Twelve examples of mFT vs FF with equal exposure and DOF

Started 8 months ago | Discussions thread
ForumParentFirstPrevious
Flat view
gardenersassistant Veteran Member • Posts: 6,053
Twelve examples of mFT vs FF with equal exposure and DOF
37

This post is long. If you are only interested in my overall impressions from this exercise please see the final section of this post, headed My overall impressions. Please read on here if you want to know why I did the exercise, how I did it and some of the problems I encountered along the way.

Several weeks ago I bought a Sony A7ii full frame (FF) camera. I wanted to see if I could get improved image quality by using a full frame camera, particularly for botanical subjects. Compared to my micro four thirds cameras I was hoping for finer, more subtle, more realistic renditions of colours, textures, light and details, and more “pop” to the images.

In this post I presented comparisons between the A7ii and a micro four thirds (mFT) Panasonic G80 for eight botanical scenes, with both cameras producing the same depth of field (DOF). When viewed as presented at around my normal output size of 1400 pixels high I could not see any consistent differences that favoured the full frame versions; for the whole images, and sometimes for particular parts of the images, sometimes one, sometimes the other looked better. All but one of those who commented expressed views along similar lines.

During a discussion on a similar subject in another thread, Great Bustard pointed out that the A7ii images had been underexposed compared to the G80 images, and so (in my words, not his) they had not shown the full potential of the A7ii. Here he wrote “you might want to try shooting photos of static scenes with FF and mFT, both using the same DOF and exposure (meaning 4x the exposure time for FF), and comparing the photos, both displayed at the same size.”

I have now done this, for these 12 scenes. (Scenes 9 and 10 look the same, but scene 10 has twice the DOF of scene 9).

Capturing the images

I used an EF mount Sigma 105 macro lens on the A7ii using a Sigma MC-11 adapter, and an Olympus 60mm macro on the G80. These are the most similar lenses I have for these cameras. They do not have the same angle of view, although it is similar, and the cameras have different aspect ratios. This means that a choice is needed between having the same angle of view vertically, horizontally or across the diagonal. I started off intending to use a vertically similar angle of view. However, my attempts at this were imperfect.

Also, I had intended to make the centre of the image the same for both formats, but I messed up the horizontal alignment badly for at least two of the scenes.

I used a tripod for scenes 1 to 7. For these I decided to simplify by using the tripod in the same position for each camera despite the difference in vertical angle of view this would cause.

Fortunately I don’t think these various misalignments compromised the overall usefulness of the exercise, or even the usefulness of those particular, misaligned scenes.

Shooting raw, I captured each scene three ways, two ways for the A7ii and one way for the G80.

One of the A7ii shots for each scene was as suggested by Great Bustard – it had the same DOF and 4x the exposure time compared to the G80 shot. It used the same ISO as the G80 shot. I’ll refer to this as the “equal exposure” A7ii shot. (See the response titled “A note on equal exposure” for more about equal exposure.)

The other A7ii shot also had the same DOF as the G80 shot, but it had the same shutter speed too. The A7ii shot used two stops higher ISO than the G80 shot so as to give it the same lightness as the G80 shot. I’ll refer to this A7ii shot as the “equivalent” A7ii shot.

For each scene I captured a ColorChecker Passport reference shot with each camera.

I wanted to have enough DOF to have enough in focus to make meaningful comparisons, and that meant using some middling to small apertures. I wanted to keep the ISOs within the bounds I generally use with the G80. This meant that some of the shutter speeds were on the low side, especially for the A7ii equal exposure shots which were 4x slower than the other shots. That did not matter for the indoor shots, for all of which I used a tripod. But it did matter for the outdoor shots, either because of risk of blur from hand-shake and/or the movement of a close-up subject or distant foliage in the breeze. (And it was distinctly breezy.) The upshot was that two of the A7ii equal exposure shots suffered from noticeable subject-motion blur and presumably there are lesser amounts of this and quite possibly also hand-shake blur spread around the other images, especially the A7ii equal exposure shots.

The captures were exposed so as to protect highlights and this meant that some of them looked rather dark “out of the camera”. This is how they looked on being imported into Lightroom with no adjustments.

