The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

OM-D E-M5 Mark II Pre-Order

scott kirkpatrick

Well-known member
I'm tempted (but I have three perfectly good bodies already).

Has anyone seen a description of how the 40 MPx process works? From the B&H descriptive article, I see that you turn off image stabilization, lock the camera to a tripod, and take 8 pictures at sensor shifts 1/2 pixel apart. Why 8? I would think 4 are needed, so maybe the next 4 are for averaging to reduce noise background.

scott

(edit): Have a look at http://www.dpreview.com/previews/olympus-om-d-e-m5-ii/4 , where they give a pretty detailed discussion and compare M5II output with Nikon 810. It turns out that the Olympus system shifts 4X by a whole pixel step to sample all four Bayer filter values!, then it shifts half a pixel in both x and y directions in a single step and repeats the 4X process at that new origin. In the comparison, the 810 has some horrible Moire, which the M5II eliminates, but the resolution of the 810 is better.

(second edit): Check out further comparisons at http://www.imaging-resource.com/PRODS/olympus-e-m5-ii/olympus-e-m5-iiA.HTM . Dave Etchells (who is an imaging engineer) explains a bit further. The comparisons are with D810 and Sony R7s. The whole process takes about a second to happen.
 
Last edited:

Annna T

Active member
I'm tempted (but I have three perfectly good bodies already).

Has anyone seen a description of how the 40 MPx process works? From the B&H descriptive article, I see that you turn off image stabilization, lock the camera to a tripod, and take 8 pictures at sensor shifts 1/2 pixel apart. Why 8? I would think 4 are needed, so maybe the next 4 are for averaging to reduce noise background.

scott

(edit): Have a look at Olympus OM-D E-M5 II First Impressions Review: Digital Photography Review , where they give a pretty detailed discussion and compare M5II output with Nikon 810. It turns out that the Olympus system shifts 4X by a whole pixel step to sample all four Bayer filter values!, then it shifts half a pixel in both x and y directions in a single step and repeats the 4X process at that new origin. In the comparison, the 810 has some horrible Moire, which the M5II eliminates, but the resolution of the 810 is better.

(second edit): Check out further comparisons at Olympus E-M5 II Review: Now Shooting! . Dave Etchells (who is an imaging engineer) explains a bit further. The comparisons are with D810 and Sony R7s. The whole process takes about a second to happen.
I have read these two reviews and some others during the last hours and I think that DPreview 1) hasn't used the optimal lens for their shots. 2) the raws looks softer than the jpegs, as if they were in need of further contrast and sharpening. 3) they used PS or LR for conversion, while Adobe has not yet issued a definitive support for that camera.
I found that the Imaging Ressource samples were much more convincing. Surprisingly good in fact.

Still the number of situations were you will be able to use the 40mb resolution aren't particularly common. Any small winds moving foliage, any moving waters will ruin your effort. And what about changing skies with clouds ? Forget about streets unless you are only looking at the higher floors (and provided no one show up at a window or a balcony... No flapping drying laundry, no flags.. Etc..
May be it will gain some new customers for Adobe CC : those wanting to take out blurred things from the HI Res scenes..
 
Last edited:

k-hawinkler

Well-known member
Any wind shaking your tripod ... as well might ruin the 40 MP shot.

I think the real improvement seems to be their IBIS with better sensors.
 

scott kirkpatrick

Well-known member
Thinking about the 40MPx -- effective in JPG, it's 64 MPx in RAW -- I can understand why it blows away the D810 in Moire resistance but does not keep up as well in sharpness. It's because the process starts with 4 full pixel distance displacements, each one putting a different Bayer filter element in front of the virtual pixel where the data is accumulated. That gives an honest 24 bit Foveon-like color signal at every pixel with no aliasing. Then they repeat the four steps displaced in both the x and y directions (i.e. at 45 degrees) by a half-pixel. If you wanted to get the best information about contours to a sub-pixel resolution, you would take three images, the original shot and one displaced in x by half a pixel, then one displaced in y by the same. A 45 degree contour aligned in the +x+y direction would not be seen at all with this single displaced set of images.

