Quasi-gratuitous header image: large format golfball, anybody?
I start this article with a deliberately provocative title, at the risk of being taken for one of those forum sensationalists that proclaims OMG NEW BEST CAMERAR EVARRRR SINCE THE SECOND COMING OF SLICED JESUS!! . But as always, there are caveats: I’m examining the situation under practical implementation, practical shooting constraints, and real world limitations: i.e. non-ideal circumstances, which I believe to be fair since this is how most photography takes place. And since we’re interested in hardware towards the practical application and implementation of photography, this is a fair approach to take. The crux of the argument is this: we have now reached a point in technology where the tradeoffs associated with upscaling your format do not translate into significant gains in shooting envelope or even practical output most of the time. Actually, I’d go even further and say that your hardware choices really hinge on a few factors, which we’ll discuss shortly.
Not so long ago, technological development at a given format size wasn’t symmetric: you’d see bigger advances in certain formats/ sensor architectures that heavily favoured one over another. The same was true for lenses, as new formats got filled out. A good example of this was medium format: the latest generation CMOS 35mm FF stuff was coming very close to the large CCDs in resolution and color accuracy, and beating them hands down in low light performance and dynamic range. Then we got a generational jump, and CMOS 44×33 opened the gap. Current 35mm 42/45/47MP sensors are very close to the 2013-generation 44×33 50MP CMOS; but we’re about to go to the next generation there. However, somewhere along the way, the camera’s potential became much greater than ours, and our shot discipline went back to being the limiting factor again.
Important note 1: we are talking about shooting handheld here, with a given practical size and weight budget: I make no assumptions about the spending power of the individual, but regardless of the depth of your pockets, each of us has a fixed tradeoff between gear weight and how long we’re willing to schlep it.
Important note 2: by ‘output’, I mean both the technical quality of the image and chances of getting something exceptional: this could be because your camera is fast enough, because the dynamic range is high enough, or because you were willing to carry it and walk for longer and happened to be in the right place at the right time. I think it’s quite obvious that if you have a slow large format setup you’re not likely to get a split second documentary style grab shot, as arresting as it may be.
There are always going to be some fundamental tradeoffs that we can’t solve because of limitations of physics and optics, and the fact that technology is equally applicable regardless of format size:
- Larger sensors will require larger lenses for a given aperture and angle of view. The tradeoff is that you compromise aperture, which reduces physical iris opening requirements, the degree of optical correction required, and thus the amount of glass – this brings relative overall size down. But, your depth of field is effectively the same as the faster lens on the smaller format. And since you’re gathering less intensity of light, this has to be compensated for by the larger, more sensitive pixels. Net result in practice? 54x40mm, 50mm 1/100s f4 ISO 3200, or M4/3, 14mm 1/25s f1.4 ISO 100. These are practical numbers, taking into account a) minimum shutter speed to avoid camera shake; b) stabilisers; c) DOF. An ISO 3200 file from a full 645 100MP sensor is great, but then again, M4/3 files at ISO 100 are pretty clean, too. Which brings us to the next elephant:
- Real resolving power doesn’t climb linearly with pixel count; not even by area. Why? Because of the relationship between angle of view, real focal length and sensor size, and then depth of field, aperture and diffraction. It means that if you can get everything in focus in the previous scenario at say f2.8 on M4/3, you’re going to need f8-f11+ on 54×40. And beyond that, you start to actually lose resolving power at the focal plane as you stop down thanks to diffraction.
- Most of the time, stabilisers and faster apertures compensate for smaller sensors. One stop of light is one stop of light: if you gain two stops on the stabiliser, and another two on your lens, that’s four stops: or the difference between ISO 100 and ISO 1600.
- Newer sensors are good, but even the best of them are linear. This means that for every stop increase in sensitivity, you lose a stop of dynamic range. Your M4/3 sensor with 12 stops at ISO 100 will still have 12 stops in this situation; 54×40 with 15 stops has lost four or five of them: it’s a wash, or worse.
- Bigger is harder to make. This is a practical manufacturing limitation: larger sensors require more power to move and stabilise, and inertia means they won’t compensate as fast; the larger the sensor, the less effective IBIS seems to be given the power constraints in a compact body, and further compounded by size meaning resolution requiring more precise control control of a larger mass than a smaller sensor. Larger lenses require higher tolerances across much larger elements, plus mechanical strain on helicoids and the like.
