Hand-Held High Res: A Practical Guide
May 15, 2020

Now that hand-held high res (HHHR) has made it to a more mainstream body, I thought I would revisit how/when I’m using this and differentiate from using tripod high-res (HR) aka traditional pixel shift. The requirement that the camera be stationary for HR is perhaps a bit restrictive for those that prioritize mobility. Now there’s an alternative on the EM1X and EM1mkIII, but as with any feature there are caveats to its use.

If I’m honest, 20-something MP, or even 10MP, is by and large sufficient for much of my work given the ways media is consumed today. But in the event you want a little more, there are new and convenient possibilities with these particular cameras. I outlined some general guidance with HR in a separate thread and will focus on HHHR here with a few comparisons to highlight differences between them.

Which Do I Use For ___________?

It’s important to understand the different approaches involved with HHHR and HR, despite the naming convention being similar, they are unique in how they achieve their objectives. While HR, like other pixel shift implementations, moves the sensor in a fixed and pre-determined pattern to boost IQ, HHHR takes a traditional brute-stacking approach, relying on subtle and natural movement of your hands between frames to provide added detail to the scene. I only mention this because it is integral to why the results are different and it may guide you into which approach is best suited for a given setting.

When viewed at the same size as HR, it's clear the advantage of traditional pixel shift (far right), but HHHR certainly offers improvement over the single 20MP frame without requiring the camera to be stationary.

In general, HHHR will provide excellent noise reduction and great detail whereas HR will provide great noise reduction and excellent detail.

This, albeit subtle, distinction will likely guide which strategy you employ. As in the first image above, both are better than standard 20MP frames when they can be used. The choice between them for me falls something like this:

  • Resolution is the #1 priority: HR > HHHR
  • Noise is the #1 concern: HHHR > HR
  • Can the photo be taken from a stationary camera position?: HR > HHHR
  • Is it a long exposure?: HR >>>HHHR
  • Am I in a hurry?: HHHR >>>HR

I should mention on that last bit about timing that both modes require a “BUSY” period for the camera to stack images, so if you need photos in rapid succession, you’re in a bit of a bind. I’ve timed HR “busy” time at ~ 6s. HHHR tend to be a little more variable as it can throw out frames that don’t stack well but expect to wait ~ 10s. I suspect this is directly related to the number of 20MP frames being stacked in each scenario (8 for HR, and 16 or so for HHHR).

Scenes With Motion

Yes, you can have motion in your scene, but the type of motion might dictate which method you use (if either). If it’s a long exposure or slow, consistent moving scene, try HR first.

HR on compact tripod. The lower 1/3 of this image was constant motion as the waves struck the support pillars. HR does much better in this type of environment whereas HHHR will lead to strange stacking artifacts (if the process doesn't fail right away).

If it’s a scene where the subject is relatively still, try HHHR first. I’ve found, HHHR is more forgiving with slight movement (an insect on a flower petal, or a leaf twitching in the breeze). Those mostly still scenes tend to favor HHHR whereas HR can result in artifacts along the moving edge.

Shot at 300mm equivalent - the wind was moving the flower gently but was able to stack successfully.

If there’s a lot of movement in the scene, like waves beneath the pier example above, I use HR and let the moving portions blur out as in the case of a long exposure. HHHR does not do well in these scenarios for two reasons: (1) as the shutter speed gets lower, getting 16 or so consecutive images good enough to stack becomes increasingly difficult. The slowest SS I have managed with consistency is 0.5s but that was indoors in a resting state. (2) if a significant portion of the image changes between frames in the stack the camera will error out. Don’t worry – there’s a fix for this scenario, it’s called Live ND 😊

As with HR, artifacts that might appear around unexpected movement can be masked out by overlaying the standard resolution ORI file (rename to ORF) if need be.

Environmental Conditions

In my previous primer on HR , I mentioned I avoid certain environmental conditions: unpredictable, gusting winds, pedestrian/auto bridges, etc. This is one area hand-held high-res is really handy. Ultimately, the steadier you are the better, but this opens up several possibilities that were not available with HR. My advice is to shoot with as fast a shutter speed as you can get away with – ISO 6400 in HHHR looks very much like ISO 400 in standard resolution.

How Do Results Compare?

You’ll notice from the first image above of the audio recorder, the 50MP HHHR files do not render as much detail as the 80MP HR files but both are better than the 20MP files. HHHR files, just like HR, provide for additional latitude in post. In fact, the HHHR files look cleaner in the shadows but you’ll notice there is less detail visible when compared to HR as in this example.

ISO 200, viewed 1:1. Significantly under-exposed intentionally to demonstrate recovery in post. The HHHR shot appears to have less noise on the clock face but also less detail than HR. Both are better than the single 20MP RAW boosted 5EV in post.