The extreme underexposure for scene 1 was deliberate, and not to do with protecting highlights. I wanted to see how the cameras responded to strong shadow lifting of the type that I sometimes want to do.

I often work in light that is not very bright, photographing plants that move in a breeze, and so I often end up using ISO 800 on my G80, which would mean using ISO 3200 on the A7ii for the same depth of field and shutter speed. That is why the 12 comparisons use ISOs for the A7ii Equivalent shots up to 3200.

In order to choose a starting level for the exposures I chose an aperture for the A7ii and then found a shutter speed and ISO which was practical to use and which just avoided any highlight zebras. I used this for the A7ii Equal exposure shot and worked out the settings for the other two shots. As it turned out the ISOs were nicely spread, with the G80 having four shots each at ISO 200, 400 and 800, matched by ISO 800, 1600 and 3200 for the A7ii Equivalent shots. Like the G80 shots, the A7ii Equal exposure shots had four each at ISO 200, 400 and 800.

Processing the images

I used Lightroom CC Classic to process the raw files, loading them with Lightroom’s Auto Settings which makes image-specific adjustments to the lightness (called “Exposure” in Lightroom), contrast, highlights, shadows, whites, blacks, vibrance and saturation.

I adjusted the white balance for each image for each camera from a grey square on the ColorChecker reference shot for that image and camera. I wanted to make the colours as similar as possible for each three images of a scene, and to that end tried using camera profiles generated from ColorChecker reference images. I had also intended to use DXO PhotoLab, in particular its PRIME noise reduction, exporting from there to Lightroom using DNG and then using Auto Settings in Lightroom. This is a workflow I use a lot, particularly for invertebrates, and now that PhotoLab can use DCP camera profiles I wanted to incorporate these into the workflow. It was not successful. Somewhere along the line in and between PhotoLab and Lightroom (which can also use DCP camera profiles) some of the colours diverged rather than converged between the three images in a set (whether using a generic or camera-specific profile).

After much experimentation and frustration I ended up not using PhotoLab and using the default, generic “Adobe Color” profile in Lightroom rather than a camera-specific profile. There was still some variation in the colour rendering, most notably in Scene 2.

I applied the same mild Clarity enhancement and the same sharpening to all the images. For Scene 2 I thought that the A7ii Equal exposure shot did not need any noise reduction and that the A7ii Equivalent shot and the G80 shot should have the same noise reduction. I decided to use the same sharpening values for Scenes 3 to 12, including none for the A7ii Equal exposure shots. (Scene 1 was an exception, having be subject to strong shadow lifting.) The only other adjustment was to reduce the luminance of the reds in Scene 7 by the same amount for all the versions because the red channel was clipping and rendering the petals with large featureless areas of an unrealistically pure red.

I produced two sets of images, one at 3400 pixels high and one at 1400 pixels high. I produced two versions of the 3400 pixel set, one without the noise reduction and sharpening adjustments and one with them. The 1400 pixel high versions had noise reduction and sharpening applied.

The reason for producing the 3400 pixel high versions was to get the most detailed view possible while viewing the images at the same vertical pixel dimension. The reasons for producing a version with no noise reduction was that Great Bustard had said (my words, not his) that the equal exposure shots would have less noise than the other shots and this might explain some or all of what had been described in another thread as a “creamy” look of some full frame images. I wanted to see the extent of the difference in noise.

The reason for producing the noise reduced 3400 pixel high versions was to see how the difference in noise looked after applying noise reduction, and the extent of any side effects such as a reduction in detail.

The reason for producing the 1400 pixel high versions was that I wanted to see the extent of the differences between the three versions, after post processing had been applied, at my normal output size.

Comparing the images

I used 3-way side by side comparisons in Faststone Image Viewer to compare each set of three images. I viewed both the 3400 pixel versions and the 1400 pixel versions at 100%. (I did not use the 3400 pixel versions at 42%, which would have produced around 1400 pixels high, because the Faststone resizing had adverse effects which could have invalidated the comparisons, including making some of the fine detail appear to be oversharpened and making some of the plain backgrounds appear additionally noisy.)