- Higher resolution requires more precise AF. Since your effective DOF decreases proportional to the sensor’s circle of confusion, the actual DOF plane gets thinner and thinner as resolution increases (independently of sensor size): you need more precise mechanisms to focus the lens such that the focal plane and the sensor plane overlap perfectly, and the maximum resolving power is achieved. Moving things in decreasingly smaller increments is harder.
- The final kicker: more pixels aren’t even always visible. Assuming critical focus is achieved, you only see the difference in resolving power at the subject plane: the OOF areas do not require high resolving power even for decent transitions (just good color/luminosity bit depth). So if you’re a shallow DOF shooter, your gains are not necessarily from sensor size or resolution, but access to lenses.
If you have sufficient light that ISO doesn’t have to be raised above base to achieve a hand-holdable shutter speed, you will see an increase in image quality proportional to sensor size providing depth of field is sufficient, and the technical factors (like focus accuracy) aren’t an issue. Whilst earlier cameras were not very demanding on shot discipline and minimum shutter speeds – I remember the 12MP D700 being good down to 1/0.5x shutter speed, or about 1/40s-1/50s at 85mm – you’re deluded if you think that’s going to work with a 50MP body. Correction: you’ll get 12MP of real resolution, because the limitation on resolving power is now your handshake or rate of instability. Raise the shutter speed to compensate (i.e. decrease the displacement of shake at the sensor plane to below the half a pixel width) and you soon find yourself having to raise ISO. Practically, I need 1/200-1/250s with the 100mm and H6D-100c to consistently achieve 100MP worth of resolving power. Less than that still makes an image, and depending on how far you downsize – the results may still look okay. But 100 imperfect MP downsized to 20 decent ones may not view or print as well as 20 perfect ones; this is a function of visual acuity and how our eyes interpret hard edges and resolution.
In short: the absolute output quality is limited because we either hit the limits faster if we want to make the most of the resolution (thus limiting the shooting envelope). Here’s an illustration: the following combinations are roughly the same size and weight, and all have AF. I pick these because I’ve used something close to it extensively and am familiar with the hardware capabilities:
Olympus E-M1 mark II, 17/1.2.
Nikon Z7, Z 35/1.8 S.
Hasselblad X1D, XCD 45/3.5.
Outdoors, in bright light or with a tripod or flash – the Hasselblad is going to wipe the floor with the other two. Under a typical indoor shooting scenario, to handhold for a sharp image at the pixel level, we would be looking at the following parameters (taking IBIS into account):
Olympus E-M1 mark II, 17/1.2: 1/30s f1.2 ISO 200, with some envelope to spare because you don’t yet need to rely on IBIS.
Nikon Z7, Z 35/1.8 S: 1/30s f1.8 ISO 400. IBIS here is less effective, but perhaps 1/15s if you have steady hands.
Hasselblad X1D, XCD 45/3.5: 1/60s f3.5 ISO 3200
The Nikon probably wins on IQ here, but as light falls further – that gap closes rapidly to the point where the Olympus may well land up with a more usable file:
Olympus E-M1 mark II, 17/1.2: 1/8s f1.2 ISO 200
Nikon Z7, Z 35/1.8 S: 1/30s f1.8 ISO 1600.
Hasselblad X1D, XCD 45/3.5: 1/60s f3.5 ISO 12800
At this point the Hasselblad file is the least desirable; no matter how much cleanup it won’t beat the Olympus. The Nikon is about the same for all practical intents; that’s no surprise as both the Nikon and Olympus have roughly the same generation of sensor architecture. And we haven’t even talked about other factors such as focusing and battery life, or that there are f0.95 lenses for the Olympus and f1.4s for the Nikon (at the cost of size, weight, AF etc.)
So what are the few hinge factors? Given we are discussing practicality, the first two are 1) cost and 2) weight. The old engineering adage of pick two of ‘size, quality and cost’ holds true: large and cheap is large format film; expensive and small is premium compact or M4/3, and so on. Most choices are a balance of these factors. The next one 3) is depth of field: do you tend to prefer more, or less? We then have 4) output medium and 5) magnification. That’s really about it. Notice I haven’t said anything about low light capability, AF speed, system completeness, haptics or lenses? That’s because no matter what you choose for 1) through 5), there’s an option that sits somewhere in the size-quality-cost triangle that lets you pick what you want from all of the remaining factors. For example, there are cheap compact consumer zoom M4/3 lenses, just as there are f1.2 and f0.95 primes; there are compact bodies and G9-sized behemoths. The same applies all the way up the tree to medium format: you’ve got older second hand DSLR style bodies, or newer mirrorless. Or newer and larger sensor DSLR, and so on. We have so many lenses to choose from now, and modern ones are so good thanks to computerised design that what you want either exists (possibly at a price) or can be adapted.
Given that stability is the limiting factor, a tripod is always the easiest solution (and has other intangible benefits such as forcing you to slow down and think). Larger format on a tripod will always be better than smaller format on a tripod, but the size of the tripod has to increase commensurately for stability. The other option to bring you back to ideal conditions is use flash or external lighting such that you’re at base ISO and optimal apertures; this of course isn’t always practical. But it is why older cameras do just fine in a studio – you’re able to extract maximum performance, and this has been sufficient for several generations of hardware now.
We haven’t even talked about practical considerations such as haptics, UI and UX: there’s such a thing as too small, and such a thing as too large. The sweet spot for usage comfort probably depends on your physique, but for a given size camera, there’s only so many controls you can cram in before they become too confusing or difficult to handle, or tradeoffs between format and power consumption (and battery life). But the very bottom line is this: for ‘ideal condition’ shooters – tripod or flash or both – decide how much you’re willing to carry and spend, and buy the largest format and highest resolution you can. If you’re shooting handheld, decide on size/price and then DOF – then pick the one that feels right. In practice: unless you have exceptional shot discipline, the difference is much less than you might think and the main equalising factor of light level kicks in much sooner than you might think, too. MT
__________________
Visit the Teaching Store to up your photographic game – including workshop videos, and the individual Email School of Photography. You can also support the site by purchasing from B&H and Amazon – thanks!
We are also on Facebook and there is a curated reader Flickr pool.
Images and content copyright Ming Thein | mingthein.com 2012 onwards unless otherwise stated. All rights reserved
Ming Thein | Photographer 
Very interesting article Ming (and many others I’ve read so far).
It took me quite a while to read the article (including all comments though), really trying to understand. Your general tenor is clear but not so much youre examples and explanations prior these.
Although you try to balance and be neutral regarding formats and their practicability, given current hardware, and you say choose the right tool for the job, I still somehow feel your examples and the “tenor” in the comments section (less so your posts) is somewhat biased. Here is why:
Example 1 you say the Nikon probably wins. Sadly no assessment about Olympus vs. Hasselblad.
Example 2 you say the Olympus beats the Hasselblad and the Nikon is rougly the same (as Olympus I assume)
Take both examples together then the Nikon is the winner.
More so if you took into account that many handheld indoor images capture people / moving subjects, at least as far as I am concerned. And this is not possible with 1/8th of a second e.g. if they are talking let alone they are moving. A missed moment of a person (or a shaky image thereof) is a missed moment, period.
(If one deliberately wants to slow shutter speed for artistic purposes one can do this any time, but also with the Nikon or Hasselblad).
So, if I take this into account your examples should tell us that Nikon is the winner actually. If you would have elaborated more on the Hasselblad, maybe even this would be the winner, I don’t know. Btw, I actually would prefer 1/60th over 1/30th indoors with moving or talking people.
Interestingly a big portion of m4/3 owners seem to use this article to backup their bias that m4/3 is the better package overall. I really doubt this. There are many reasons for m4/3 but not image quality even in non-favorable conditions, with the exception (maybe) of shooting handheld indoor static subjects.
That is not bashing or ranting, I do own and use an Olympus m4/3 by myself, and together with their medium sized lenses, and for example their portrait lens, this is magnificence par excellence in a very small package, very portable and unobtrusive.
But not better.
And given the favorable conditions the bigger formats shine.
In my book, the title would be “Smaller isn’t always worse, or why you can’t see the difference sometimes” together with better examples.
My other “system” is a APS-C Sony with adapted manual lenses, Would this camera have IBIS I preferred it over my Olympus any day, if size would not matter at that day. I will “side”-grade my APS-C tool soon (avoiding up- or downgrade in this context), it will have in-camera IBIS but it will be FF.
Also, I believe the differences in DOF are not so vastly different for equal f-numbers, and depend much more on print size, viewing distance and on resolution/resizing to that print size than on the size of the sensor. Maybe format and resolution should have been better distinguished I believe.
All my above is meant with all respect to you Ming and the commenters.
Best regards
Marc
einer der Gründe warum ich zu einer zB. Lumix G9 würde ist : Serienbilder: 60,0 Bilder pro Sekunde, und ob die dann etwas taugen, keine Ahnung ( im Bereich Fußball und Motorsport zu nutzen ) so habe ich eine D 850, aber Belichtungszeiten: 60s bis 1/32000s gegen Belichtungszeiten: 30s bis 1/8000s
interesting article, also late to the party 🙂
i recently(well it just got out) shot a cover and 36p (20p bw, other color) special for magazine using lumix gx7mk3 with leica 15mm, d850 w 58mm 1.4g , and a7r3 w batis 85mm. so from wide to close but just with different cameras.
reason i used 4/3 was because the client wanted a film look, and a snap-documentary like feeling, yet had no time for film development.
since there was a lot of bw, and lumix has amazing bw science, i decided to give it a try. (ofcourse shooting jpg&raw)
lumix is easy to hold in one hand, just point and shoot. leica 15mm is outstanding though a little wide for my tastes.
using different kinds of film grain in post pro – i was sure i could make camera results look alike (and i did, but with some problems).
the shoot was supposed to be happening during magic hour, artist was not coming out from make room on time and it suddenly became cloudy(japan weather can do that) – so i ended up shooting in cloudy conditions, using last hour of light during the day. 15 min of iso 200-400, than almost in darkness and whatever i could squeeze out using environmental lights.
on a subject of 4/3, panasonic was not in precise focus in like half of the pictures, though it showed i was in focus all the time(it makes you feel scary after. the talk that box on the face does not actually means that camera is in focus are true). sucks in low light coniditions, but manageable. d850 ruled the world, af haunted once when subject was moving. a7r3 feels like a piece of sh while shooting, worst experience ever with evf, ergonomics, especially in the same working time frame, when quickly changing between lumix and nikon. with sony i am always not sure if i am in eye focus when shooting in dark conditions cause camera hunts so much and i am not sure if it actually captures what i need it to capture. i end up not using eye af function.
yet in 95% cases it produces reliable results. but its just not great shooting experience. (imho sony is great in low pace, manual lenses kind of style). also sucks at iso higher than 540 when it comes to skin retouching, just takes more time. btw, the same goes for x1d, if only it had a proper evf for proper portrait/studio work.
its funny, but we ended up choosing between two pictures for the cover, one from nikon, other from lumix – ended up using lumix, mainly because of the richness of color! i dont believe even myself saying that, but thats how it appeared in print, yet it looked less sharp at close look.
m4/3 is bare minimum for a4 editorial, (though bigger poster end up looking well, but posters usually done on better equipment/paper in print house than actual magazine). i used grain so that it appeared sharp, in fact, on monitor lumix never felt sharp unless stopped a little ( maybe its the lens?)… nikon color just not appeared as rich in print (on monitor and data they looked almost identical). as with sony.. sony color is so far from other two, that i could not match them in color no matter how i tried( now i remind you we went for film look, agfa, 80s sakura stock look).
now coming to black and white. m4/3 will always have this jpg-ish look when croped just a tiniest bit, but when used properly it becomes a “film” look that everybody crave for(bless the pana’s bw modes and leica lenses). still, in print some pictures end up looking washed out. not enough information? maybe i need a better m4/3 lens but than it will be as big as any other camera, not point and shoot anymore. without grain, in bw m4/3 is enough for some editorial work (wont get close to lindbergh though, but can go close enough lol). but in color – will suck a lot. especially if you need to do some heavy skin retouching.
d850 raws are soooo flexible, i can make them look like anything and they never show me hard time. sony ends up looking like sony…its a good thing, but a bad thing also. i had to really put effort to make it look like the other two. even in bw. sony and lumix tend to look more washed out in print, than nikon on the same iso.
i wished i waited for z6(z7 is a bad joke for me) , but now i am all invested in a7r3 as sub camera, gotten used to its video features and can get enough courage to sell it for… yeah, no great alternatives at the moment.
anyways, thought maybe you find this info interesting. sorry for bad english.
i wonder if newer, next gen m4/3 sensors will perform better and look less jpg-ish in bigger prints. or m4/3 achieved its peak?
This article was delightful, Ming!
You basically summarised why M4/3 is superior in many areas. I love shallow DOF. I only have one camera and lens (Canon 5D MKII & 50/1.2). I use it strictly for my family and some street photography. Something happened that made me want something else. See, I borrowed my friend’s Olympus E-M1 mark II and I fell in love with it. It was crazy how the shooting envelope expanded especially when light slowly went away. The IBIS was amazing in combination with fast glass. I quickly picked-up how I can use slower shutter speed, low ISO and panned the subject and used the blur as bokeh! Pure isolation! Amazing! I’m hooked! Then things get more interesting when I borrowed a Nikon Z6. It made my Canon feel like it came from the stone age! The cost of it is not too far of a difference compared to the E-M1. It also has the shallow DOF that I was used to BUT this time it has IBIS! The review by Mr. Rajsingh was reinforcement. I wish I have deep pockets to buy both. With a limited budget, I can only maintain a single system. I think I’m moving on with the Z6. I’ll wait for the Ricoh GRIII when it hits the used market.
I’d go with the Z6, too; you gain a stop or two on IBIS on the E-M1.2, but you also gain many stops off the larger sensor…Praneeth has shown me results at 12.8-25.6k ISO that are more than passable (i.e. comparable to 1600-3200 on the Olympus; he should know since he was using one before the Z6).
I enjoyed reading this article, not least because it provides a clear and detailed explanation for my own experiences with APS-C and FF sensors. As much as I think I would enjoy using medium format for landscapes, my output (usually max’ing out at 17×22) dictates the lack of necessity for it. I second the amazing qualities of the new Nikon lenses for the Z line, I can hardly wait for a 28mm to come out so that I can compare it with my beloved Zeiss 28mm ZF2. I recommended this article to a budding young photographer who wanted to talk cameras, along with sternly admonishing him to buy equipment according to output and not according to features. I enjoy these kinds of articles much more than gear reviews. Thanks Ming.
No problem. Looks like you and I are out of luck for the Z 28 though – we have 35 already, 24 is next and 20 next year. 28 doesn’t even seem to be on the roadmap anywhere…
Sean, I broke down laughing when I got to the end of your comment and read this sentence – “I enjoy these kinds of articles much more than gear reviews.”
Quite!
The one thing no one ever talks about is noise reduction software; it’s pretty great nowadays and I wonder whether It has already closed the gap between m4/3 and at least APS-C….
NR software is format-independent, which means that it’s moot: M4/3 goes to APSC, but APSC goes to FF and FF goes to MF…relative capabilities remain the same.
When designing products for eventual retail sale, our old engineering adage was: Good/Fast/Cheap–pick two.
Nothing has changed 🙂
But should your initial-listing “17/1.8” be changed to (match the others’) “17/1.2” ? –fyi
Done, thanks.
Sorry but, what kind of photografy are you talking about?.. Seem statics or not moving subjects… Speed shutter at 1/8 s? To compensate iso and aperture? Only if you take a building photos.. Not a car moving or bus or a person running in your shot
Actually, you can use motion/panning to isolate your subjects rather than relying on shallow DOF…
You had me at sliced Jesus LOL!!!
😉 Not everything works better when combined!
So adorable how the inferiority complex of m43 owners have them trying to measure up to FF all the time. The fact is FF has superior IQ and that cannot be argued about. You chose m43 for other things like the smaller size or good stabilization but not for best IQ. In fact, m43 sensor is so outdated that even 1″ sensors are beginning to look close to that m43 can do.
So stop the nonsense and accept you bought the system compromising IQ for other things and be happy.
a) I think you’re missing the point of the article entirely
b) I also have a H6D-100c, with a full 645 MF sensor. So: what inferiority, exactly? 🙂
Your point 1 – ‘Larger sensors will require larger lenses for a given aperture and angle of view.’ Is not true. Yes, it is often repeated, but it does not stand up to scrutiny. Some examples:
The new CV M 50/1.2 is faster and smaller than any M43 rough equivalent. CV M 50/1.2 49mm long 344g. CV 25/0.95 70mm long 435g. This is a significant difference, and its aperture is not as fast an equivalent. All of the fast 1.2 CV M FF lenses are smaller and lighter than their M43 (less than equivalent) M43 f0.95 lenses.
