Best (situational) example of Hand-Held High Resolution in E-M1 Mark III

Joined
Jul 31, 2013
Messages
441
Location
UK
Real Name
Ovidiu
This evening, while I was going to work, I chanced upon my first deer sighting of the year. It was quite dark, at 19:50, and behind a tree line with clouds covering the sunset as well so high ISO and the slow shutter speed were unavoidable. After making about a dozen single shot ISO 6.400 at 1/125 sec and f 4 from 200mm end I decided to try HHHR since there was nothing to lose at this point (I couldn't get any closer because of the fence with a barb-wired in front of me and I was running out of time to get to my workplace as a night carer). I only changed the drive mode since the exposure was telling me I was about 2 stops underexposed at the above settings and did not want to risk going slower than 1/125 sec if the deer decided to move too fast ... luckily it did not move at all ... well until I took the HHHR shot, it took off like a shell out of a shotgun, I don't know why.
These are the original .ORF files and the .XMP from my Lightroom adjustments, these are as far as I would comfortably go on such high ISO output:
https://1drv.ms/u/s!AksBB2IBTehLhJMYE7m80AsexaCQlQ?e=WYHBt3
The OOCR have a magenta tint to them in the WB which I assume the Auto WB is a bit goofy in this situation (never seen it behave this way before, or to this extent) but it was quite fixable in Lightroom:
Subscribe to see EXIF info for this image (if available)


Subscribe to see EXIF info for this image (if available)

At 100% magnification there are significant IQ differences even OOCR and what took me by surprise was not the less amount of noise or the higher level of detail on the HHHR 50 MP image but on how smoother the OOF area is compared to 20 MP. I would have been tempted to believe that the DOF is somehow narrower on the 50 MP which, by all-optical physics, it shouldn't be that. Maybe the 6.400 IOS noise acts as a sharpener at 20 MP that destroys the smoothness of the bokeh? I haven't noticed anything like this on my previous tests of the HHHR (but I admit that I have the Mark III for barely one month now).
Subscribe to see EXIF info for this image (if available)


Subscribe to see EXIF info for this image (if available)

Of course when edited the 50 MP file has a lot more room to work in a bit more IQ, of which I am not surprised. Making the image, compared to 20 MP, a lot more useful and printable by a decent margin. Given that the RAW file is 50% larger at 39 MB I would be tempted, from these results, to give HHHR a try even for wildlife images IF the subjects is large enough and still enough after I got the single 20 MP results I needed first (deer, foxes and maybe even bunnies should work when they stand still to listen for predators or prays). I would tend to agree that HHHR gives about 2 steps improvement in IQ over single shot, making ISO 6.400 look more like ISO 1.600 (but in the shot above I would say even closer to ISO 800) and that's before scaling it down to 20 MP.

The 50 MP file seem to handle like a champ even some extra sharpness and slight EV bump with local brush adjustment (+0.33 EV, +25 Texture, +25 Sharpness) on the deer over the base edit of:
+1 EV (didn't want to push to hard since I'm not planning on printing it) ,
-100 Highlights (the green in the grass was getting to bright),
+25 Contrast (to keep the shadows from popping up like christmas tree and to keep the overall darker mood),
+33 Texture (to bring a bit of details in the furr and the grass),
+15 Clarity (to add even more local contrast on fine details,
+15 Dehaze (to clear up any warm waves that might haze the atmosphere, it was quite warm today at noon),
+100 Sharpness (the 50 Megapickles needs quite a lot of sharpening to push the file to its true potential),
+1.5 Radius (my default setting for all images),
+50 Details (makes the sharpness a bit more effective)
+33 Masking (this will keep the sharpening away from completely smooth OOF areas, especially important on High ISO or textureless areas)
And finally I added 2 graduated filters: one from the bottom, convering 15% of the image, and one from the top, covering 25% of the image, with -50 Clarity and +50 Noise to smooth out the OOF area and the bokeh and the noise where it needs to be smooth.
Subscribe to see EXIF info for this image (if available)

100% crop :)
 

Bushboy

Mu-43 All-Pro
Joined
Apr 22, 2018
Messages
1,174
Technically impressive. :)
I can’t get my M5ii high res to behave nicely. To much movement.. Looks like a nice Roe buck? You’ll have to go back...
 
Joined
Jul 31, 2013
Messages
441
Location
UK
Real Name
Ovidiu
Technically impressive. :)
I can’t get my M5ii high res to behave nicely. To much movement.. Looks like a nice Roe buck? You’ll have to go back...
I pass that field every day I go to work and since the human malware problem I'm working about 5 to 6 nights in a row... I can't wait to get lucky again and the sun to set later so I can get better light.
Last year around the same time of the year I got lucky with 2 Roe bucks, one male and one female but I was stuck at 16 Megapicles so less cropability (until Olympus launches that 150-400mm Pro). I have barely used the HR on my E-M5 Mark II because I have a allergy for tripods (they make me tired to carry, annoyed to deploy and less enthusiastic to make pictures).
 