Here is a screenshot of one of the 3-way comparisons, a 100% comparison of 3400 pixel high versions in this case. Like all of these 3-way comparisons it has the G80 version on the right, the A7ii equivalent version in the middle and the A7ii equal exposure version on the left.

I have provided my observations on each of the 12 scenes in responses titled “Comparing the images - Scene 1”, “Comparing the images - Scene 2” etc. These posts include screen shots for each scene like the one above for images that are 3400 pixels high and 1400 pixels high.

In the next section I summarise my overall impressions. Your impressions may of course be different (I am often surprised by the different aspects of images that we variously home in on, and how others notice things that I don’t, or don’t see what I see). Because of that I have provided a lot of images in this album at Flickr. It contains

  • The 3400 pixel high images, with and without noise reduction and sharpening
  • The 1400 pixel high images with noise reduction and sharpening
  • 100% 3-way comparison screenshots for 3400 pixel high images, with and without noise reduction and sharpening (for 9 of the 12 scenes there are two of these looking at different parts of the image).
  • 100% 3-way comparison screenshots for 1400 pixel high images with noise reduction and sharpening.

You should be able to find particular images/screenshots from the file names.

Here is a contents list for the Flickr album.

The look of the images obviously depends on the raw conversion and post processing I applied to them. I can put some raw files up at DropBox if anyone wants to look at them for themselves using their own preferred software and techniques to produce their own preferred output size and style and do their own comparisons.

My overall impressions

When capturing the same scene (shooting raw) with an A7ii and a G80 with the same DOF and shutter speed (to produce “equivalent” images), the results generally look rather similar. They look similar before noise reduction and sharpening is applied and they look similar after the same noise reduction and sharpening is applied to both images.

The major differences are that the backgrounds of the A7ii images tend to be very slightly noisier, while the clarity, detail and sharpness of the A7ii images is better in some cases. (For example see the response titled “Comparing the images - Scene 5”.) These differences are most apparent when viewing the images at the pixel height of the G80, 3400 pixels. When looking at my usual display height of 1400 pixels these differences (which are only visible in some scenes at 3400 pixel height) are sometimes still visible, and sometimes not.

The colours of the A7ii version can occasionally look richer compared to G80 colours that look a bit washed out. (See the response titled “Comparing the images - Scene 5”)

The A7ii handles deep shadow recover far better than the G80, which exhibited severe (I think unrecoverable) issues with noise, especially colour noise, with the deep shadow recovery test in this exercise. (See the response titled “Comparing the images - Scene 1”)

When capturing the same scenes with the A7ii using the same exposure as for the G80 (and the same DOF) the A7ii produces results which are significantly less noisy, with smoother looking backgrounds. The difference between these “equal exposure” shots and both of the other two shots is much larger than the differences between the other two shots. (See the response titled “A note on equal exposure” for more about equal exposure.)

However, using the same exposure means using an A7ii shutter speed four times longer and where the G80 (or the A7ii in equivalent mode) only just has a fast enough shutter speed the A7ii equal exposure shot can suffer from significant blur caused by subject movement and/or hand-shake. (For example see the response titled “Comparing the images - Scenes 9 and 10”)

From a previous comparison exercise I thought that using the A7ii might solve a problem I had been having with “difficult” colours of some flowers. This exercise showed that this was not the case. (See the response titled “Comparing the images - Scene 7”)

Despite my best efforts using a ColorChecker Passport and carefully setting white balances from reference shots, the colours varied between the A7ii and the G80, more for some scenes than others. Where the colours varied I could not assess which was the more accurate. I certainly did not get a feeling that the A7ii was giving me consistently more accurate/subtle/pleasing colours.

In closely comparing the images I did not notice any other benefits in terms of image quality from using the A7ii rather than the G80; no extra “pop” for example.

Panasonic G85
If you believe there are incorrect tags, please send us this post using our feedback form.
ForumParentFirstPrevious
Flat view
Post (hide subjects) Posted by
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
ForumParentFirstPrevious
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow