Olympus High Resolution Mode test w/60mm macro

Lou Jost · Post by **Lou Jost** » Sat Nov 19, 2016 6:04 pm

Slightly cropped from the original 10368 x 7776 pixel image.

A 100% crop:

This is another example of the High Res Mode of the new Olympus MFT cameras. It is incredible to see an inexpensive little camera like this produce sharp 80Mp images straight out of the box.

This is done with the Oly PEN-F and Oly 60mm internal-focusing macro lens. Elsewhere in the forum I had shown that this is sharper and has less chromatic aberration than the Micro-Nikkor 60mm D lens.

I was lazy and did focus bracketing by hand. This is a wing of Urania leilus which I found laying on the ground in the Peruvian Amazonian rainforest. It is a relative of the Sunset Moth that you all know.

I transformed the RAW images to Tiffs before stacking, using ACR, process 2010, linear gamma, with default sharpening. It is a Zerene DMap stack of about 6 images. Unsharp sectors are my fault for making too-large steps while focus-stacking by hand. I sharpened the stacked image slightly and also retouched out the dust that rains down on everything in my house. Here a crop without sharpening or retouching after the stacking:

rjlittlefield · Post by **rjlittlefield** » Sat Nov 19, 2016 6:59 pm

Impressive. What f# was this shot at?

--Rik

zzffnn · Post by **zzffnn** » Sat Nov 19, 2016 7:10 pm

Great images.

What is the cheapest useful Oly micro4/3 body that has this high res mode? The Pen F is not cheap for a casual hobbyist, at $1300 new, though its spec's look impressive. Such as the shutter speed.

I assume the Oly 60 mm macro is the F/2.8 one that costs around $500 USD and close focus distance of 18.8 cm?

Lou Jost · Post by **Lou Jost** » Sat Nov 19, 2016 7:48 pm

Thanks Rik. This was at f/4, one stop down from max. I did a few shots at f/5.6 and they seemed softer, probably diffraction setting in.

Lou Jost · Post by **Lou Jost** » Sat Nov 19, 2016 7:51 pm

zzffnn, the OM-D E-M5 Mark II has this mode also, though with a less dense sensor. When I said "cheap" I was speaking in relative terms, since the only native 50-80Mp cameras are far more expensive.

Yes, that is the macro lens I used. Very good and feather-light. Ideal for field work.

These cameras also have in-camera focus-bracketing, which is amazing, though stupidly it doesn't work in High res mode.

rjlittlefield · Post by **rjlittlefield** » Sat Nov 19, 2016 9:36 pm

Lou Jost wrote:Thanks Rik. This was at f/4, one stop down from max. I did a few shots at f/5.6 and they seemed softer, probably diffraction setting in.

Diffraction is definitely a factor.

To get some gut feel, http://www.cambridgeincolour.com/tutori ... graphy.htm and scroll down a little less than half way, to the section titled "VISUAL EXAMPLE: APERTURE VS. PIXEL SIZE".

There, select the camera type "Canon PowerShot G9, S100", which will give pixel diameter 1.9 microns. That's a little bigger than the effective pixel size on your Pen F, which computes out to 1.675 microns given the sensor size of 17.4 mm x 13 mm with 10368*7776 pixels.

Then, select f/5.6 to see Airy diameter 7.5 microns, which would be almost 4.5 pixels wide. Even at f/4, Airy diameter is 5.3 microns, over 3 pixels wide. The Airy disk is center-weighted, so it's not like you're splatting every point evenly across 3 pixels, but still that's broad enough to soften the image quite a bit.

Looking at it another way, using the MTF formulas at http://www.photomacrography.net/forum/v ... 831#124831, with lambda = 0.00055 and f/4 (NA=0.125), the optical MTF is just under 23% at the Nyquist sampling limit of one cycle for each pair of 1.675 mm pixels. At f/5.6, it's barely MTF = 2.7% . (!)

You are completely justified in doing some pretty aggressive sharpening at pixel level with these images...

--Rik

Lou Jost · Post by **Lou Jost** » Sat Nov 19, 2016 10:14 pm

Rik, yes, I've been using that page a lot to figure out the constraints on my lenses and apertures. That's one reason why I have been so interested in coupled lenses rather than using extension--- it theoretically gives me a given m at significantly wider apertures.