Last edited:
Joined
Jul 31, 2013
Messages
441
Location
UK
Real Name
Ovidiu
Nice! That's a great example. Too bad my E-M5III can't do that... :(
I always like to have 2 camera with me and after E-M1 III I will wait for Olympus to make a more compact (Pen F II or E-M5 IV maybe) version with HHHR because I want feature parity between the two. (Yes, I know it will take about 3 to 5 years until then)
 

c0ldc0ne

Mu-43 Regular
Joined
Oct 9, 2012
Messages
93
The HHHR feature on the E-M1iii and E-M1x are completely different implementations to the HiRes on the E-M5ii, Pen-F, and EM1ii.
Is that an official Olympus statement, or did you determine this empirically?
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
895
Location
New Hampshire
Real Name
Rich
I think it is too bad that Olympus made HHHR mode have the same MPs as the regular Hi-res mode (both 50MP), making people think they are the same.

HHHR mode seems almost like something you could do in post-processing via software, assuming the camera got the images ready for it (some type of bracketing). That way, even lesser cameras like the E-M5III could have it because they wouldn't be doing the actual image creation in-camera (which apparently requires a lot of processing power).
 
Joined
May 18, 2017
Messages
120
Location
Phoenix, Arizona USA
Real Name
Jeffrey Behr
I think it is too bad that Olympus made HHHR mode have the same MPs as the regular Hi-res mode (both 50MP), making people think they are the same.

...
In the EM1-3, HHHR produces 50MP files while tripod-HR still produces 80MP.. From what I've read, the two modes vibrate the sensor differently, and the former uses the natural vibration of the hands and camera in its processes.

I LOVE HHHR; it's my default hand-held shooting mode.. I have LOTS of examples if anyone wants to see some of them.
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
895
Location
New Hampshire
Real Name
Rich
In the EM1-3, HHHR produces 50MP files while tripod-HR still produces 80MP
For tripod-HR, jpgs are 50MP. The RAW files are 80, but the jpgs are 50, just like the HHHR. As far as how the modes differ, have you watched the YouTube video I link to above? I think he explains it pretty well.
 

Zairski

Mu-43 Regular
Joined
Jul 23, 2017
Messages
169
Location
Minnesota,USA
I would like to see some.
In the EM1-3, HHHR produces 50MP files while tripod-HR still produces 80MP.. From what I've read, the two modes vibrate the sensor differently, and the former uses the natural vibration of the hands and camera in its processes.

I LOVE HHHR; it's my default hand-held shooting mode.. I have LOTS of examples if anyone wants to see some of them.
 

pake

Mu-43 Hall of Famer
Joined
Oct 14, 2010
Messages
2,146
Location
Finland
Real Name
Teemu
...ahem...who on earth uses JPGs for hi-rez stufff? 🙂
Weeeell.... As it is much cleaner than a "normal" JPEG and in same cases can even beat a (smaller) raw image I would argue that at times the HHHR JPEG is good enough. If you're looking for the best possible IQ I agree RAW is the way to go but surely the JPEG is good enough for most use cases.
 

pdk42

One of the "Eh?" team
Joined
Jan 11, 2013
Messages
6,875
Location
Leamington Spa, UK
In the EM1-3, HHHR produces 50MP files while tripod-HR still produces 80MP.. From what I've read, the two modes vibrate the sensor differently, and the former uses the natural vibration of the hands and camera in its processes.
Yes, they are quite different implementations, albeit with similar end results.

HHHR is essentially an image stacking and upresing exercise that can be fully emulated in post processing from multiple images from any camera (if you can be bothered to do it!). It works by filling in gaps caused by fine detail getting lost across the matrix of RGB pixels that make up the colour filter array (CFA). Our 20Mp sensor is about 5200x3900 pixels, but these pixels are not full-colour pixels - they are arranged as repeating sets of 2x2 groups, each with 2 green pixels, 1 red and 1 blue. When the image is processed, a process called de-mosaicing is applied which converts these repeating groups of four points into an array of single points each with an RGB reading. The algorithms that do this have varying levels of complexity, but the net result is that our 20Mp sensor reduces to something much smaller when we consider it as a full-colour array. Fine details that are smaller than the 2x2 matrix will be lost to some degree.

If you take multiple shots, natural movement of the camera will cause fine details to shift in their positioning within these 2x2 groups. When demosaicing runs, slightly different outputs will be produced for any particular fine detail. If you upscale these images and then align and average them later in post processing (in camera or otherwise), these details will be revealed and so we can produce an overall higher resolution image than a single shot would give. As a bonus, the stacking uses a "median" mode, which also gives a significant improvement in noise.