If I had to draw similarities, I would say detail in HHHR files typically resembles a 30-40MP FF image (think EOS R, D810). Maybe more toward the 40MP end of that range with a great lens or in a controlled environment. I would love to see Imatest results from HR/HHHR someday, my assessment is purely observational from shooting side-by-side with other cameras.

1:1 views at equivalent AOV/DOF. All handheld at lowest shutter speed I could manage for the FF cameras. The EOS R was shot with the 24-70/2.8 IS @ 35mm f/5.6, ISO 800, 1/6s. The A7R2 was shot with the Sigma Art 35mm @ f/5.6, ISO 800, 1/6s. The Olympus was shot with the 17/1.2 @ f/2.8, ISO 200, 1/6s. The Sony probably should have been at 1/8-1/10th as it still looks a little soft compared to the others.

What Lenses Do I Use?

As with HR, whatever you got! It helps to know the apertures your lens performs best. Often the subject will determine your aperture for you but don’t be afraid to use the lens you have with you. Even a relative mediocre performer like the Olympus 17/1.8 will benefit, the 17/1.2 even moreso.

Original and 1:1 views for Oly 17/1.8 vs 17/1.2 - both handheld

Smaller primes are often easier to balance and stabilize but I have experienced positive results from larger lenses as well. At the extremes of focal length (ultrawide and super-tele) HHHR becomes more challenging, usually requiring higher shutter speeds. Shallow DOF (see yellow flower above) is also challenging as slight shifts in distance can have rather dramatic impact on detailed subjects.

Example at approx 420mm AOV. The 40-150 + 1.4TC isn't the sharpest wide open but worth a try.

What Do I Use HHHR For?

I use it primarily for landscape, architecture, urban scenes, still life, product photography outside the studio, outdoor macro, real estate. If we were to contrast this with how I use HR, you will notice I have eliminated the long exposure scenes like seascapes, light trails as well as tasks like archiving where I am in a well-controlled environment with access to a tripod for HR which provides more detail. If I am shooting long exposure scenes without a tripod then I turn to LiveND but let’s not dive into that here.

Why Do My RAW Shots Look Soft?

The 25/50MP JPG’s produced in camera are excellent but if you want to further play with the RAW, a word of caution; while Adobe Camera RAW will open a high-res ORF file, it’s probably not where I would start. The DPR Studio Scene “simulated RAW’s” are a great example of this. Download the ORF’s and open them in Olympus Workspace – you’ll see the difference. If you’re just after the noise improvement, it likely won’t matter which RAW developer you use on HHHR files but if you want to maximize detail then start with Olympus Workspace.

Same HHHR file converted with Olympus Workspace on the left and ACR on the right.

How Should I Sharpen for Final Output?

HHHR images, just like HR images require more aggressive sharpening than you might expect. Every photograph is different. I have a preset saved as a starting point in Lightroom for both HR and HHHR: Amount: 75, Radius: 2.7, Detail: 3, Masking: 50 and will adjust accordingly from there. I have seen great results from Unsharp Mask in Photoshop and the AI Module from Topaz as well. This part takes some experimenting to find what works for you.

Can’t I Just Upscale the 20MP File?

I’ll leave it to you to decide. Here’s an example of (1) a standard 20MP ORF viewed at 160% to match HHHR size. (2) The 20MP file upscaled using Photoshop’s Preserve Details 2.0. (3) The 20MP file upscaled using Topaz Gigapixel AI. (4) the HHHR ORF converted to TIF in Olympus Workspace. While the Gigapixel image looks pretty good, it does not match the detail in the HHHR file – most evident in the brush strokes on the wooden shelf and the paint texture on the wall above the photo frame.

Can’t I Just Stack 16 images in Photoshop from Any Camera?

Yes, and I’ve done this myself with success. Let’s look at the difference between each approach using the same camera, same lens, same settings, but manually stacking 16 RAW frames in Photoshop (using this guide) vs the out of camera RAW from Olympus HHHR.

1:1 view at ISO 1600. 16 image manual stack in Photoshop on the left, HHHR ORF on the right.

Both are very good at suppressing noise at ISO 1600 and I’m not sure I could tell them apart in real world use. The big difference is how we arrived there. The Olympus in camera HHHR provides a ready to share ORF or JPG in ~10 seconds, no computer, cables, $oftware, or additional time involved.

Final Words

HHHR in a relatively compact camera like the EM1mkIII is exciting for a number of reasons. Mainly it broadens the envelope for what a crop sensor camera can do on the go. What I’ve learned is that like any camera feature, there are circumstances that will benefit and those that are best avoided. Standard 20MP frames will provide more than enough utility for most use but having a 50MP option for some uses is certainly attractive as are the noise benefits. Ultimately, with these features in the hands of more users I hope to see more creative ideas and uses shared here. Feel free to post questions, suggestions and tips below. I do ask that we try and keep commentary constructive and on topic. Thank you. -IM