In part one yesterday, I looked at the camera as a standalone device with few references to its predecessor or competition; today we’re going to examine some of the technical differences in a bit more detail against two benchmarks: the outgoing OM-D E-M5, and the Nikon D600. Both are 2012 cameras, and cameras that I’m intimately familiar with because I use them heavily in the course of my normal work – the E-M5 as my travel/teaching camera, and the D600 for video and backup to the D800E. The former is a no-brainer; the latter is perhaps a bit more of a stretch: not only is there a significant price difference, but the sensor goes up in size by two whole notches – it’s effectively four times the size of that in the E-M1. Surely this is an unfair fight?
Update: ISO comparison chart mislabelling fixed, and I am checking on the 12 vs 14bit issue.
E-M1 against older brother E-M5 with grip extension and full vertical grip; the E-M1 seems larger but is in fact just slightly wider, taller with its vertical grip, and actually shorter if you compare it to the E-M5 with the grip extension piece only.
Here’s my rationale behind the comparison: we want to know if the new camera is better than the old one – it is – and more importantly, by how much. But at the same time, the E-M1 really has no direct competition at the moment: the only other camera that comes close spec-wise is the GH3, and that doesn’t have the same level of build quality, PDAF, frame rate, or EVF quality; it’s really built with a different purpose in mind, too. Bottom line: there are no real pro-grade* DX or compact system cameras out there at the moment; the E-M1 is pretty much it. However, that brings us to the relative price point: in Malaysia, street price of an E-M5 body is about RM2,800; the E-M1 is ~RM4,500; the Nikon D7100 RM3,900; the Canon 7D RM4,000, and the Nikon D600 RM5,500. The Canon 6D is a bit more at RM5,900. I don’t have a D7100 or 7D handy; in any case, I don’t think they’re quite the same level of camera – the E-M1 trades a little resolution for a higher frame rate, significantly larger buffer, and much better build quality. In fact, the on-paper spec is much closer to the flagship D4 or 1Dx (16MP, 10fps, 51 shot RAW buffer, full environmental sealing) than the D7100. Bottom line: if you’re going to spend this much on a camera and you’re not committed to a system, you are probably going to be considering most of these options; especially when another RM1,000 (US$300) gets you full frame, and the D600 is the lightest and smallest there is at the moment.
*Built without compromises, or to protect a product higher up in the manufacturer’s line.
Before we get into raw performance, let’s talk a bit about use in the field and system completeness. The Nikon and Canon full frame lens lineups are undoubtedly mature; there is pretty much a lens for everything, including special purpose optics like supertelephotos, tilt-shifts and macros that exceed 1:1 reproduction ratio. There are also several grades of lenses to suit all budgets and durability levels. M4/3 has the most choices out of all of the mirrorless systems, but we’re lacking the tilt shifts, the pro telephotos, and the weather sealed high grade primes – surprisingly, lenses like the 12/2 and 75/1.8 are not sealed, (though I’ve had no problems with operating them in harsh environments). Flash solutions are a wash for both systems; there are wireless TTL/ commander options and heads of different power outputs.
Mirrorless of course has an enormous size advantage; having undertaken plenty of reportage and travel photography with both, I can tell you that there’s absolutely no question or shade of doubt in my mind – if I have to carry it for any length of time, I’m going with M4/3. Image quality is already more than good enough for large prints – and that’s with the previous generation of sensor. The one final element missing from M4/3 was continuous autofocus capability – and we’ve seen that’s just been addressed by the integration of phase detection photosites on the new sensor. So does there remain any solid reason to pick an APS-C or FF DSLR other than absolute resolution or extreme low-light? I opened a can of worms with an earlier article on the demise of the DSLR; now I’m going to pour that can out onto the table and spread it around a bit.
Here’s how the core feature table looks; I’ve thrown in the D4 for comparison, because I think this is the E-M1’s natural full-frame competition: a tough-as-nails, pro-grade, speed-focused general purpose photographic bludgeon. Green is a decisive win, red is a decisive loss. There’s one thing missing here – that’s the E-M1 and E-M5’s stabilizers, which is a definite advantage. You’ll note that the E-M1 actually seems to have the best balance of compromises, unless you’re a videographer – in which case the Canon (or the GH3) would be a better choice. In fact, what stands out is that the older E-M5 loses on a lot of the categories – yet that was reflected in no real disadvantage in its ability to make great images. More, is of course better. This is of course not a complete spec sheet, but I think it goes to show how difficult a comparison we have on paper. The price is undoubtedly steep – and that’s something I criticised the GH3 for, given that the camera was still lacking PDAF and had a subpar viewfinder even compared to the E-M5; the E-M1 pushes the boundaries higher still. It’s a shame, because I think Olympus must be torn: price it like the tool it is, or go for volume because a lot of consumers are still motivated by size?
