13, 2017permalink

Reactive

Has anybody here stood very close to a 4K TV and tried to see the individual pixels? Want to try 8K? Whether its a 60" home TV or a 60 ft stadium screen, 8K is pointless for human eyes. Its only purpose (just like the '3D revolution') is to give something 'new' and 'better' for manufacturers to sell us. I'd much rather Olympus stopped at 4K and devoted their R&D budget to improving image noise and dynamic range.

Mar 13, 2017*permalink

totally agree. even 4K is overkill for a TV if you're sitting watching rather than standing up close.
8K "full dome" video does have a purpose if you're doing VR.

Mar 13, 2017permalink

MarshallG

I demonstrated an 8K TV which was over two meters tall to many experts at a major motion picture studio, and they couldn't see the pixels from even a few inches. As for 4K, the resolution is not the only factor in commercial content (blu-ray, etc) -- it's 10-bit, wide-gamma color with high dynamic range. I would rather have a 1080p, 10-bit HDR video camera than a 4K, 8-bit camera. The market hasn't figured it out yet.

Mar 13, 2017permalink

indeed... I have a, now ageing, 42" 1080p Plasma. One day I'll replace it with a 55" OLED set with 10-bit HDR or Dolby Vision etc, but whether it does 4K won't matter because I simply won't be sitting that close to it.

Mar 13, 2017permalink

MarshallG

That television will support at least 4K, because 55" 1080p panels are no longer being manufactured.

Mar 13, 2017permalink

gkdiamond

I agree that an 8K TV is not needed. For most viewers 1080p is more than enough. As far as 8K, I haven't seen the calculations for distance vs. screen size for 8K. An 8K 60 foot stadium screen may not be pointless.

Mar 14, 2017permalink

Mar 14, 2017*permalink

Reactive

@gkdiamond: Let's get metric here. A 60 foot diagonal stadium screen has a base length of almost 16 meters. If it were a monster 1920x1080 screen, each pixel would be only 8.3 mm (0.3") wide. I challenge anyone to distinguish those from their seat in the stadium! So an 8K stadium screen (2 mm / 0.08" pixels) is clearly pointless.

Mar 14, 2017*permalink

caver3d

What makes you think they are not currently working on image noise and dynamic range like everyone else? Of course they are. Just because they are exploring 8K does not mean they don't have the resources for other things. Take a deep breath and relax.

Mar 14, 2017*permalink

Stubb

Would rather they work on a stabilized, telephoto, macro lens, say a 150 f/4.0. Let's get some working distance while ditching the tripod.

Mar 13, 2017permalink

agachart

this might be new 8K sensor on 4/3 format.
i wish panasonic develop this sensor and bore sony sensor.

That mean olypus/panasonic had share sensor on top model again.
* i think sony don't play with 8k 4/3 sensor.
* olympus still has no semicon factory.

Mar 13, 2017permalink

LJohnK2

Specs for most consumer's are truly meaningless other than for marketing fodder....perhaps 8k video means something to a pro or to the measurabation crowd, but in terms of visual IQ a FF 12 mp sensor still produces more pleasing print output than a 24 MP APS-C or a 16 MP mft....argue away, but it's the same with audio...hence the refusal of vinyl to die....we are not digital beings....wow, I sound like an Artsy sort.

Mar 13, 2017permalink

Mal69

Its not the same with vinyl and let me explain why, its all in the mix, many CD's dynamic range was altered and watered down by the companies who produce the music, it's happening right now to CD releases, the original studio dynamic range is altered resulting in loud sounds but no dynamic range, a CD can easily beat vinyl for sound quality if they do not alter the dynamic range.

Now for camera's as long as the lens is very good then the extra pixels are worth it, it's a garbage in and garbage out thing, poor lens will produce poorer images than a 12 mega pixel camera with a great lens but stick a great lens on that 33 mega pixel camera and assuming everything else within the camera works then it should be much better with more fine detail in the image.

8K will have enormous benefits as you will be to able to shoot high resolution video that can then be downconverted to 4K for the latest televisions and projectors, this always produces more fine detail than a straight 4K shot footage, of course it is going to need a powerful PC to edit such footage, other benefits include the ability to have a still photograph taken from that moving footage.

Resolution alone does not interest me, I want to see 960fps and higher so we can get great slow motion on these camera's and I want to see expanded colour gamuts, technology is getting better and I see no reason why we should not embrace it.

Mar 13, 2017permalink

LJohnK2

.....but in terms of pixels ......phyical size does seem to matter.... from my perception/perspective....a 12 MP FF sensor beats a 12 MP APS-C (or 24 for that matter) when it comes to the "look" of the output....the color transitions and overall image rendering just look better...more natural....perhaps it's partially processing but I think it's more than that.....as this is my observation across brands

Mar 13, 2017permalink

https://www.cnet.com/news/why-ultra-hd-4k-tvs-are-still-stupid

Mar 12, 2017permalink

Andreas Voigt

In a recent interview Olympus suggested, that they have maxed out m43 sensor performance. Scaling that sensor to 33-36 MP means noise performance from hell.

Mar 12, 2017permalink

Of course all 20mp 1" sensors are terrible as well right... Oh wait...

Mar 13, 2017permalink

Mariano Pacifico

Photos has now become an afterthought. Digital camera's primary function nowadays is video that is why DSLR behemoths are losing sales over cellcams because cellcams take better videos and photos and readily available at moments notice. There are technologies that are present in cellcams that are bewilderingly supiciously missing from DSLRs. Anyways only DSLR owners care so much about aberrations not the cellcam users.

Mar 12, 2017*permalink

some makes have taken video seriously for longer than others. Panasonic have had models capable of taking 1080p60 or 1080p50 for years now in even their budget micro4/3rds models.

Mar 16, 2017permalink

If there is a linear, inverse correlation between the number of pixels being read out and the frames per second (meaning: If total pixel readout x max. fps = constant), then

250 million pixels with 5 fps

would equal (same pixel size)

250 million pixels / 8 with 5 x 8 fps = 31.25 million pixels with 40 fps

But this was 1.5 years ago.

Mar 12, 2017*permalink

If I recall this news from Sept. 2015, nothing seems impossible:

"Canon turbo boosts the pixel race with 250-million-pixel APS-H sensor"
https://www.dpreview.com/articles/9255490800/canon-turbo-boosts-the-pixel-race-with-250-million-pixel-aps-h-sensor

Calculated pixel pitch:

wide: 29200 / 19580 = 1.49 µm
high: 20200 / 12600 = 1.60 µm
If the given numbers are correct, the pixels aren't quite quadratic.

Max. frames per second (full sensor readout): 5 fps

For comparison:
Sony HX90 Travel Zoom Cam (30x)

1/2.3 inch BSI-sensor with 18.2 Megapixels (4:3):
Calculated pixel pitch (1inch=16mm) : 1.14 µm

Canon writes about its (almost) 250 Megapixels sensor:
".. despite the exceptionally high pixel count, Canon applied its sensor technologies cultivated over many years to realise an architecture adapted for miniaturised pixels that delivers high-sensitivity, low-noise imaging performance. .."

=> Sounds good !

Mar 12, 2017*permalink

and how does that resolution compare with the capability of their best lens?

Mar 16, 2017permalink

https://2.img-dpreview.com/files/p/TS5368x7152~sample_galleries/0138426498/2926430396.jpg

Good question.
But I don't have the answer.

Generally speaking (I know, that's not enough in this case): If there wasn't a Canon lens to match or better the resolution of this sensor, the sensor wouldn't make much sense, would it?

If I remember right, there was a promotional image from Canon showing a large telephoto lens attached to a protoype box containing this sensor.

An idea: One could do some Google search for satellite optics. What makes their resolution so exceptional? Is it the optics, is it the sensor, or is it the image processing (or all three together) ?

Mar 19, 2017*permalink

Favorable Exponynt

8K from a m43 sensor is 33 mp. Since video isn't 4:3 since the tube went out, even more mp. Totally senseless with current technology. Resolution will get lost in noise and diffraction. Better use larger sensors.

Mar 12, 2017permalink

Dillon Frazier

I don't think a sensible company like Olympus would say this if they couldn't actually do it. Panasonic is also pointing at future 8k. There's clearly still room for sensor growth in m43.

Mar 12, 2017*permalink

Fujica

Of course they'll can get 36MP on that tiny sensor.
The question is will the IQ stand up to it.

Individual pixel size at 36mp will come very close to what we see on phone camera's. So that is about the IQ you will get out of your 4/3 camera.

If you are satisfied with that then fine.
But I'll think many people whom are not pure 'point and shoot' photographers will want something better.

Mar 12, 2017permalink

Favorable Exponynt

Right. m43 is just too small to get the full potential out of 8K. Improvement over 4k will be too small. We already need very big lenses to outresolve 36mp Full frame. On m43 pixels will need to be 3.8 times smaller for the same resolution. Remember that the potential of the cmos sensor has already almost run out because of hitting physical limits of the number of photons that can be collected, and there is diffraction (directly related to pixel size) to take into account.

Mar 12, 2017permalink

(unknown member)

Mar 12, 2017permalink

alexpaynter

You cannot take 8K video on a MFT sensor with just 33 MP. MFT has a 4:3 aspect ratio. So given a 1920 width, the height would need to be 1440.

8K needs some 32 MP in its own aspect ratio. This would equate to about 43 MP if expanded vertically to fill the 4:3 ratio.

Am I missing something here?

Mar 11, 2017permalink

mxx

Holding out for 64K.

