New Olympus EM1 Mark 3 will have 80mp jpeg for High Rez mode

[1of1snowflakes]

I currently use the Olympus EM1 mark 2 along with the MC-14 teleconverter, two 10mm extension tubes, the Olympus 60mm macro lens and a Raynox 202 to do high resolution stacks of grains of sand. So far, I have liked the results better than what I get with the Mitutoyo 10x - in most cases.

I sell the pictures I take of the sand grains in 11x14 or 16x20 mats with my signature on the bottom right, and place the sand grains with a tiny amount of super-glue next to my signature - so anyone who buys a print gets the actual sand grains in the picture for a unique piece of artwork no one else will ever have.

Here are a few examples of the images I have obtained. First the original file and then the cropped file. The original file is 8160x6120.











The new High Resolution mode on the EM1 Mark 3 will be 10368x7776 pixels, which should only help with the detail on my photos. I didn't know what your thoughts were on continuing with this type of stacking?

I use the Olympus Capture program on my computer to move my focus small increments instead of using a rail, and then taking a high resolution shot at each increment. Each sand grain takes around 15 minutes to capture the 35-50 shots needed for the stack in Helicon.

Here's an example of the same sand grains - the first with the Olympus setup as described above, and the second with the same camera, but with 160mm of tube and reversed Raynox 150 with a Mitutoyo 10x.




.

I see things I like in both shots and things I do not. No one else is really photographing sand like I am on an ongoing basis....
so I am kind of unique in what I am doing. My customer base really prefers overall the Olympus setup instead of the Mitutoyo.

[rjlittlefield]

That's great news about the 80mp for in-camera JPEG.

https://www.ephotozine.com/article/olym ... new--34470 explains it this way:

Previously, the largest size of image you could shoot on the OM-D E-M1 Mark II was 50-megapixels, using the high-res multi-shot mode. Now, you can create 80-megapixel images using the high-resolution tripod shot mode, as well as capturing 50-megapixel images hand-held.

This is the first time that the 50mp restriction has really made sense to me. I have no hands-on experience with Olympus pixel shift, and it had not occurred to me that they were thinking it would be used handheld as well as completely static.

--Rik

[Antal]

Those lines in the out of focus parts in your image: Are they a result of stacking?

[1of1snowflakes]

Those lines in the out of focus parts in your image: Are they a result of stacking?

yes,
the sand that I use in my background can cast shadows that the stacking software has trouble resolving in the stack. So something it will be wavy lines....at times I will fix those but not always depending on what it looks like.

[Pau]

Beautiful micro garnets

[Lou Jost]

the sand that I use in my background can cast shadows that the stacking software has trouble resolving in the stack.

I think the wavy dark band in the out of focus area of your shot is the result of a drop in light intensity during one frame of the stack.

The new High Resolution mode on the EM1 Mark 3 will be 10368x7776 pixels

The website you refer to must be wrong. This is is the same resolution I have always been getting on my lowly PEN F for many years. I have trouble believing that the much more recent flagship Oly EM1 M2 would self-cripple itself. 50MP would have been its resolution for shifted jpgs, but the RAW image would have been 80Mp just like the new M3.

This is the first time that the 50mp restriction has really made sense to me.

It always made sense to offer that resolution for the jpgs because the pixel-shifted 80Mp image has some redundancy in it. There really only is about 50Mp of information in it, in most situations. But the RAWs were/are 80Mp and I always preferred my stacking results from 80Mp transformed RAW images over those made from the 50Mp jpgs.

[rjlittlefield]

It always made sense to offer that resolution for the jpgs because the pixel-shifted 80Mp image has some redundancy in it. There really only is about 50Mp of information in it, in most situations. But the RAWs were/are 80Mp and I always preferred my stacking results from 80Mp transformed RAW images over those made from the 50Mp jpgs.

And yet, if there's only 50 mp of information, why have you always preferred working with the 80's?

I am reminded of our earlier discussion at https://www.photomacrography.net/forum/ ... 746#248746 .

--Rik

[Lou Jost]

For one thing, stacking involves image alignments that get rounded to integer numbers of pixels. More pixels should mean more precise alignments. But probably more important is that the RAW files aren't sharpened by default, like the jpgs are.

[rjlittlefield]

For one thing, stacking involves image alignments that get rounded to integer numbers of pixels. More pixels should mean more precise alignments.

True enough.

But then, the same thing is true for any camera-resolution images, and Zerene Stacker offers the option to up-sample on the fly as a way of overcoming that issue.

So now I'm curious about whether you use the up-sampling option, and if so, under what conditions?

BTW, I'm not intending to be a troll here, though I am painfully aware that I probably sound like one.

I'm just very sensitive to the fact that sometimes people do things for perfectly good reasons that are not the ones they offer as explanation, and I'm trying to ferret out whatever real reasons there are.

So yes, I am pressing you to resolve whatever look to me like inconsistencies.

Processing from raw to avoid oversharpening makes perfect sense to me. Processing from 80 mp raw to 80 mp JPEG or TIFF makes perfect sense to me in terms of preserving all the resolution that the sensor captured. Processing to 16-bit TIFF instead of 8-bit whatever makes perfect sense to me in terms of preserving the gray scale. But the image alignment thing seems rather shakier.

My own guess is that the 50 mp images do not capture all the resolution in the pixel-shifted raws, and that the loss of resolution just naturally feeds through to produce lower quality stacks. If that's not the case, then I would be very interested to learn why, because I'm not aware of any reason in sampling theory to think that 50 mp should be enough.

--Rik

[Lou Jost]

For my most important stacks of non-high-res shots from MFT, I do sometimes (not often) upscale them to double their size.

I always agree that more pixels is better. But they come at a cost, in acquisition time, processing time, and storage space.

I had read studies of pixel-shifting showing that on normal subjects the shifting doesn't actually double the resolution over normal shots. The exact resolution advantage is subject-dependent. 50Mp is about enough to capture the average resolution gain. (Enrico S. on this site has made his own estimate of about 40Mp if I recall correctly) 80Mp certainly captures a tiny bit more. Whether that tiny bit of improvement is worth the extra time and storage space depends on what I will do with the image.

[Scarodactyl]

It seems to me that a mitutoyo should be getting better results than this (in terms of technical stuff I mean), certainly better than a macro on extension. That's not to say anything to the detriment of the photos because they are obviously coming out beautifully, it just feels like the details should be a lot crispier with the 10x given my own experiences with it.

[Lou Jost]

It seems to me that a mitutoyo should be getting better results than this (in terms of technical stuff I mean).

I thought the same thing but was hesitant to say it.

[Macrero]

I received the Panny G9 a week ago, haven't had much time for photos, but did a few quick HR test stacks, and I can say I am pretty impressed with the result.

I honestly lol'd at "only 50MP". Yep, that's about the actual Hi-Res, but "only"? A tiny, easily coverable sensor with 3.33 µm pixel pitch , that can output a 50MP full of detail and free of false colors images. Who actually needs more? I certainly don't.

The 40MP RAW was misleading though, that's for JPEG. The RAWs are 80MP (128MB approx). I convert and stack at full-res and downsample at 45-50MP latter.

Best,

- Macrero

[Lou Jost]

macrero, yes, hi res is wonderful! If it were faster I would use it more often. How long did the capture and write process take on the G9?

[Lou Jost]

The 40MP RAW was misleading though, that's for JPEG.

If this comment refers to my parenthetical comment about Enrico's estimate of 40Mp, this did not refer to jpgs, it referred to what he thought was the size that the RAW file needs to be reduced to in order to appear sharp at the 100% level, or something like that. I don't recall his exact criteria, maybe he can be add more explanation here.