The tripod HiRes mode does something similar in that it stacks multiple images, but the images are precision-shifted by the IBIS rather than random movements by hand shake. The movement algorithm uses a sort of circular dance that ensures every point in the scene gets exposed through each of the pixels in the 2x2 matrix. So each point is captured as a successive reading of its luminance in each of the R, G, & B channels. This adds an important extra feature - we don't need to de-mosaic. Not only does this increase the resolution, but it delivers much improved colour accuracy and reduces artefacts caused by demosaicing - especially moire. It's better than HHHR, but of course it needs a perfectly static camera since the IBIS can't do the precision sensor shift and the image stabilisation at the same time.

I LOVE HHHR; it's my default hand-held shooting mode.. I have LOTS of examples if anyone wants to see some of them.
I'm getting really close to pushing the button on the mkiii - entirely for HHHR. I'd love to see some real-world examples - esp landscape or cityscape.
 
Joined
Jul 31, 2013
Messages
441
Location
UK
Real Name
Ovidiu
At the moment this is pure conjecture on my part but I am starting to notice that HHHR does not work very well with Panasonic O.I.S.
I am noticing slightly (not sure if it's really of any significant) sharper images using only the IBIS compared to prioritizing the O.I.S. on the Panasonic Leica 50-200mm f 2.8-4. I haven't made any comparative tests yet but I will try and look into it. There are a lot of variables to account for, like the aperture, the focal length and there is also the, invariably, the randomness of HHHR between each shot (it will be extremely hard to change Image Stabilization settings in the camera menu and retain a 100% identical framing between each shot).
Any Panasonic lens users that have tried HHHR on Olympus E-M1 Mark III and X have noticed something similar?

And in the same area, I don't think Panasonic Leica 50-200mm f 2.8-4 is really sharp enough to handle 50 Megapickles images, I want to try the 80 MP HR to see if it's the same or worse. I find the Olympus 12-40mm f 2.8 (at the telephoto end) handles HHHR a lot better.
 

pdk42

One of the "Eh?" team
Joined
Jan 11, 2013
Messages
6,875
Location
Leamington Spa, UK
At the moment this is pure conjecture on my part but I am starting to notice that HHHR does not work very well with Panasonic O.I.S.
I am noticing slightly (not sure if it's really of any significant) sharper images using only the IBIS compared to prioritizing the O.I.S. on the Panasonic Leica 50-200mm f 2.8-4. I haven't made any comparative tests yet but I will try and look into it. There are a lot of variables to account for, like the aperture, the focal length and there is also the, invariably, the randomness of HHHR between each shot (it will be extremely hard to change Image Stabilization settings in the camera menu and retain a 100% identical framing between each shot).
Any Panasonic lens users that have tried HHHR on Olympus E-M1 Mark III and X have noticed something similar?

And in the same area, I don't think Panasonic Leica 50-200mm f 2.8-4 is really sharp enough to handle 50 Megapickles images, I want to try the 80 MP HR to see if it's the same or worse. I find the Olympus 12-40mm f 2.8 (at the telephoto end) handles HHHR a lot better.
That makes sense actually - Oly cameras with HHHR turn IBIS off since they want the random movement between frames. OIS on the other hand will be permanently on and will be stabilising the image before it hits the sensor - this will probably remove some of the random re-positioning on which HHHR stacking depends.
 

Ruairi

New to Mu-43
Joined
Jun 3, 2019
Messages
9
At the moment this is pure conjecture on my part but I am starting to notice that HHHR does not work very well with Panasonic O.I.S.
I am noticing slightly (not sure if it's really of any significant) sharper images using only the IBIS compared to prioritizing the O.I.S. on the Panasonic Leica 50-200mm f 2.8-4. I haven't made any comparative tests yet but I will try and look into it. There are a lot of variables to account for, like the aperture, the focal length and there is also the, invariably, the randomness of HHHR between each shot (it will be extremely hard to change Image Stabilization settings in the camera menu and retain a 100% identical framing between each shot).
Any Panasonic lens users that have tried HHHR on Olympus E-M1 Mark III and X have noticed something similar?

And in the same area, I don't think Panasonic Leica 50-200mm f 2.8-4 is really sharp enough to handle 50 Megapickles images, I want to try the 80 MP HR to see if it's the same or worse. I find the Olympus 12-40mm f 2.8 (at the telephoto end) handles HHHR a lot better.
This raises another issue I've noticed - lower keeper rate using PL 200mm on E-M1ii with Lens IS priority switched on. I see motion blur very similar to what I see when leaving FPS Priority on and IS priority off. I wonder if Panasonic lenses don't communicate when the stabilising lens element is in motion to an Olympus camera - which would account for loss of resolution using HHHR with OIS lenses.
 

Latest posts

Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2009-2019 Amin Forums, LLC
Top Bottom