However, after lots of experiments, I think the figures on that page are overly pessimistic. Maybe it has something to do with the way they take into account the Bayer pattern. Actually the effect of the Bayer pattern is subject-specific. If the object being photographed is blue or red, then the pixel size is double the size that the table calculates. [Edit: should be "four times the pixel size that the table calculates". So the effective number of pixels on a 40Mp camera would be just 10Mp for a pure red or pure blue object.]

I find that at m=0 with my best lenses at their best aperture, I start seeing diffraction in the image around f4. It is as if the effective number of pixels is 40Mp. And in fact if I reduce my big images to 40Mp, for apertures faster than f4 the reduced image is sharp at the pixel level.

rjlittlefield · Post by **rjlittlefield** » Sat Nov 19, 2016 10:46 pm

Lou, I admire your enthusiasm and diligence at wringing as much performance as possible out of that Pen F.

That said, I am frequently struck that in large part you're struggling with problems that are created by the small size of the sensor.

To take this particular example... Estimating the size of your frame as 50 mm diagonal, then you're running at 0.43X. With a full frame sensor you'd be at 0.86X. Using a nominal f/4 lens in both cases, focused by extension, the effective f-numbers would be 5.72 and 7.44, a ratio of only 1:1.3 in effective f-number versus 1:2 in sensor size, giving an improvement ratio of 1:1.54 in favor of the larger sensor.

If you have to reduce your images to 40 Mp in order to get an image that is sharp at the pixel level, then your Pen F would be handily outdone by a 50 Mp full frame sensor.

Given that you have the Pen F, I think you're doing a great job of figuring out how to exploit it.

But to the extent that readers may think you're recommending the Pen F as a preferred way to get high megapixel images, I feel compelled to present the other side of the coin.

--Rik

austrokiwi1 · Post by **austrokiwi1** » Sat Nov 19, 2016 11:41 pm

I have been searching through my photo files without success. Some time ago I assessed the resolution and CA of the Oly 60mm macro and was actually quite disappointed. I recall it was much better than the Olympus ED 50mm F2(Which I ended up selling). Again going from memory, the SK APO Componon-S 40 F 2.8 HM enlarger lens was heaps better( mounted normally its optimal magnification range matches the MFT sensor nicely). The Oly 60mm macro is the best native lens available for the MFT system.... but it isn't the best lens available. I would go so far ( and this is a really "stretched" analogy) as to say the SK40/2.8 is , for the MFT system, the "Printing nikkor"

I would tend to agree with Rik, and add that my Oly ( doesn't have high resolution mode) is a great back up camera and that's fortunate as that is what I purchased it for. However if I have a challenging subject and need to use magnifications above 0.6X then its my FF Sony every time. The huge issue for me with the OLy is the amount of noise produced by the MFT sensor(IMHO for me to notice it has to be a significant issue).

I believe the small sensor of the MFT system is great for low magnification work. However, the limited dynamic range and noise issues combined with the pixel packing really recommends larger sensor systems when 1-1 and higher magnification is required.

Lou Jost · Post by **Lou Jost** » Sun Nov 20, 2016 6:40 am

Rik and Austrokiwi, thanks for those observations.

Rik, I actually am suggesting that this camera is on the same level with FF and is ideal for microphotography. I am careful NOT to get my magnification from extension. As I mentioned earlier, I try to use coupled lenses when possible because of the advantage in EA. In that case, the EA is proportional to m and for a given FOV, I need only half the m of a FF sensor. So the faster EA exactly compensates for the smaller pixels, right?

I think hi-res mode, downnsampled, makes smoother, less noisy images than FF images of 40-50Mp.

Even though not perfectly sharp, an 80mp image still can outresolve a perfectly sharp 40Mp image, by containing some info about lines that are spaced more closely than the 40Mp sensor can resolve.

Edit: The image in this post was made with continuous light at a shutter speed of 1/4 s. No settling time between shots was needed, since the camera is mirrorless and has no mechanical second curtain.

The advantages of the MFT are economic (huge difference; not only in prime lenses and camera body but also in the ability to use cheaper, lower m objectives for a given FOV) and ease of vibration control (flash + fully electronic sensor in PEN-F; few FF cameras have that).