Rendering the 64 MPx raw files requires software that nobody seems to have -- I would expect all the reviews are based on in-camera jpegs.

scott
 

Knorp

Well-known member
It'll be perfect for stills under controlled conditions, but the improved IBIS is probably its most interesting asset.

All the best.
 

raist3d

Well-known member
Add me to the list !

As for electronic shutter, 1/20 sec. Is the limit. (Read that on a Spanish review, in DSLRmagazine or Quesabesde)
You mean electronic shutter goes to 1/16,000... I guess you are talking about the 40 MP mode?

- Ricardo

PS: My interest with this camera has nothing to dow with the 40 MP mode, as it does nothing for my photography, but the electronic shutter :)
 

scott kirkpatrick

Well-known member
Here's how Hasselblad does the pixel-shifting stuff to reach 200 MPx in their $45k H5D 200C MS digital CCD MF cameras, released last summer: http://www.hasselblad.co.uk/media/4808915/h5d-200c ms_datasheet_en_v2.pdf .
They start with the same four shots spaced one pixel apart. I think of that as just lifting the Bayer filter up and shifting it each time so that you get 4X as much color information. That's their 4-shot option. The 6-shot option then takes a single exposure with the chip shifted half a pixel to the left and another with the chip shifted half a pixel up. If you really wanted perfect non-interpolated color and twice the linear resolution, I guess this would take 12 shots (4x3), but even the H5 doesn't have buffers that big...

scott
 

raist3d

Well-known member
Well, this camera looks real nice but I want to play with some raws on sharp lenses. There's nothing particularly wrong, but I have really become more and more sensitive to the effects of an anti- alias filter now that I have been using AA-less cameras for a while and as hard as I try to look for detail, there is this "micro smear" that seems to kill some details. It's been always there on AA cameras in general.

So the question it, at least for night shots, at what point the necessity to do more noise reduction on a smaller sensor matches this one with the AA filter.

Ah well, will wait for more raws.

- Ricardo

PS: Iridient can open the RAW files of this camera. It's not officially supported but the current unofficial support is actually pretty decent color/exposure wise, so we can start seeing some raw files. It certainly does not support the high-res mode.
 

Knorp

Well-known member
Those files are huge ... :confused:
What will support the 64M RAWs ?
What about your SD cards ?

When shooting in RAW+JPEG mode, the camera will save a 40M JPEG file, a 64M RAW (ORF) file and a 16M RAW (ORI) file. 64M RAW images may be processed using Photoshop CS4 and later with a required plug-in.
One shot ... 120M :wtf:
 

bavanor

Member
Raist,

From what I have been reading it sounds like the new Em-5 mkII does not have an anti aliasing filter.

Aaron
 

raist3d

Well-known member
Raist,

From what I have been reading it sounds like the new Em-5 mkII does not have an anti aliasing filter.

Aaron
Thanks Aaron- do you have a link with an official Olympus statement or specs sheet where it mentions it? Thanks, much appreciated.

- Ricardo
 

ptomsu

Workshop Member
From what I know already the EM5 and the EM1 do not have AA filters!

So why would Olympus go back to AA filters in a newer model?

Peter
 

raist3d

Well-known member
I know the EM-1 doesn't have one, but I was never able to fully corroborate the EM-5 doesn't have one. Do you have a link that proves the E-M5 does not have an AA filter?

The look on the shots I am seeing of the EM5 MKII seems like it has one.

Now, why would Olympus go from EM-1 with no AA to EM5 MKII with one? Simple- it can be part of a way to draw the line between the two tiers. It's like with the old 4/3rds - even though the e-620/e-420 used a "similar sensor to" the high end E's, the higher tier always had an extra something.

Basically, if it has no AA filter, why is it nowhere on the specs or marketing materials? I certainly could have missed something.

- Ricardo
 
Top