[Steinar1]

There are consistent rumours of a new OMD body with the same sensor, but the ability to do 40mp images through sensor shifting. I don't quite know what it is, but apparently 5 photos are taken to create a 40mp image.

What do you guys think? I think it could take MFT out of the "high mp" quandrum, if it is true and if results are good!

[Klaus]

Qs:

- how much light does this require (supported shutter speeds) ?

- can the mechanism also be used to increase DR and/or decrease sensor noise ?

- can it be used in conjunction with IBIS ?

[Klaus]

FWIW, this is supposed to be "it":

http://www.43rumors.com/ft5-first-images...us-e-m5ii/

[JJ_SO]

Sensor shift - does this mean the sensor is moved to a couple of positions to gather more photons? Then it's easy. I duplicate a D810 picture three times and get also a 138 MP    neat idea, good luck with that.

[Klaus]

Why shouldn't this be possible ? Assuming that the thing has an electronic shutter and sufficient light you could take numerous images without even noticing. Essentially this is like an IBIS-supported 4.8K video burst. I could easily imagine that this can work out except in low light.

[JJ_SO]

And on faster shutter speeds - so instead of 1/8000, four movements would make the whole time 1/2000 (4 × 1/8000), as each exposure position would count additionally, wouldn't it. Instead of higher resolution you'd get higher motion blurr if things move.

 

But:  I confess, I've no knowledge about that, only an idea of what could happen. Yet. I will have a look to the link.

 

Next point, how fast would the fps be? It's not only saving one picture but multiexposure also takes a bit processing power to bring the pictures together in the right order. x-times more heating of the sensor and therefore more noise?

 

Edit: so I read the link. "up to 8 shots", so 1/8000 per shot becomes 1/1000 in the final shot, plus some time for shifting the sensor - goodbye fast shutter times, but then it also becomes a philosophical question: The shutter is not fully open at 1/8000 in my 36MP, so I also don't get exactly what happened in that fraction of time (like an ultrashort flash) but a 1/250 storyboard.

 

Now, when it comes to lowlight, a 1/60 becomes 8/60 = 1/7.5 sec - I expect interesting pictures  :blink: Interesting effort from Olympus to go high MP!

[felix]

Qs:

- can the mechanism also be used to increase DR and/or decrease sensor noise ?

 

 

Isn't more resolution kind of the same? Scaling down would give you a cleaner file with less noise (which is also sort of the same as an increase in DR (?)).

 

I'm just hoping this feature will not be limited to JPGs 

 

I guess we'll see soon enough. 

[Steinar1]

I think we are guessing right now, but it would be interesting if the MFT system can find a way to survive. I love the weight and size advantage, but at some point it will have to be possible to get higher resolution. (if the system is going to survive and become accepted by more pros.

[dave9t5]

And on faster shutter speeds - so instead of 1/8000, four movements would make the whole time 1/2000 (4 × 1/8000), as each exposure position would count additionally, wouldn't it. Instead of higher resolution you'd get higher motion blurr if things move.

 

But:  I confess, I've no knowledge about that, only an idea of what could happen. Yet. I will have a look to the link.

 

Next point, how fast would the fps be? It's not only saving one picture but multiexposure also takes a bit processing power to bring the pictures together in the right order. x-times more heating of the sensor and therefore more noise?

 

Edit: so I read the link. "up to 8 shots", so 1/8000 per shot becomes 1/1000 in the final shot, plus some time for shifting the sensor - goodbye fast shutter times, but then it also becomes a philosophical question: The shutter is not fully open at 1/8000 in my 36MP, so I also don't get exactly what happened in that fraction of time (like an ultrashort flash) but a 1/250 storyboard.

 

Now, when it comes to lowlight, a 1/60 becomes 8/60 = 1/7.5 sec - I expect interesting pictures  :blink: Interesting effort from Olympus to go high MP!

 

I wonder if it's the opposite; 8 shots of 1/8 the exposure time could be enough to get the proper exposure?  Just as if you were doing multiple exposures on film?   I'm sure there is plenty of exposure information in 8 shots to give a processor some way to interpolate a correct exposure.  8 exposures at 1/8000 would give you super duper HDR, no?  

 

I guess we can stop speculating soon enough when they officially announce the camera.

 

I don't see this feature as a negative.  Even if the high res feature is limited, it's still like having two bodies in one.  

 

If you're shooting reportage or something similar then, then resolution is not so important and you don't need to use the feature.  It's like a Nikon D4s or Sony A7s.

 

If you need very fine detail, you're probably already using a tripod and rigourous techinqye (landscape or product photography) and then this feature is a useful extension.  It's like a Nikon D810 or A7r.

[Scythels]

Sensor shift technologies have been around for a while.  In essence, for perfect results it requires a static subject.  The lens' aberrations turn a dot into a blur disk (the airy disk), the sensor shift moves the sensor around so it independently samples multiple points of the airy disk.  This allows it to roughly identify what the original point should have been and recover details that fell between photosites.  The result is an image of the same pixel count, but much higher per-pixel detail.  Of course one can digitally enlarge the image to a higher resolution and get similar per-pixel detail, but in sensor noise limited situations (i.e high ISO) you will not see much better detail due to the noise.