Mar 11, 2017permalink

pjl321

You joke but the amazing thing is at some point we will get cameras that do that, let's just hope memory storage and all other techs keep up as 2.1 billion pixels per frame is going to take its toll!

Mar 12, 2017permalink

mxx

Absolutely.

Mar 12, 2017permalink

https://www.youtube.com/watch?v=55SWtb1Hm7w

Sure they will make it, because there will enough people buying it, just like 4K. 99.9% of the population cannot even see the difference between 1080p and 4K at normal viewing distance on TV. When you do see a difference, it most probably means you sit way too close to your TV to enjoy the show. Consumer TVs even have trouble showing sharp 1080p video without artefacts. But as long as the industry can fool people to believe they are actually making better video, it sells like cakes and they are happy to take your money :o) "It's the economy, stupid." There are of course people that understand that clean video with decent signal to noise ratio and high dynamic range is much more important, on lets say 1080p. But why would companies produce that kind of cameras for consumers and at the same time justify 10x the prices for pro equipment ? Wake up you people, consumer 4K/8K/16K/whatever video simply sucks compared to pro 1080p equipment. For consumers, content is what really matters !

Mar 11, 2017*permalink

Melchiorum

I am sorry, but you are either trolling or are willfully ignorant. It is very flattering to think that the entire industry that uses 4K is stupid and you are the only one who "sees the truth" about 4K being unnecessary. Yet the truth is very simple: almost everyone today films 4K for production because it offers a number of clear advantages. In 80% of cases when you see a good quality crisp Full HD video on YouTube, you are seeing down-sampled 4K.

Another huge advantage is the ability to zoom and pan without actually zooming and panning. This has allowed thousands of beginner filmmakers shoot better video without employing expensive fluid tripod heads.

As for "10x prices for pro equipment", your point is just plain silly. You have clearly never worked with video professionally and are completely oblivious about what makes a price of the gear.

Please educate yourself.

Mar 11, 2017*permalink

Ebrahim Saadawi

The OP is not trolling, he's actually correct, but you too, are also correct! You two brought out the advantages and disadvantages of 4K vs. HD.

Advantages:

-Higher resolution/detail viewing
-Cropping ability if mastered to HD
-Finer noise structure/grain, given everything else is constant
-Gave us real 1080p video resolution (C300 like) on affordable cameras by downscaling in post in an era when 1080p HD was nearly 720p lines of resolution worth (5D like)
-Has a higher marketing appeal for clients and customers

Disadvantages:

-Most people don't need or see the extra detail
-Shifted companies from making perfect 1080p cameras/modes (C300 like) to regular 4K ones.
-Can lead to sloppy shooting relying on post-reframing/cropping ability
-Lower lowlight performance, everything being constant
-Needs more powerful computers and storage vs. HD

__________________________________

One thing needs to be made very clear: image resolution is one of more than ten elements that define image quality, not the only factor, far far from being it.

So if someone prefers a 1080p camera with higher dynamic range than a 4K camera, it's not a wrong or bad choice. It's personal. If someone chooses a 4K camera over a 1080p camera for the better colours that particular 4K camera has, it's not a wrong choice. You get the point.

It's a lot of factors that make up a good cinematic image: resolution, dynamic range, highlight falloff, noise structure, lowlight performance, colour science, colour compression/codec quality, motion integrity (rolling vs global shutter), ability to capture pleasing images fast, LOG encoding vs Linear, compression artefacts, macroblocking, motion compression/edge smearing, etc

I mean there's a SH@$LOAD of factors that make up a high quality video image and resolution is simply one (yet big) factor, not the only one, at all. That's why most hollywood films are shot on the 1080p/2.8k Arri Alexas or why most of tge world's broadcasters shoot on HD Canon C300s, etc.

So yes there are situation where HD cameras are preferred over 4K cameras, but the reverse is mostly the case simply because of technological advancement being all applied to 4K while HD has been abandoned.

One thought I have is that they should not have abandoned HD, abandoning the world's most consumed/used video resolution is kind of an illogical move, so I wish cameras were offering really high quality HD capture modes, as the highest resolution or simply next to 4K resolution modes. What we see today on consumer cameras is a 1:1 high quality 4K video signal with line-skipped low quality HD modes. They should all offer the same HD performance as if they were taken and downsampled in post-production. The C100 uses a 4K image/sensor and internally downsamples for a very good example of what I think should continue on consumer cameras who want to capture high quality HD video and not 4K.

Mar 12, 2017permalink

trungtran

When someone claims 99.9% of the population saying they can't tell the difference between 4k and 1080p I stop reading anything else they have to say.

If you are not blind check this out.

Mar 12, 2017permalink

Melchiorum, so all beginners have the money and know how to pan and zoom in post and to downsample to post high quality 1080p ? Best joke ever ... A beginner has no clue what he or she is buying, and that does not matter because it is all about leaning how to shoot video for a beginner. But please keep on buying high res cameras so you can keep fooling yourself.

Mar 12, 2017permalink

Trungtran, as you said yourself, you stopped reading my post too early :o) I said 99.9% does not see the difference at normal viewing distance. Thats not saying that 99.9% does not see the difference ... You understood this time ? Btw, watching youtube movies lowers my expectation for video quality at the bottom. Youtube compression simply destroys the video quality no matter the resolution. In fact, the higher the resolution on youtube, the more precious bandwith and disk space is wasted ...

Mar 12, 2017*permalink

https://upload.wikimedia.org/wikipedia/commons/thumb/f/fa/PureView_Pro_Sensor.png/800px-PureView_Pro_Sensor.png

Ebrahim, nice post. But do you really think these companies will offer a 4K consumer camera that offers high quality down sampled 1080p recording in prores or dnxhd format at consumer prices of 1000 to 2000 USD ? And at the same time selling pro cams like the C300 mark II at 15000 USD ? So yes, 10x the price Melchiorum, whether you like it or not (or are smart enough or not) !

Mar 12, 2017permalink

(unknown member)

Youtube link for video quality?

How about you would upload the footage straight from the post to Google Drive, Mega or even OneDrive and give the link?

Streamin services KILLS almost all quality benefits you can get.

Mar 12, 2017permalink

Ebrahim Saadawi

@Zakzoezie There are very few perfect 1080p cameras that do offer what we are talking about (internally downsampled signal).

-Canon C100 Mark I was actually selling new for 1700$ before being discontinued! That camera does one of the best 4K->HD images ever made, and can record to a small ninja ProRes/DNxHD easily.

-Sony A7s Mark I, the Mark II versions only.offer good 4K modes but the a7s offered the absolute best HD image from any of these ILCs, strangely. The 1080p internally is as sharp and detailed as the 4K downsample and is as detailed as a C100/C300.

I can't think of more cameras that do give us that immediate perfect HD image, they all give good 4K modes and crappy HD modes! All of tgem. The new 5DIV, A7s/rII, GH4/GH5, Em1II, XT2, A6300, 6500, simply all of them.

If price is of no restriction, some perfect 1080p shooters are also:

-Canon 1DC in S35 crop mode. A downsample from a 3K image that gives perfect internal.HD files in C-LOG with exceptional colour and lowlight.

-Sony f35: not joking but this camera had an almost 6K sensor especially designed in a bayer pattern to give the perfect 1080p image for cinema needs. Back when it came out was a 250.000$ and now sells for 5000-8000$ on Ebay. Best HD image ever for FILMIC images.

-Canon XC10/20 1080p mode is a perfect downsample from the 4K stream. Great camera but pricey for a fixed lens one. Image and IS are wonderdul though.

-Hmm who else gives us perfectly downsampled HD images SOOC? Alexa 2.8k -> HD? Just kidding... I can't think of any more, I mean even the FS7 doesn't give good 1080p, just great 4K, like all the other new 4K-craze cameras.

I think I'll get a used 1DC to get both 4K AND perfectly S35 downscaled 1080p to use for most normal shoots where 4K is overkill. Or just a good'ol C100, but drop the 4K mode, yey I don't want to tell clients I CANNOT do 4K if they request, so a 1Dc is better from that prespective. Problem is a 1DC needs to be rigged up a little for comfortable use, like a Zucoto Loup for a good large EVF, a shoulder rig at the very least and I can't use my great EF-S glass so will need new FF glass (C100 takes EF-S). I do have a large EF-S collection I love. A Canon 10-18mm STM, Canon 18-135mm STM IS, Canon 55-250mm STM IS, Canon 50mm f/1.8 II, H!!!

Mar 13, 2017permalink

bleeboo

Olympus should straighten out their auto-focus disaster with the EM1-M2 in 4K first, before chasing down this particular animal...

Mar 11, 2017permalink

Melchiorum

As an E-M1 II owner I have to agree. Auto-focus is pretty much unusable in video right now, which is a shame, because otherwise it is one helluva video shooter with relatively low noise, usable dynamic range and almost steadycam-like stabilization.

Mar 11, 2017permalink

bleeboo

Agreed.

As a fellow E-M1 M2 owner, I have two hopes.

1. That Olympus can rectify this tragedy via firmware, or

2. That the GH5's IBIS is about say...80% as effective as the world class IBIS on the E-M1 M2, and that the DFD on the new Panny will be as good as promised shooting 4K 60p.

Hope springs eternal.

Mar 11, 2017permalink

Ebrahim Saadawi

Sadly it's only ONE camera company that provides usable video autofocus, and that company is very stingy on offering video features. For a good stabilized video image with AF and great colours your only choice is a Canon 80D + an 18-135mm USM IS which does have crazy stabilization and the camera has crazy AF. They just don't give huge video features which is a shame because DPAF would have been a great companion to an affordable camera that shoots in LOG, in 4K, has higher DR, has higher frame rates, has peaking/zebras, has an EVF, etc...

BUT, an 80D in my hands is an extremely good video machine that makes ANYONE capable of capturing high quality, stable, in-focus HD video content. The cheaper 800D should do the same, and 77D, M5 and M6.

Mar 12, 2017permalink

bleeboo

True, however, no 4K, so that's a showstopper.

Yes, I need 4K. Have an LG OLED 55" 4K set, and yes, I can see the difference.

Rented a Canon EOS-1DX MII, and was flabbergasted at video IQ, especially 1080P 120fps, however, the codec for 4K, while providing stunning quality, is inefficient and virtually unplayable without significant PC upgrading.

And, the stabilization there is also a showstopper.

So, fingers crossed and hopeful for the new GH5. Very disappointed no auto-focus when using variable frame rates for slow-mo, but...nothing's perfect.

And, I hope one day Olympus can fix the auto-focus in 4K video catastrophe now inherent in my E-M1 M2 via firmware.

Mar 12, 2017permalink

trungthu

If the dimension of the Olympus's sensor is 18 X 13,5 mm. And when they make a sensor for 8K, that means 18 mm / 8000 px = 2.25 um per pixel.
And I think with that pixel dimension, Olympus can not say anything about the detail or the dynamic range of the picture or motion picture.
The lens can not "see" that so tiny pixel... or they must use software to reproduction the detail. Not the real detail from the sensor.

Mar 11, 2017*permalink

(unknown member)

So how do you explain the smaller sensors that has a smaller pixel pitch than such 4/3" sensor would have?

Remember that 808 Pureview?
The Nokia 808 features a 41 megapixel 1/1.2 in (10.67 × 8 mm) sensor pixel pitch 1.4um.

Total sensor size 7728 x 5368
16:9 mode 7728 X 4354

Multi-aspect sensor even....

And no... not either that sensor was Sony designed...

And that sensor toke great resolution for its size, optics just weren't so great from Zeiss.
https://2.img-dpreview.com/files/p/TS7728x4354~sample_galleries/0138426498/2363618701.jpg
https://2.img-dpreview.com/files/p/TS5368x7152~sample_galleries/0138426498/2926430396.jpg

Mar 12, 2017permalink

trungthu

I don't know you want to say about the "number" of pixels in a sensor or the "quality" of the picture using that tiny pixel pitch,,,
Someone has said about the quality of the Nokia (41Mp), maybe 7 pixels equal one pixel in a normal DSLR... (when saying about the quality of lens plus lens's resolution...)

Mar 15, 2017*permalink

Debankur Mukherjee

Will wait for 16K............before I blow my eyes off.............

Mar 11, 2017permalink

(unknown member)

2007 Olympus had 4K cameras for industry
https://farm4.staticflickr.com/3192/3110575804_a949926f2a_m.jpg
And costed $25000 to rent one of the three cameras for a day.

So why not to have 8K in the designs as they already have a sensors in research and development for next 10 years to come?

Mar 11, 2017permalink

cropedview

where is the part thay they are working or even concidering to work (on that 8K) ? I hear only "yea - anything is possible, look how far tech has come since 2003"... well then read mine statement: "16K is possible, in year 2003 we were on 480p - look at the progress" (so I guess you can assume that I am working on 16K right now)

Mar 11, 2017permalink

Stemu

Much more important for most of us would be if Olympus would give us better 1080p. Now many people shoot 4k in order to scale it down to good 1080p. Why can't cameras give us good 1080p in the first place? It would make storing and postprocessing so much easier. And would be enough quality for most.

Mar 11, 2017permalink

Asking as a non-videographer:

Isn't there a good reason for shooting at 4K and then down-scaling and cropping etc, e.g. better-looking images, image stabilisation etc?

What sort of improvements do you think should be made to standard 1080p?

Mar 11, 2017permalink

(unknown member)

The 4K benefits really are that you can crop the video to create a multiple field of views from same perspective. So get a single camera to record a interview and then get multiple framings as needed.
You can stabilize the footage but only to specific point, and need to be calculated in for field of view.
You get away from partial moire and better colors by this way too.

But it is all a lot time consuming, requires extra planning and considering when to do it.

It is just easier and better mostly get the camera, position as wanted and change focal length (/lens) as wanted and shoot it.

And then comes the opinions about 1080p quality.... Some are never happy and want to see a 4K in form of 1080p.

Mar 11, 2017permalink

Which camera is the 1080p issue with? Sensors up to and including the GH4/ E-M1 generation didn't use all the pixels for 1080p (31% in the 16:9 crop were ignored) so you got issues.

Mar 11, 2017permalink

Stemu

Most of the cameras shoot poor quality 1080p. For example my Lumix GX7 and Olympus EM10 II. Lots of aliasing which makes jagged lines, moire and details dancing. Lumix GH5 and Sony A7S shoot good 1080p but not perfect either compared to 1080p resulution jpeg pictures or video downscaled from good 4k.

Mar 11, 2017*permalink

Unfortunately, cropping 4K video to reframe a scene results (really) inferior quality, I was so disappointed whenever I did this. So, I'd really appreciate an in-body 4K-HD downscaling camera. I used to have an FZ2000, which was able to downscale from 4K to HD over to HDMI, and I was able to record with an Atomos Ninja2. It was really a brilliant option, but it required an external recorder. If you ever worked with 4K, you experienced that downconverting on an i7 computer takes about four times processing time of the actual playtime: that is, a 10 minute 4K video downconversion takes 40 minutes So, we want this processing in-the-camera, right?

Mar 11, 2017*permalink

Miki, are you saying that if you crop a 4K image (to something still higher than 1080) and downconvert to 1080, it's worse than plain 1080? That seems counterintuitive?

Tommi K1, I'm referring to cropping-for-stabilisation in those cases where you don't have the luxury of a tripod etc. I was assuming that the resulting cropped/downconverted-to1080 video would still be higher quality than plain 1080 video, but maybe that's not the case.

And what do you mean about multiple fields of view, just from a higher-res video? What are the use cases for that? For example, that's not going to handle (say) shooting and interviewer and interviewee.

Mar 11, 2017permalink

RPJG, "are you saying that if you crop a 4K image (to something still higher than 1080) and downconvert to 1080, it's worse than plain 1080?" This is just my feeling after working with 4K from my FZ2000 and A5100. The 50Mbps HD from my A5100 is really high quality, when I just cropped out a 1920x1080 portion from a 4K video in my Premiere Elements, it didn't look as great. I need more testing and experimenting to support this observation, though. If you have a 4K camera just give it a try, please, I'd love to learn, too. The well-appreciated 4K UHD (3840 x 2160) video from my iPhone7 is only 50Mbps, the FHD (1920 x 1080) from my A5100 is 50Mbps, too. So, in the 4K video only a quarter of the bitrates are used for a 1920x1080 portion vs the native FHD file; so, I can see it counter-counterintuitive :-)

Mar 11, 2017*permalink

I took the time, since I was interested enough, too. This experimental video shows that https://youtu.be/PHZv8BIHTfc the 50Mbps HD video is (significantly) better quality than the video cropped from a 50Mbps 4K (iPhone7) video. Maybe the situation is different when the 4K is recorded with 200 Mbps. However that would result in hilariously large video files. The 50Mbps image quality from Sony A5100 is absolutely great one of the best HD image quality. I was blown way by the excellent image quality of this movie: https://www.cinema5d.com/exclusivesony-a5100-video-review-footage/

Mar 14, 2017*permalink

Thanks Miki. However, I don't think it's a big surprise that the iPhone footage is nowhere near as good as the "real" camera's footage.

In general, I think we're all discussing the difference between HD and cropped/downscaled-4K from the same camera/lens, in which case the cropped/downscaled video *should* look at least as good, if not better.

Mar 15, 2017permalink

RPJG:"iPhone footage is nowhere near as good as the "real" camera's footage"
Miki: I wouldn't underestimate the iPhone's 4K, it's far the mostly used and popular 4K on Earth today (check out the last part in my demo video); and it is really not bad at all.
RPJG:"Difference between HD and cropped/downscaled-4K from the same camera/lens, in which case the cropped/downscaled video *should* look at least as good, if not better"
Miki: It depends on the size of the crop; if the crop is a 1920x1080 portion from the UHD frame, the resulting video might be (much more) inferior than a 50Mbps HD video recorded with a camera like the A5100. The A5100 50Mbps image quality is stellar, really good enough for most videography purposes. I love 4K, nevertheless, but definitely I wouldn't use it for cropping; this was my practical experience. I'd use proper videography techniques and framing right when shooting; I am really not a fan of post processing, anyway.

Mar 15, 2017*permalink

Sure, I'm not hating on the A5100 at all :-)

But your raw iPhone footage shows that it doesn't hold a candle to the raw A5100 footage, so naturally the cropped/scaled version is going to look even worse.

Somewhere on the interwebs there'll be a like-for-like 1080-vs-downscaled-4K comparison, which will (presumably!) show the (maybe slight) superiority of the downscaled version.

Mar 15, 2017permalink

Franz Weber

Olympus has always been known for the best video quality out there
Incredible 1080p with 24 Mbit/s when others wasted their time with 200Mbit/s and 4K

Mar 11, 2017permalink

photofan1986

What ? You must be talking about panasonic or...

Mar 11, 2017permalink

(unknown member)

Sarcasm.... ;)

And even today we enjoy most of the movies and videos with 720/1080p via 2-6Mbits, while others waste their time with H.264 codec that was designed for 1-6Mbits bandwidth output (not for recording) over slow internet connections.