Austrokiwi, I can't speak to the SK 40mm. I can only say that the Oly is better than the Nikkor 60 D which is a highly ranked lens, and it allows extremely fast automatic in-camera focus bracketing (one minute to take a stack of a couple hundred shots!) which the SK does not.

So yes, you could say I am a fan! Having come from Nikon, this is another world. A Canon or Sony user may feel differently. But I for one don't miss that flopping mirror and the worry of how many shutter actuations I am running up! On my camera there is NO wear even if I run very large stacks.

austrokiwi1 · Post by **austrokiwi1** » Sun Nov 20, 2016 9:43 am

I can't speak with regards to the Pen - F. I can say that I too am a ICL fan. My Sony is a full frame mirrorless and the IQ, at the same magnification using the same optics is significantly better than the Oly OMD-E M10II. Like your camera the sony has a silent shutter.

{I have also moved to Coupled lenses for up to 5X magnification. My major challenge has been in finding suitable lens pairings}

Lou Jost · Post by **Lou Jost** » Sun Nov 20, 2016 10:08 am

Austrokiwi, I am sure that a mirrorless FF camera could have better image quality than an MFT camera. I would be surprised, though, if this could be true at the same magnification. The FF image would put far fewer pixels on a given feature than the MFT sensor. Did you maybe mean "at the same field of view"?

Suppose your Sony FF has a 40mm sensor and you are shooting at m=2. My Oly would get the same FOV at m=1. If you used m=1, in order to get the same FOV you would have to take 1/4 of your image, or 10Mp. I would bet a lot that your 10Mp crop would not match my 20Mp image of the same subject, and it would certainly not compare well with an 80Mp hi-res image. That's why I think you may have meant "at the same FOV" rather than at the same m.

Yes, finding suitable pairs of lenses to couple is a big challenge. But theoretically there could be a big payoff for m less than about 3.

Lou Jost · Post by **Lou Jost** » Sun Nov 20, 2016 10:20 am

Austrokiwi, in your comparison of the Oly 60m and your SK 40mm, what m did you use? [Edit to delete this sentence: "As you know, to compare the two fairly at m greater than 1, the Oly needs to be used in reverse, and should be focused to a magnification of 1/m.". I am sure you know that, and you probably compared them at <1x. I'd love to see those tests if you find them.]

Lou Jost · Post by **Lou Jost** » Sun Nov 20, 2016 11:09 am

Rik, I should add that there are other reasons why I switched from Nikon to MFT. I do a lot of field work in remote places that require hiking for many days, and the weight of APS or FF lenses is too much for me. If money were no concern, sure, a nice set of FF gear for microphotography and a nice set of MFT gear for fieldwork would be ideal. But if money is limited, MFT is a good compromise.

austrokiwi1 · Post by **austrokiwi1** » Sun Nov 20, 2016 11:24 am

How I measured the resolution:

I projected the image of a 1951 USAF 1X resolution target by the lens being tested through a 10X microscope objective ( at 1X magnification).

I still haven't found my tests of the 60mm I will redo it it later in the week (I currently have a new toy that I am focusing my attention on (Canon (RF) Dream lens) So it may take a while)

Here is a comparison of the oly at 5.25X and the A7rII at the same magnification ( I have cropped the Sony image to the same FOV as the Oly. To my eye the Sony is still producing the better image. Optics are the SK 40/2.8 coupled with an SK 210/5.6 Large format enlarger lens( A very workable combination that happily covers a full frame sensor) Lighting was the same in each case and I set custom White balance separately for each camera.

I am not convinced the 80mp high res mode of a MFT camera will be better than that from a FF or even an APSc camera. Its not just resolution ta makes a good picture, Dynamic range, contrast, and noise levels would all be different. I believe the Sony A7rII with the back illuminated sensor may well perform better. The really big question, for me, is how much of an 80 mp MFT image is unnecessary/wasted resolution? I have compared the output of my 42mp ff camera to that of a 80mp Medium format... and I have to say in that comparison Medium format images were better. However this was just comparing different images... I have not had the joy of using a medium format camera.

For me the advantage of MFT comes with low magnification shots( less than half life sized)