I don’t think there’s any doubt that if the core sensor technology is of similar vintage, then the larger sensor will be better on all technical measures; the more pertinent question is just how much better, and more importantly, how far up do you have to climb the diminishing returns tree to see the difference? And that’s what I’m going to try to answer today.
Important testing notes: we do not yet have ACR support for the E-M1, and I’m not familiar enough with the Olympus software to be confident of extracting the most out of the raw files, so testing will be done with JPEGs – I’ll update this portion as soon as an update is released. For all cameras, sharpening will be set to the optimum for the camera – generally about halfway between default neutral and maximum; saturation was reduced slightly, contrast was set to minimum and noise reduction off; basically, I tried to create as good a starting point as possible for processing – in effect a ‘quasi-raw’ JPEG. I’ll use the 85/1.8 G on the Nikon, and the 45/1.8 on the Olympus cameras – stopped down to f8 on the Nikon, and f5.6 on the Olympus to balance off optimal sharpness, depth of field and diffraction. The cameras will be locked down on on a heavy Gitzo 5-series tripod and Arca-Swiss Cube geared head. You can click on any of the relevant links following the images below for 100% crops.
Low ISO comparison – 200-1600 – click here for 100% crops.
Let’s talk about the easy stuff first – low ISO noise. I’d say the most obvious thing that’s apparent from the above swatches is that the D600 has more dynamic range than the M4/3 cameras; unsurprisingly, the ensuing images look flatter and lower contrast (lighting conditions and exposure times were identical for all three cameras). All three cameras are very clean to ISO 800, though it’s also clear that the E-M1 has the weakest (i.e. none) AA filter of the three; the D600’s AA filter is fairly strong – look at the dot pattern in the white CD case. The E-M5’s JPEG and NR engine is noticeably coarser than either the E-M1 or D600; there’s just a hint of splodginess creeping in at ISO 1600. Still, I wouldn’t hesitate to use any of these cameras at any of these settings.
High ISO comparison – 3200-25600 – click here for 100% crops.
Higher ISOs are a different story – the E-M5 is looking very ropy by ISO 6400; I try not to exceed 3200 on this camera. The D600 is still pretty smooth and retaining fine detail well (look at the logo in the black CD case) though color is starting to get very flat and chroma noise is dominating the shadows past ISO 6400. This is actually a little surprising as Nikon’s forte has always been keeping noise in the luminance channel. I wouldn’t use this camera past 6400. The E-M1 is actually keeping pace with the D600, and trades chroma noise for a bit more luminance noise; there’s not a lot of difference in resolution to ISO 6400, but above that the D600 pulls away. On an absolute basis, I think the E-M1 has pulled out a stop from the E-M5 – I’d use this camera at 6400, but no higher. What’s really impressive about the E-M1 is that there are no odd colour/hue shifts going on as the sensitivity increases – look at the red swatch, for instance. (Actual color in real life is somewhere between the D600 and E-M1; neither camera gets it right.)
In this scene, we look at dynamic range. Each camera was exposed until the any of the individual channel highlights just clipped; I used the live highlight warning on the M4/3 cameras, and the playback highlights on the D600. It appears that the E-M1 has a slightly brighter highlight rendition than the E-M5 – the same amount of clipping was visible in both at these settings. I suspect it might have something to do with the color rendering: even though all cameras were manually set to the same Kelvin WB, the E-M1 has the most accurate color of the three; the E-M5 is too yellow, and the D600 is too green. On the E-M1, the histograms for each individual color channel are much closer together – resulting in slightly brighter highlights.
Highlights. 100% crops here.
Shadows. 100% crops here.
No question that the D600 has the most dynamic range of the three; that’s to be expected given that it has the larger pixel pitch by some margin. I’d call it half a stop in the highlights and perhaps a stop in the shadows; I suspect it would be a lot closer in RAW however – the E-M1 seems to have very clean shadows, potentially hiding quite a bit of usable latitude. The E-M5 noticeably trails both cameras again – look at the folded cloth, and the hat band. Out of the three, I prefer the E-M1’s rendition of the scene; I think it’s the mixture of getting the color almost spot on, as well as the added punch from the lack of AA filter; though the two M4/3 cameras are quite similar here, note the softness in the D600 image – even at the point of focus (look at the hat).