Consumers don't care... They just want the movie.

Mar 11, 2017permalink

junk1

Maybe they'll shift the sensor back and forth and up and down to get 8k...

Mar 11, 2017permalink

PhotoKhan

On a completly unrelated piece of news, I have just been assured by Mr. Akura Wannabe, the famous Japanese chef, that there is no problem making stinking Stilton cheese out of seagull's droppings.

Mar 11, 2017permalink

Roland Wooster

Yeah, show me 4K/60 first then I might believe 8K is realistic.

Mar 11, 2017permalink

Boissez

Well, the EM1 mkii is capable of processing 20mp frames @ 60 fps. Bumping that up to 32 mp @ 30 fps seems entirely feasible.

Mar 11, 2017permalink

Indeed, Boissez, And processing and storage speeds are only going to improve over the next couple of years.

Mar 11, 2017*permalink

(unknown member)

At the some point a SATA slot is required in these cameras. Really a eSATA or something that allows to attach a SSD and never worry about card speeds for video.

Make a slot and case where you can insert any 2.5" SSD and then that to be inserted to camera or attached with eSATA (case being with external power source) and it is smaller than external recorder, so allowing still gunning and running.

Mar 12, 2017permalink

photomedium

A tide that could lift all boats...meaning 4K would be faster & better.

Mar 11, 2017permalink

trungtran

I did not know Olympus made sensors.

Mar 11, 2017permalink

(unknown member)

They don't fabricate them for mass production.
But they do have own lab to do the lithography and own teams for research and design. The E-M1 Mk2 sensor is designed by Olympus.
Olympus has well lots of patents for sensor technologies and they have many research papers as well.

Olympus main business is medical division, and the camera division is participating to the optics part as well. They share the R&D.

Mar 11, 2017*permalink

I'd still like to see some evidence that Oly's design work on the E-M1 II sensor consisted of more than specifying things like PDAF, as it's quite different to the sensors they have done more serious work on.

Mar 11, 2017permalink

(unknown member)

And it is far more different than what a Sony has, if it is even fabricated by Sony.
We see what comes in GH5 as well.

So far we need to accept to facts:
1) Olympus has a sensor lithography capabilities
2) Olympus has said multiple times it is their designed sensor (for camera thats development started 2011-2012)

But who ever fabricates it for mass amounts is totally another story then and doesn't matter.

Mar 12, 2017permalink

http://ir.ultratech.com/releasedetail.cfm?releaseid=222154

Link showing Oly has a Fab please?

Mar 12, 2017permalink

(unknown member)

"The Olympus order validates the success of our strategy to develop application-specific production tools to meet customers' unique needs in the MEMS/microsystems arena. The NanoTech 160's projection lithography technology was specifically designed to handle a wide range of substrates and materials and deliver superior dual-side alignment capabilities in thick-resist, high-topography conditions. The tool is designed to produce diverse precision products ranging from sensors to sophisticated optical devices."

Mar 12, 2017*permalink

MEMS - completely different to image sensors.

Mar 12, 2017permalink

agachart

yeah,they can design but don't made.
something on design can't made on factory.

i think this is news of panasonic's remade image sensor,why not Sony.
Assure Sony can 't supply 8K sensor on 2018.

Mar 13, 2017permalink

Panasonic is a mega-corporation with a bunch of Semiconductor fabs, including IIRC 4 image sensor fabs, which they sold a half-interest in to Tower Jazz. It's a huge cost buying an up-to-date Fab.

Mar 13, 2017permalink

Joe Niepce

If an 8k sensor can have the same fast readout speed as the EM1 mark II it would be a real evolution.

Mar 11, 2017permalink

Ok, besides the question: How much sense does this make on a m4/3-sensor? (perhaps Olympus has some special techniques like multi-exposure in reserve to solve the "small pixels = more noise" problem):

If Panasonic claims a 6K-mode for its GH5 with 5184 pixels wide, then Olympus needs for its 8K not more than 7000 pixels. Maybe this helps a bit! ;-)

Mar 11, 2017*permalink

Panasonic claimed the 18MP 6k frames were the same as 6k x 3k resolution, so 6k, despite not offering a 2:1 video mode or horizontal resolution above 5k. It's really amazing they said it with a straight face. Then again I guess my Coolpix 950 takes 12k stills, as it's 1600 x 1200, which is the same pixel count as 12,800 x 150 so 12k.

Mar 11, 2017permalink

“we can assure you that there is no problem in developing sensors at 33 million pixels for filming in 8K" - Well as they say, there may be no problem doing it, but how usable would the resulting images/videos be?

Unless this can be overcome with better technology somehow, the problem I see with 8K on a Micro 4/3 camera is how small the pixels will have to be to fit 33MP (or more) on a 4/3 sensor. Smaller pixels mean more noise and less dynamic range, greater diffraction (I think) and currently Micro 4/3 cameras are not known for great dynamic range compared to larger sensor cameras, especially for video.

I am far from an expert in sensor technology, so it will be interesting to see what Olympus eventually comes out with, but right now for M4/3, I'd rather have for video, improved better dynamic range and better colors and low light performance over more resolution, but there is only so much you can do with a sensor that size.

Mar 10, 2017permalink

brycesteiner

It's funny you say this since the best digital 4k video of any system (until you get to RED) is M4/3. Why is that if it's so bad?
We've been hearing naysayers for years say "It can't be done" yet resolution keeps increasing and noise continues to decrease across all the sensors. One can easily take 6400 ISO on m43 and still have good images.

Mar 11, 2017permalink

Right now I think the best camera under the cost of a RED for 4K is the Blackmagic URSA line, but for hybrid cameras I think m4/3 are very good for 4K, especially the GH5.

To record 8K, you need at least 33.2MP and on a M4/3 camera you need to make the pixels smaller than they are now, That does two things as I mentioned above:

1. Dynamic Range. Larger pixels will give you better dynamic range. Many wanted the GH5 to have more dynamic range for video because of harsh highlight rolloff (Better in V-Log L) that the GH4 has. Panasonic said they couldn't increase DR because of the 20.3MP, pixel size they had to use. How adding a lot more MP to a same sized sensor, solve that? Adding BSI?

2. There is only so much in camera noise reduction and in editing you can do to lower noise and smaller pixels don't help when it comes to noise so increasing the pixel count and making them smaller doesn't make less noise.

Hopefully to make a 8K sensor Olympus can find ways around those two issues.

Mar 11, 2017permalink

(unknown member)

If people can buy and watch millions of HD and 4K videos from cameras like the HERO, then sure, no reason why Olympus can't do 8K.

Mar 11, 2017permalink

(unknown member)

Sure it is technically possible. Ie. Even when my smartphone can record 4K 30p, it ain't better looking than a 1080p version. If I downscale the video, then I can get it to look little better, but it is just no worth as it looks anyways good on the TV.

Mar 11, 2017permalink

I have a GH4, but it isn't the best 4k below a Red but quite some margin. Nice though, especially if you keep the ISO down and use an external recorder.

Mar 11, 2017permalink

I am not saying Olympus couldn't do it or that it isn't technically possible if they say it is, and that is fine as I quite like the idea of 8K M4/3 cameras, as over the years camera tech has been getting better and given us the better IQ we wanted at higher resolutions and I can't see that changing any time in the future.

I am just questioning how good the 8K results would be, based on camera tech now as the pixels need to be smaller than they are now to fit them all onto the 4/3 sensor and currently doing so causes issues as I mentioned already and one reason many prefer larger sensor cameras over M4/3 for some of those reasons.

Mar 11, 2017*permalink

JoeWall

And who will watch 8K videos? Nobody who uses a m43 camera has an 8K monitor

Mar 10, 2017permalink

Dale Baskin

Probably nobody. But that's not the point of 8K video. The people who are really excited about 8K video are content creators. The extra resolution will result in better quality HD and 4K rendering, as well as the ability to crop, pan, etc. without any loss in resolution.

There are already 8K monitors and displays (and they are beautiful), but even the companies making them don't anticipate them becoming consumer-oriented products in the near future. (Though I'm sure we'll get there at some point.)

Mar 10, 2017permalink

Dale Baskin

Also, stills from 8K video are really impressive. At NAB last year Canon had a gallery of prints made from 8K frame grabs. They obviously used scenes that translated well into still images, but with over 30MP of resolution in a single frame (roughly Nikon D810 territory) the images were stunning.

Mar 10, 2017permalink

Give me an 8k global shutter camera with some DR and kiss my DSLR good bye.
I already shoot with a 4k global shutter camera that shoots RAW.
If DJI introduces their 5.2k RAW m4/3 camera on a handle my wallet will be in trouble.
In a few short years SLRs as we know them as still centric cameras will be history.
It will be the other way around. Video cameras that can do stills.

Mar 10, 2017permalink

yortuk

That's already happened in phone cameras -- the Google Pixel camera is constantly recording by default and then just combines a few frames when you press the shutter and saves it as an HDR image. It has its limitations, but the results are usually better than I would have thought possible for such a small sensor. Applying the same approach with a m4/3 sensor could be really impressive.

Mar 11, 2017permalink

yortuk, that's true, but surely that always-recording approach would just kill battery life (even mnore) on lager-sensor camera like m4/3 and up? One day, we'll get that huge leap in battery performance that so much depends on...

Mar 11, 2017permalink

(unknown member)

First, m4/3 ain't the sensor but mount. 4/3" is the sensor.
Secondly, E-M1 Mk2 already has this "always-recording" in a limited manner.
Enable the "Pro-Capture" mode and the camera will buffer 14 frames at your selected frame rate (like 60fps) from the moment you half-press the shutter. Then when you fully press the shutter, those 14 frames are stored to card and you can select the missed moment. Pro-Capture H is S-AF/MF limited but gives you that 60FPS for 14 frames. Pro-Capture L is C-AF capable but gives only a 15 or 18FPS for 14 frames.

Mar 11, 2017permalink

I really don't think 8k will stop people shooting stills unless you can have Raw video, which would be an insane bandwidth. Current 6k photo is 210Mbit/s for 25 frames/s, which is basically the same amount of data as a single Raw from the same camera.

Mar 11, 2017permalink

yortuk

RPJG, it's not just the battery capacity, but also the efficiency of the sensor and processing. That's where a lot of the advances have been coming lately, particularly with Sony's stacked sensors. Also, with live view displays we're doing a certain amount of continuous readout already.

Mar 11, 2017permalink

RAW video can be compressed.
The bandwidth isn't that bad.
SSDs and hard drives are cheap.
I have 2 cameras that shoot RAW video.

Mar 11, 2017permalink

yortuk

Tommi, the big difference with the Google Pixel is the automatic processing it does to combine the frames to get a higher quality image than is possible with a single frame. This is possible in other cameras with post-processing or even in-camera options, but requires the user to make those decisions and is not a seamless, automatic process. The point is that computer processing can really do a lot to increase the image quality that's possible with a given sensor, and it's becoming more automated.

Mar 11, 2017permalink

FLruckas - remember we're talking 8k video here...
Also lossless compression doesn't save you that much. Lossy compression helps, but you lose stuff.
Sony F65 4k Raw is 2Gb/s, so at 4x the data 8k Raw would be 1GB/s (8Gb), which strikes me as a lot. Canon 12-bit 4k Raw is 2.3Gb/s.
The 8k version of the F65 they showed a long time back was 5Gb/s IIRC.

Mar 11, 2017permalink

With 4k lossless raw I get about 1TB an hour.
And that's not with the fastest SSD.
I've tested cheap M2 SSDs and get about 3TB an hour.
So we're not really that far off from supporting good 8k in affordable devices.
DJI is already shipping a 5.2k RAW camera with M2 recording.

Mar 12, 2017permalink

Tommi, pedantry regarding sensor/mount names aside, I realise that the Mk2 has this feature, The massive difference is that it only happens when you ask for it; it's not the battery-sucking always-on feature that was being discussed.

yortuk, sensor/processing efficieny is always improving, but at the same time they're being asked to do more, e.g. processing more pixels at higher bit depth at faster frame rates and faster storage speeds etc. So that *roughly* balances out. Hence, I think an always-on mode for a large-sensor camera with large EVFs/displays will need much-higher battery capacity than we have today, unless other features are allowed to stagnate.

Mar 12, 2017permalink

FLruckas - thanks, although 1TB/hr is a huge amount of data even before you multiply by 4 for 8k. It works out as 2.2Gb/s, so basically in agreement with my Canon Raw figure. I'm not sure non-Pros will want much to do with 8k if they get 1.1GB (8.8Gb) of data per second, especially for burst mode photos. My Atomos has a 540GB SSD in it, which is 8 minutes of 8k Raw video for about a hundred quid of storage (vs. 6h 52m of Prores 422 4k).

Mar 12, 2017permalink

(Oops - 2.5 hrs of ProRes 422, forgot it showed 1080p capacity remaining by default. I should see if I can change that...)

Mar 12, 2017permalink

garyknrd

Olympus will be the next 800lb. gorilla in the room IMO.

Mar 10, 2017permalink

But they are contracting and selling less cameras?

Mar 11, 2017permalink

barrym1966

Cool working on 8k but no longer making any new lenses?

Mar 10, 2017permalink

Dale Baskin

Just to clarify, Olympus has not stopped making lenses. It has discontinued its original line of Four Thirds lenses, but not Micro Four Thirds lenses for mirrorless cameras.

Mar 10, 2017permalink

barrym1966

ok that makes more sense :)

Mar 12, 2017permalink

BostonC

Akira Watanabe: "12mp is, I think, enough for covering most applications most customers need" - said in 2009. He knows his stuff, he is the designer of EP1.
How time changes perspectives. The cam manufacturers have to run for the hills, to stay ahead of the smartphone flood. A M43 sensors w high pixel count is one way to beat it.
8k in 16:9, the sensor would be better be in multi-aspect format, so the 4:3 portion is more manageable at 37mp w pixel pitch ~2.4µm--exactly the same as on Sony 20mp 1" sensor.
If not using multi-aspect, the M43 sensor needs to have 44mp.

Mar 10, 2017permalink

myn

I think the quote still holds. 4000×3000 px covers UHD video resolution, can be printed sufficiently big, so for some foreseeable time it will be enough for covering most applications most customers need. The question of selling cameras is fulfilling what people want, and this results in better and better specifications.

Mar 11, 2017permalink

(unknown member)

How many megapixels do you need for an web use?
How many megapixels do you need for an most common print sizes (up to A3)?
How many megapixels do you need for most common photo usage?

A way less than 12Mpix in most cases.

Mar 11, 2017permalink

Depends on how much you want to crop...

Mar 11, 2017permalink

BostonC

Higher mp count helps cropping obviously. I argue that viewing habit has changed. Most ppl use smartphones/devices that feature screen zoom by finger pinch. Some of you seem not aware the feature is used very often, much more so than cropping. Newer phones/devices have 5X magnification. Only high res images can hold up.

Mar 11, 2017permalink

myn

Sure, more pixels are better for cropping. But I think in many cases people use pinch to zoom because there are practically only wide angle lenses on their phones and the cameras have enough resolution at those 12 MP.

Mar 11, 2017permalink

falconeyes

An 8k sensor in 4:3 aspect ratio has 44.2+ MP. µFT won't be the most economical form factor for such resolutions, taking lens cost into account (because of the excessive resolution requirement).

Mar 10, 2017*permalink

Current m43rds lenses can quite easily use 40mp. The High res mode is a great stress test for resolving power, and they can resolve as much as a Pentax 645D... I know it seems hard to believe, but it is true.

I suspect it is the cost of the processing power that holds them back.

Demonstration of the high res mode vs the pentax.
https://www.dpreview.com/forums/post/58984009

Mar 10, 2017permalink

(unknown member)

Combining several pictures is not the same that taking 1 high resolution picture, so that high ress mode say nothing about if the lens is capable or not. Besides, people forget that 44mp on m43 sensor size means that the difraction threshold is at f4,4, so a large amount of the current m43 lens lineup will not be able to ever achive that sensor potential because the resolution will start to get nerfed by diffraction even wide open.

Mar 10, 2017permalink

Hi-Res mode increases sampling density by shifting the sensor. If you increase sampling density, then it doesn't matter if you take all the samples at once, or half the samples at one time then the other half at a later time.

Mar 10, 2017*permalink

Tomi,
It is exactly like a high resolution shot. As Barty points out.

But don't take my work for it, look at the samples I linked to. at f5.6 (DPRs aperture setting) the Pen F out resolves the pentax 645D.

If you are wondering why DPR tens to not notice, they haven't really explored the function as well as someone like myself (who shot 10 commercial images today alone with it).

Most people would probably be surprised how much resolving power most modern lenses have today, especially stopped down.

Mar 11, 2017permalink

(unknown member)

That has nothing to do with 8k video because when you take video you NEED ALL THE SAMPLES AT ONCE and most likely at least at 30fps. Copy from Em 1 MkII review: "In High Res mode the camera shifts the sensor eight times, taking a photo in every position, and combines them into a single 50 or 25 Megapixel image". The camera take 8 low resolution pictures and then it combine them, but it never do 1 single high resolution picture. So the answer is NO, having a high res mode doesnt mean that any of the current lens is able to deliver the full potential of a 44Mp sensor. If you take a look at DxO you will see that most of the m43 lens don't reach the 60% potential of current 20Mp sensor (also the one that get there are very packed around the 25mm, so not that much diversity), so they are really far from reaching full potential of a 44Mp.

Mar 11, 2017permalink

Tomi,
You keep saying the lenses cant resolve 44mp. Yet I have shown you the PEN F with 45mm f1.8 out resolving the PENTAX 645D... Clearly you not believing your lying eyes.

If the lenses couldn't resolve the potential (in my sample example it was an 80mp raw with the pentax upressed to the 80mp)... They simply would have looked like mush, no matter how many shots you took and blended.

Now DXO mark has a few quirks most people don't understand. Their final score is an average across the range of available apertures. As m43rds lenses often go to F22 (f64 equiv on FF) this reduces the average resulting in a lower score. Combine this with the fact that there is only currently a 20mp sensor available is a double whammy.

But again, we can test this. DxO rates the 45mm f1.8 at 11mp on the EM1.2, and the 85mm f1.8 on the nikon D610 (21mp) as 19mp... Yet DPR contradicts this showing the D610 losing out to the PENF:
https://tinyurl.com/hs2s6r3

Mar 11, 2017permalink

(unknown member)

You didn't show that at all. You just show that after good post proccesing you can merge 8 files into a single bigger one. Thats not resolve nor best, that just show you how a good post process can do starting with 8 different pictures that are taken at 8 different times with a low time lapse between them. Now even when the time lapse is small all the pictures are still different pcitures and not one single picture, so NOOOOOOOOOOOOOOOOOO the PEN F can't out resolve the Pentax on a single picture with a proper lens. Your thread just don't show that because you just can't do that. You can repeat and post your results many times and they will continue to be 8 different pictures that were merged all together but each single picture will not outresolve the Pentax.

Basically, you need 8 Oly EM 1 MKII to try to make for a Pentax thats what your thread prove.

Mar 11, 2017permalink

https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=nikon_d610&attr13_1=olympus_penf&attr13_2=sigma_dp1m&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=100&attr16_1=200&attr16_2=100&attr16_3=100&attr126_1=normal&attr171_1=off&normalization=full&widget=28&x=0.8191165489114827&y=0.6650299401197604

Mar 11, 2017permalink

falconeyes

Hi everybody. My point actually wasn't that mFT cannot resolve 44MP. But that it isn't economical at that. The total cost (with the ever decreasing sensor cost) of body plus lens will be higher than, e.g., full frame. You see that already today with the best FT lenses.

Also, there is a difference if a lens resolves 44MP in the center at an ideal aperture, or in the typical case. Moreover, superresolution IS a lens stress test, yes, but I know of no correct comparison (MTF) with single frame tests. One caveat is sharpening which always happens as part of processing superresolution but not normally in a standard MTF test. Another caveat are non equivalent lens settings. A correct test would have to be against like an Otus or Art lens on a 42 or 50 MP body.

Mar 11, 2017permalink

(unknown member)

Tomi. It doesn't matter. The function of HR mode is exactly same as the sensor would be a 80Mpix (using 20Mpix 4/3" sensor).

The lens needs to deliver the sharpness. No matter how much you move the sensor or put a high res sensor behind it as you get soft image if you have soft lens. You need to resolve the difference first.

Resolution means what? It doesn't mean how many pixels there is in the sensor. It means how much can be resolved.

If I take 50 shots with 16Mpix from same position and stack the shots. I am not increasing the resolution, I am removing the noise. In a hence spirit I am increasing the dynamic range as dynamic range is noise limited as well. I get much sharper photo, but the resolution is still 16Mpix, with huge increase in detail, good increase in dynamic range and a huge decrease in noise.

Now if I would divide the view to 4 parts, pan the camera and shoot 15 frames of each parts and then stack + stitch the photos. I have all above + increase in resolution.

Mar 11, 2017permalink

Again Tomi.
You might need 8 shots from an em1 to out resolve a Pentax 645d but that is a 20mp sensor. If you had a 50 no sensor what it does prove is the lenses can cope just fine... despite what you believe.

The issue I suspect is processing and battery power.

So carry on being now informed if you like it actually expand your understanding of new imaging techniques and their applications.

Mar 11, 2017permalink

(unknown member)

Tomi K1, you are again wrong. The HR mode take 8 different pictures taken a diferent sites but very close one to the other because the sensor shift but not by much, so basically each single pixel of the end results is looked thru 8 different sectors on the lens, then the camera compare the 8 pictures and decide which one is more accurate or just merge the 8 pictures.
Taking 8 pictures and merge them doens't mean one lens can resolve 80Mp in one single shot, because the camera is using 8 different sectors of the lens, so its good post processing and not a lens resolving issue.

Mar 11, 2017permalink

I can't believe the misinformation you are writing here.

I actually use the system every day and the high res mode. I have also done tests to show the better the lens, the better the result.

You just keep writing nonsense. If the lens can't resolve 50mp. It won't. Regardless of how many shots you take resolution is limited by the weakest link in the imaging chain.no magic will improve on that.

You should try to understand, especially when you have no experience using a particular technology.

Mar 11, 2017permalink

Revenant

The hi-res mode takes four exposures shifted one pixel-width horizontally and vertically. The sensor is then shifted half a pixel-width to a new starting position, and then takes four new exposures shifted one pixel-width horizontally and vertically. This results in both full colour information for every pixel position (because of the four exposures), as well as the capture of more detail (because of the half pixel-width shift).

However, you can only capture more detail, if in fact the image cast on the sensor by the lens contains that detail. So, if the hi-res mode does result in images showing more real detail (and it does), this means that the lens is resolving that detail. No amount of clever processing can reveal detail that isn't resolved by the lens.

Mar 11, 2017permalink

(unknown member)

Its resolved after you move the sensor to 8 different positions and you recomposed the image with those 8 images at different places, but as I've told you before none of those 8 images needs the lens to resolve a 80Mp. As you express it very clear, in order to get that image you need to oversample that same spot 8 times and after the camera watch the same spot from 8 different angles it decide what is going on there.
Anyway, just to make your things more complicated the m43 are designed to use a central part with 4:3 ratio and video is taken at 16:9, which means that you are going to need 27Mp sensors for 6k (6000x4500 to get the 4:3) or 48MP for 8k (8000x6000), if you actually believe that Olympus is close to that then I can just aplaud your optimism. Even if they get those sensors then the difraction starting at f5.6 for 6k and at f4.5 for 8k will remain and that issue will make a really large amount of the current m43 lens line up just not ready for that new technology.

Mar 11, 2017*permalink

"I actually use the system every day and the high res mode. I have also done tests to show the better the lens, the better the result."
This doesn't prove your claim. Of course better lens will resolve better. Because a better lens will also resolve better when it is one single image,duh! Using technology doesn't mean you understand the science behind it.
Tomi O is right and pretty logical.
In every image you took by using HR mode, the lens resolves ONLY 20 mp each time (for E-M1 Mark II). But since the sensor is moving slightly, every time it matches other fine details from the scene.

Mar 12, 2017permalink

teslacoil

Tomi O and Xentinus, I don't think you understand how the sensor shift technology works. The sensor shifts a miniscule amount... Within the same image circle projected by the stationary lens. The sensor can't resolve detail that isn't projected by the lens. If you move the lens and combine the shots, the lens only needs to project 20 MP of data each time. But if you keep the lens stationary and move the sensor within the same image circle, you cannot resolve detail that the lens couldn't resolve given a sufficiently high resolution sensor.

You can't invent data that isn't already there. The hi-res mode doesn't just uprezz the image, it actually resolves more detail than a single shot can. And if the same image circle can resolve 80 MP by shifting a 20 MP sensor around, then it can also resolve 80 MP in a single image if you place a 80 MP sensor behind it.

Mar 12, 2017permalink

https://www.dpreview.com/reviews/olympus-pen-f/10

You guys have a claim, but all you can do is to say "HR 50 mp, so the lens can resolve 50 mp". No physical or logical evidence.
In HR mode, every image is just 20 mp. I don't understand how you can claim that; the lens can resolve 50/80 mp, with simply zero evidence.

I will try to write logical evidences down.

-When Nikon released 36 megapixel D800, Zeiss announced Otus 1.4/55 in 2013. Because they knew to resolve 36 mp sensor, requires new lenses.
But you claim an Olympus 45mm 1.8 (which is announced in 2011) can already resolve 50/80 mp.
-"The other advantage towards this pixel-shift mode is it also reduces noise. At ISO 1600, the Olympus shows slightly less noise when compared to the a7R II (despite the Sony's much larger sensor)" DPR says in this page;

So would you also claim that a 50/80 mp 4/3 sensor can give you clean output at high ISO?

...................................

Mar 12, 2017*permalink

...
-Why modern lenses are pretty big, comparing old prime lenses? Why companies started to produce telecentric lenses?

Mar 12, 2017permalink

(unknown member)

The m43 fans are funny, they claim that a lens that is already giving just 55% of a 16Mp with or without AA filter and stay at 55% when it moves to a 20MP is going to miraculously fully resolve a 50Mp sensor. Just check the results at DxO and compare the Tamron 85 1.8 VC agaisnt the Panasonic 42.5 1.2 (the sharpest of the m43 lenses). The Tamron fully resolve the 36MP on the D810 and any other FX sensor that doesn't have AA filter and it drop only to 90-95% when it move to the high pixel density DX cameras, meanwhile the Pana that cost 2x more just reach 80% on any modern m43 camera without AA filter.

Mar 12, 2017permalink

The only thing that is funny here is the persistence of those with little to no experience and a poor understanding of spacial resolution being so adamant, to the point of stupidity, and clinging to a position that is... with a little thought, clearly WRONG.

I am chiming in one last time. So can the PEN F resolve more than a Pentax 645 Z in a single image. NO. Noone claimed that. What is clear is that with the 45mm f1.8 in HR mode the PEN can out resolve the 645Z.

The question raised was if the 43rds lenses can resolve 40mp. Well, the answer is clearly YES because the HR mode shows MORE detail from the PEN at 80mp than the 645Z.

This is not the same as stitching as the first step moves the sensor ONE pixel. Leaving FULL RGB colour in 20mp (like FOVEON), THEN it shifts is HALF a pixel, this samples for detail that IF NOT THERE BETWEEN the pixels of a 20MP shot, you would only get a soft looking 20mp shot.

So please stop taking about things you have not thought through properly.

Mar 12, 2017permalink

Tomi, you also clearly dont understand DxO's lens measurement. So please stop invoking them.

Mar 12, 2017permalink

https://www.dpreview.com/forums/post/58984009

Xentinus, There are a number of reasons for larger lenses. The first is resolving power wide open. Of course, the m43rds lenses probably can't resolve 40mp wide open, but neither can most lenses available from most mounts... Hence the re-release as MP get higher.

The second is optical corrections. the m43rds system relies on digital correction of some elements (mostly CA and distortion). This works as they can correct for the distortion in the EVF, while on DSLRs and an optical finder it would not be pleasant viwing every image with a lot of barrely distortion.

Did you look at the samples I linked to? Stopped to f5.6 the 45mm f1.8 can quite obviously resolve MORE than the 50mp 645Z... in case you missed it.. here is the link again:

Mar 12, 2017permalink

Tomi O,
What you have just said, is a physical evidence. Beside that, it is easier to prove via Pentax cameras and lenses. Because Pentax has a high resolution full-frame sensor, while there is also APS-C sensor with lower resolution.
For example;
43mm 1.9 Pentax Lens + K-1 VS 43mm + K-3 Mark II VS 43mm + K-3 Mark II HR mode
IF*** the result with 43+K3 HR mode shows more details comparing 43+ K1, the question will be simple:
why 43mm can not resolve 36 megapixel?

Mar 12, 2017permalink

Ab Latchin,
Thank you for the link.
Above I sent the link which shows amazing resolving power of foveon sensor!
Foveon sensor.
Again foveon SENSOR. SENSOR...See the thing which shows amazing details is not lens! Sensor!
Did you finally understand? You are insulting people but you can not understand one simple thing!
HR mode works just like that.
http://www.imaging-resource.com/PRODS/pentax-k1/ZPIXELSHIFTRES.png

So again it shows more details comparing full-frame or even medium format (in the link you have shared above). This doesn't prove it can resolve 50mp!

Mar 12, 2017permalink

"I am chiming in one last time. So can the PEN F resolve more than a Pentax 645 Z in a single image. NO. Noone claimed that. What is clear is that with the 45mm f1.8 in HR mode the PEN can out resolve the 645Z."

The thing we are discussing is not this. Don't twist the words!

The question is, IF it was a 50mp sensor, would 45mm 1.8 resolve? the answer is NO!

Mar 12, 2017permalink

Xentinus, the foveon sensor wouldn't be able to resolve the detail it demonstrates WITHOUT the lens permitting it.

Everything projected onto the sensor first has to pass through the lens. If the lens cant, the sensor wont show high detail.

Then please explain HOW, if the 45 1.8 CANNOT resolve 50mp, it actually does? Because the sensor cant record what the lens didn't project.

Mar 12, 2017permalink

Foveon shows the true detail of the lens, not robbed of the de-mosiac process from the bayer pattern.

This does the same, and ON TOP OF THAT does a half pixel shift sub sampling the detail.

Look at the Leica Monochrome sensor for what type of detail is missing with a bayer filter array. That shows the potential resolving power of the lens without even increasing the sensor resolution.

Mar 12, 2017*permalink

http://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm

In HR mode, the lens has to resolve less pixel for every single image (comparing 50mp native sensor). But since the sensor shifts, it obtains different information for each image. Then processor does the job.
The logic behind the foveon sensor is similar.In HR mode, the lens doesn't have to resolve all pixels AT ONCE.
Really claiming that "It shows amazing details in HR mode, so it can handle 50mp sensor too" is silly (Im being nice).

Mar 12, 2017permalink

(unknown member)

AB Lachin, do you even understand that if a m43 lens needs to be stoped down to f5.6 to try to resolve a 50Mp sensor the lens is going to fail because the difraction start affecting the image 1 stop before that? The situation you have in your mind is not even possible with a perfect lens. Please, give a read to this and take a look at the difraction section:

You pretend to understand better than anyone else DxO scores, but you miss the point that the 45 1.8 resolve only 11 of the 20MP on the Oly Em1 MkII at f2,8 (the best setting for sharpness on that lens) and it goes below 10 at any other setting. Anyway, there is no point to continue this discussion until m43 get a 50Mp sensor, which is very likely that is never going to happen.

Mar 12, 2017permalink

Xentinus,
Lets approach this logically. Bayer 50mp has 3 to 4 pixels (approximately) making up ONE pixel of the final RGB image. Can we agree on this?

By using a Foveon sensor you NO LONGER have to use 3-4 pixels to make that final 1 pixel. This is why Foveon presents an actual MP and then an equivalent to bayer resolution. Give or take a few MP they are correct. a 20mp Foveon massively out resolves a 20mp Bayer. However for the lens, instead of a point of detail covering 4 pixels, it now must resolve to 1 pixel in width (for the same sensor size). Can we agree on this? This puts a greater resolving demand on the lens, right?

Now HR mode does this with the 1 pixel shift. Putting more demand on the lens to resolve to 1 pixel. It then shifts the sensor half a pixel, now occupying the space between the pixels for the previous 4 captures. This asks the lens to resolve even finer detail as the lens has not moved, and neither has the scene.

Cont.

Mar 12, 2017permalink

Cont. So the lens is being asked to resolve between the original pixels by the sub pixel shift.

This is now looking for detail between the pixels.

If you were to simply place a 40mp foveon, it would look very similar to this technique.

If the lens couldn't do it for the HR mode, it couldn't do it for foveon, and it couldn;t do it for 50mp Bayer.

But IT CAN do it for HR, so it CAN do it for Foveon, and it WOULD do it for Bayer.

Mar 12, 2017permalink

Tomi, DIffraction isn't a wall you simply hit. It is a gradual reduction of resolution.

You keep talking about diffraction like some limit. It may creep in, but at f5.6 it isn;t a problem CLEARLY. Because AGAIN if you look at my samples f5.6 using HR mode resolved more detail than the pentax 645Z... So it isn;t problem you are making it out to be.

The problem here is you don't seem to understand that HR mode demonstrates REAL resolution.

This should make you feel good, stopped down, almost all modern lenses are superb. The reason manufacturers keep making them bigger is for resolving power wide open.

Mar 12, 2017permalink

(unknown member)

Sorry to be rude AB Latchin, but if you don't understand that the diffraction is an effect that appear at the moment you take the picture as a combination of aperture and pixel size on the intitial low resolution pictures and not when you combine them, then you are a complete ignorant about why the difraction affect the pictures.

Your HR picture tell you absolutly nothing about diffraction unless you reach the difraction limit on each of the single 8 pictures you used to get make the final HR. If you would had a minimum idea about what are you talking about you should know that.

Mar 12, 2017permalink

Tomi, I dont think you are rude. Just horribly mis-informed. Despite multiple people explaining this to you over and over again you keep repeating the same nonsense, then changing the subject

All formats are aperture limited for the SAME DoF. So if 20mp on m43rds is affected by aperture at f8, then FF at 20mp would be at F16 (2 stops, shocking, relative to the sensor size and focal length relationship).

40mp on m43rds isn't MORE limited than 20mp, it is simply going to produce less pixel level detail WHEN VIEWED AT 100%. However if you resize the 40mp image to the 20mp image it will appear sharper and cleaner than the original 20mp shot.

HR mode is EXACTLY limited with the same relationship as a standard lens on a standard sensor. And I have tested this using manual lenses.. Have you?

A 40mp FF sensor shows more detail than a 24mp FF, and so would a 40mp m43rds sensor.

Try to learn something for crying out loud.

Mar 12, 2017permalink

https://www.youtube.com/watch?v=yyOmgArU0MA&feature=youtu.be

To test something scientifically,at first you can change just one parameter each time. Once you observed the correlation between 2 parameters, then you can change others.
Apples to apples please. Foveon sensor technology (despite its flaws), lets Sigma to show more detail, even if the lens is not close to be perfect. It is the SENSOR's structure which obtain the information of the scene.
Of course the lens has its own part too. But if you put a Sigma 50mm 1.4 A lens on a foveon body and compare it to smilar full-frame Canikon counterpart with the exact same lens, the result will tell you that you are completely wrong.

This video will give you ANOTHER logical evidence. But if you still can not understand, I have nothing else to do. I rest my case ^^

notice 6:25 VS 9:35!

This is the same thing what Tomi O said above. This time about Otus' results!

Mar 13, 2017*permalink

Xentinus,
YOu yourself admit in your comment above that Foveon reveals more detail from the lens. But in your previous comment you say "the SENSOR THE SENSOR".

You just don't understand what this HR mode does. It is exactly like a high resolution sensor. The only difference, and the reason the PEF F looks better than the Pentax 645 is we get a full colour 40mp file while the pentax is a bayer 50mp sensor. So when we enlarge the Pentax to the Pen RAW size, the resolution loss from the bayer array jumps out.

I don't need to argue this any more. This is a new tech that clearly neither of you own, but seem to care so much about that you want to academically (but with no experience) explain how it works.

In real life, it would be more helpful to your development if you listen to those with experience in something you have not tried.

Mar 13, 2017*permalink

Ab Latchin,
Yes, I already knew. It is foveon sensor which shows incredible detail.
I also said; if you attach the exact same Sigma lens to a Canikon high resolution camera, it will NOT able to show the same amount of details. This alone proves everything. There is no conflict in my words.

These are *your* words;

"You just don't understand what this HR mode does. It is exactly like a high resolution sensor. The only difference, and the reason the PEF F looks better than the Pentax 645 is we get a full colour 40mp file while the pentax is a bayer 50mp sensor. So when we enlarge the Pentax to the Pen RAW size, the resolution loss from the bayer array jumps out."

Did you notice something? You are talking about sensor and the sensor only. Your observations are all about the sensor, but somehow your conclusion is about the lens.
Suddenly you say "so this 45mm lens can resolve if it was 50mp sensor". You still mean bayer sensor of course. Cause Olympus doesn't use foveon sensor.

Mar 14, 2017permalink

Sergey Borachev

TALK IS CHEAP!

😄

Mar 10, 2017permalink

Suntan

Exactly what is needed in the next pocketable camera...

Mar 10, 2017permalink

ttran88

It's probably reasonable to expect $2500 or more for next EM1.3. Hooray....

Mar 10, 2017permalink

aramgrg

“we can assure you that there is no problem in developing sensors at 33 million pixels for filming in 8K"
Interesting... I though it is not you need to assure that, but Panasonic or Sony. This is simple cheap talk from Oly.

Mar 10, 2017permalink

Androole

Yeah, Panasonic announced a few years that they were restarted their sensor development to focus on 8K for 2020 (Olympics?). So it is pretty clear that Olympus would be a beneficiary of that development...

Mar 10, 2017permalink

Revenant

Olympus could very well be involved in the development of such a sensor, even if they don't have the manufacturing capability themselves.

Nikon designs some of the sensors they use (eg. for the D3/D4/D5), but outsource the production (they used to be made by Renesas). OmniVision doesn't have any fabs of their own, and neither did Aptina (before it was acquired by another company).

Mar 10, 2017permalink

Mister Joseph

Means easy, mature, smoother 4k

Mar 10, 2017permalink

Let's all say in unison....

Nyquist limit.....

:-)

Mar 10, 2017permalink

Why?

Mar 10, 2017permalink

Google....
:-)

Mar 10, 2017permalink

Is not an answer. What are you sampling at the Nyquist rate, and why is it relevant to this article?

Mar 10, 2017permalink

8k silly....
.......
The dots gotta be really close.....
:-)

Mar 10, 2017permalink

At 33MP on m43, the pixels will be approximately 2.6 microns. What 5.2 micron or greater thing are you imaging to make Nyquist relevant?

Mar 10, 2017permalink

I'm just making fun of the fact that years ago people said the nyquist limit was a lot less than 30mp for m4/3 and for some strange reason the nyquist limit has magically been repealed.
And back then it was a big deal.
Just poking fun at what seems to be the sound barrier at any given technology time.
Now we have 50MP FF cameras that are clearly over the limit, and no one is talking about it.

Mar 10, 2017permalink

I think you want Airy, not Nyquist. How much light gets blurred is purely down to the f-ratio, so small pixels really need lenses that are fast and sharp wide open.
At f2.8 the blur is 3.7um, at f8 it's 10.7um.

(You always get some diffraction blurring, but there is a point at which it becomes annoying. Also if you use the whole image and don't print insanely large for people viewing up close then you don't care, but you don't care about more than the teens of MP then either. Personally with an AA filter I'm happy up to two pixel widths worth of diffraction blur, a bit less without.)

Mar 11, 2017*permalink

I was on the point of upgrading from Nikon V1 to an Olympus (I loved PEN E-P5) in 2013 when Sony suddenly announced the mirrorless full frame A7, and I jumped over to it, Never looked back. In the meantime even Fujifilm became more popular than Olympus. Olympus once had a great range of film cameras: the original PEN F was the first APS-C, OM was full frame. One of my favorite lenses is the PEN F 38mm/f1.8 on my Sony, my daughter's favorite is the OM 55mm/f1.2 on her Sony. Still, if the new (2016) PEN F had 4K and PDAF sensor, I'd have jumped on it, I guess.

Mar 10, 2017permalink

Androole

"In the meantime even Fujifilm became more popular than Olympus. "

Citation required?

The internet is a bit of an echo chamber and empirical data is hard to find, but BCN Rankings at least have some objective methodology and a little bit of credibility. (https://www.bcnaward.jp/award/section/hard/hard94.html) They track sales from most large Japanese retailers, and they put the current rankings in mirrorless with Olympus as the market share leader (typically trading with Sony), and with Canon rounding out the top 3. Below that is Panasonic, Fuji, Nikon, Leica, and Pentax all with less than 18% market share.

Basically, at the very best, Fuji has a 50% smaller market share than Olympus. But likely much less than that. Fuji might manage to tie Panasonic for 4th place, if they are lucky.

A lot of enthusiasts like Fuji cameras, but that shouldn't be confused with popularity when the wallets actually come out.

Mar 10, 2017*permalink

(unknown member)

All I see is just some scribbling there and numerics:

オリンパス26.8％) キヤノン(18.5％) ソニー(17.9％)

So can't you give some respectful source that knows how to write?

(sarcasm).

Mar 11, 2017permalink

That;s only for sales inside Japan. World-wide Oly are third behind Sony and Canon (although above Olympus) for mirrorless sales.

Mar 11, 2017permalink

Androole's comment made my comment the first on the Most Popular tab, thank you! The only thing I can say to support my statement is that we had in my city two major photography gear expo/fair and the biggest area was for Canon, second Nikon, third Fujifilm (a huge area, I was surprised), Sony and DJI about the same size as Fuji. All others: Olympus, Panasonic, Pentax, Leica only a desk with one or two guys behind. The number of visitors were more or less proportional to the size of the show area. Fuji and Sony desks were heavily overcrowded, though. We have 5 dedicated Sony Showrooms/Centers in my city and only one Olympus.

Mar 11, 2017*permalink

pkcpga

You do realize that Japan only made up 12% of the global market last year and that was again a shrinking number from 14% the year before. And Japan is one of the few markets were mirrorless sells similar to mirrored. Globally dslr sell 2.6-2.7 times more than mirrorless. So Japan's market is no longer a good indicator as it once was a decade ago. The Asian camera market is shrinking while the European market is over double Japan's and US also double Japan's market share globally and Europe and US purchase different products than Japan. You can always check out Amazon rankings monthly for your country to see what's popular there currently. In the US I can tell you in the largest, highest volume camera store in the US, B and H, Olympus has a very small section in the retail store compared to Sony or canon or Nikon. Olympus, Fuji, Panasonic and Pentax combined barely equal to one of those other three brand space within the retail store. On a more interesting note, the compact camera area seems to be shrinking even more to expand the Medium/large format camera area. Shows which differection the markets heading.

Mar 12, 2017permalink

tinternaut

As long as the sensor technology is either the same or slightly better, I don't see a problem with pushing the pixels. Not sure I see the point of 8K though....

Mar 10, 2017permalink

No more need for stills cameras.....
30 or more FPS.

Mar 10, 2017permalink

Revenant

Extracting stills from video would take all the fun out of photography for me. YMMV.

Mar 10, 2017permalink

2 Rev...
Not me...
That's why I shoot with a video camera that has a global shutter and shoots RAW.
But what works for you works for you.
I'm good with that.

Mar 10, 2017permalink

(unknown member)

Take stills from a video camera, that is limited to 1/25-30 shutter speed?

Yeah... sure... The motion is a tricky thing...

I take anyday the still from that camera, with capability to do 30fps bursts for a 10 frames.

Mar 11, 2017permalink

princecody

Why can't Olympus & Panasonic have interchangeable firmware? This is micro four thirds after all. 🤔

Mar 10, 2017permalink

dinoSnake

Because firmware is low-level hardware (CPU, supporting systems), not attribute (shape and size of chassis, sensor, screen, etc), dependent. Micro Four Thirds is a specification of the sensor, the lenses and the interplay between those two, plus some other things like TTL flash communication (at least, between Panasonic and Olympus sharing the technology). The guts inside each and every camera is different, meaning different firmware is required.

The only way to implement what you suggest is to virtualize the hardware via a HAL (hardware abstraction layer). In other words create a VM (virtual machine) that actually controls the hardware at a base level ("bare metal VM"), then have the firmware act as an interchangeable OS above the VM. The problem? More heavy lifting for the CPU, therefore a performance hit.

So that's why they stick to bare metal firmware (firmware that directly controls the hardware) - it means no interchangeability but it means absolute best performance.

Mar 10, 2017*permalink

Glen Barrington

You might as well just have some Chinese OEM manufacturer design and build cameras for both companies and then slap 2 different labels on the generic box.

The whole point of a 4/3s system is to share certain manufacturing and development costs, but still be able to COMPETE with each other on a technology basis, and with the rest of the industry as well.

Mar 10, 2017*permalink

princecody

EM1 M3=$2700 🤔

Mar 10, 2017permalink

In Australia the MkII is $2700, in New Zealand $3700.

Mar 10, 2017permalink

Franz Weber

The decision to price the MKII in New Zealand so high was made by Olympus to prevent the New Zealanders from buying the whole stock of the world production of this cam as they did with the MKI.
There has been a terible shortage back then.

Mar 11, 2017permalink

justmeMN

Perhaps 8K will be included in a $4,000 EM1 Mark III. :-)

Mar 10, 2017permalink

Georgeee

33MP for single shot. handheld high-resolution image mode . beat FF

Mar 10, 2017permalink

NZ Scott

Problem is that pixels are so small you have to constantly shoot at high shutter speed.

Mar 10, 2017permalink

Glen Barrington

Not if you make the m43s sensor 4 inches by 5 . . . just sayin'

Mar 10, 2017permalink

Vadimka

no you don't, all you need is to shoot wide open.

Mar 10, 2017permalink

Leonp

33mp single shot ... means that from a wedding shoot of 6 hours at 60p you go home with 1,296,000 pictures to choose from.
Matbe I'll go back to analog anyway. :-(

Mar 10, 2017permalink

The wanna b m4/3 comments are supposed to be posted in the m4/3 forum....

Not in news.......

:-)

Mar 10, 2017permalink

NZ Scott

FLruckas: This is an m43-related article.

Mar 11, 2017permalink

Yeah....
But the m4/3 comments from those secretly wishing they had FF belongs in the m4/3 forum where they're wildly more popular.
:-)

Mar 11, 2017*permalink

NZ Scott

What makes you think some people "secretly wish they had FF"?

Mar 11, 2017*permalink

I think some m43 users maybe keep dolls of FF owners to stick pins into every time sensor size gets mentioned... well, it seems like that sometimes judging by the endless heated discussions on the m43 forums about how there are just the best thing ever, which people regularly start threads on. (I have three m43 cameras, but they have pros and cons.)

Mar 11, 2017permalink