Next up is a grab from out of the studio window; we will consider real world resolution and dynamic range.
Highlight rolloff – click here for 100% crops.
Practical resolution – click here for 100% crops.
Once again, the D600 seems to have the flatter image – shadows aren’t quite as dense as the E-M1, but we really need to see RAW files to figure out how much of this is the in-camera processing and how much of it is the sensor’s native response. What I do notice though is the highlight rolloff of the E-M1 seems to be the best of the three, though it shares the densest shadows with the E-M5; the E-M5’s highlights are a bit dull, and the D600 seems to clip abruptly. Differences in native tonal response? Probably. Though the two M4/3 cameras are pretty close on resolution, I’d give the E-M1 a hair in acuity and microcontrast; it must be a mix of the lack of AA filter and new image processing engine; fine detail just doens’t seem as coarse as the E-M5. The D600 is clearly resolving a little bit more than the other two – look at all of the number plates – but it’s really surprisingly quite close. If the D600 also lacked an AA filter, the difference would be much larger.
Here’s the practical challenge, though: all of these tests were conducted under optimal shooting conditions: heavy tripod, magnified live view to confirm critical focus, base ISO. In the real world, you’re not going to be able to achieve that all of the time under the shooting conditions for which these cameras were intended – handheld travel or reportage-style work – which means that you might well not be able to get the same results. I find that the current crop of 24 and 36MP cameras give up a stop or two in shooting envelope – you have to have a much higher shutter speed to ensure critical sharpness, which effectively degrades both low light capabilities and dynamic range. The OM-D twins, however, have that excellent stabilizer that allows you to stick to the 1/focal length rule or below; effectively buying you a couple of stops. Unless you’re shooting action, where shutter speed is critical to freeze motion, this makes a huge difference in practice! If we couple that with the increase in size of the larger sensored system – for both camera and lenses – then the advantage of the mirrorless contingent becomes even larger.
With PDAF on sensor, that last bastion of the DSLR is eroding, too. Practically, the D600’s AF system still tracks better than the E-M1; I spent some time shooting traffic and found that the E-M1 would perform similar to or slightly better than the D200 generation of cameras in terms of tracking ability; I think with another iteration or judicious firmware update, the gap will, shrink even further. Mirrorless will always outdo an SLR in AF accuracy; simply because the exact focus point is also the exact imaging point.
And here we come full circle: I compare the E-M1 to the D600 because it’s the cheapest entry into full frame (and I didn’t have access to a pro DX camera; in any case, none of the current lineup match it on spec either) – and whilst the D600 still holds a bit of an advantage in image quality, it’s not as much as you might think; less in practical application; far more of the difference will come down to shot discipline and how the images are processed. And that’s assuming pixels are going to be peeped: they’re close enough that even at 100% it takes a reasonably trained eye to spot the difference. Everybody will see the composition first, of course. Even if we’d had DX cameras in the mix, the results would be even closer still – if not an even match. Even as it stands, I haven’t observed that much difference in underlying sensor quality between the GR and OM-D; at stop, at most. Most of the difference is due to the optics. Yet despite its sensor, the D600 lags behind in every other specification; it’s not until you hit the full-fat D4 that you can match frame rates or environmental sealing. Bottom line: there is simply nothing quite like the E-M1 at the moment – a very compact professional system camera.
The Olympus OM-D E-M1 is available for preorder here from B&H and Amazon.
The Olympus 12-40/2.8 PRO is available for preorder here from B&H and Amazon.
The E-M5 is available here from B&H and Amazon.
The D600 is available here from B&H and Amazon.
Enter the 2013 Maybank Photo Awards here – there’s US$35,000 worth of prizes up for grabs, it’s open to all ASEAN residents, and I’m the head judge! Entries close 31 October 2013.
Visit the Teaching Store to up your photographic game – including workshop and Photoshop Workflow videos and the customized Email School of Photography; or go mobile with the Photography Compendium for iPad. You can also get your gear from B&H and Amazon. Prices are the same as normal, however a small portion of your purchase value is referred back to me. Thanks!
Images and content copyright Ming Thein | mingthein.com 2012 onwards. All rights reserved