The Canon DSLR 35/2 IS and the new RF 35/1.8 IS lenses are almost exactly the same size as the Fuji 23/1.4 lens which has no IS. Canon”s 24/70 f4 IS Zoom is smaller and lighter than Fuji’s 16/55 f 2.8 lens and the Canon has IS. There are many, many more examples.
Again: is this an apples to apples comparison? Do they all have the same resolving power? Are they designed with the same relative flange back distance/ tele centricity principles? Other features like AF, robustness, IS? Etc.
You could also argue that there are FF 50/1.1 lenses like the MS Optical version that are the same size as a M4/3 45/1.8, and the 55/1.4 Otus is a monster by comparison: but we’re not comparing the same thing. The MS Optical does not have the resolving power of either of the other two; it isn’t telecentric; it has no AF; etc. You could theoretically also get a cardboard tube, a couple of spares from a spectacle shop and put together an even lighter and smaller 50/1.1, but you might not want to use it 😛
The examples I gave are high quality, high resolving examples – in the case of the first comparrison, and all the FF 1.2 v M 43 0.95 lenses, they are even made by the same company. It could hardly be closer. All the examples, and there are many others, contradict the ‘rule’ you gave (a rule that many others also repeat). All I have done is shown that there are readily available, very, very attractive, fast FF lenses smaller than their very close equivalents in smaller formats. This is enough for me to find FF cameras that are built small, like the Nikon Z7 or the Sony FF cameras, a very viable preference to the smaller formats with their supposed advantages, which when looked into seem to disappear.
I’m not disagreeing entirely, but I am saying there are other factors to take into account making this comparison not quite exact (M4/3 flange back distance, corner angles etc. aren’t the same as Leica M and require rear telecentric groups for correction which adds to size – LM does not). Furthermore, having owned examples of both, I don’t put the Voigt LM f1.2 lenses and M4/3 0.95 lenses in the same performance category optically – and neither is in the top tier. The Pana-Leica f1.2 lenses are much higher performers wide open than the Voigt 0.95s ever manage, for instance.
It seems most people are missing the point to this post: in PRACTICAL situations, it’s closer than we think.
This of course does not mean there are some big exceptions: I’ve always said the Q is a good example, as are the Nikon or Sony mirrorless things paired with the right lenses and with certain caveats. But if we are talking a) AF; b) optical performance matching the sensor; c) IS; d) best of class high ISO performance, whatever that might be – there are far fewer exceptions than it first appears. How many really excellent f1.4 lenses (at f1.4) for FF mirrorless are there? Most are very large and heavy, and even with a small body – your practical advantage disappears.
That there are certain design parameters (like flange back distance) that make the M43 lenses larger, is just part of the observation I’ve made – that they are larger. I don’t see your point here, since all I argued is that they (the examples I gave) are larger – does it matter why?. After all, FF lenses need to cover a larger sensor and that also limits how small they can be. Each format has its restrictions.
Have you tried the new FF CV M 50/1.2 or M 40/1.2? They are very new, and are getting excellent reviews. I make no claims at all about whether the CV 0.95 lenses are top tier or not. If, as you say, they are not, I would expect top tier M43 lenses to be even bigger, thus even more strongly confirming my observation.
I do understand your general point about practical situations, and it is worth being made. I prefer, though, to look for the smaller exceptions in FF and take advantage of those, which I feel gives me more flexibility – small size and shallow DOF if I want it. If I shot sport and wildlife, I might think differently. Anyone who uses Leica Ms is already accepting a heap of compromises. Perhaps with the Nikon Z7 and a good adapter there will be fewer compromises and more advantages with M lenses from 35mm up. I will find out.
Yes, I’ve tried both of the new CV lenses – there’s a degree of micro contrast and ‘bite’ missing (as usual for CV designs, which seems to be intentional since Cosina can also make things like the aforementioned Otuses and Voigtlander made the APO Panthers).
My point is as a system whole there are more small exceptions for M4/3 than FF…and enough issues with say pairing a small M wide with FF mirrorless that it isn’t really workable at all. I’ve tried and am all for the compact solution; unfortunately no dice below 35mm, borderline at 35mm, and somewhat okay above. And then something like the Z 50/1.8 S blows even the best M options out of the water, with near-Otus performance at half the size and weight.
Careful here.
“And then something like the Z 50/1.8 S blows even the best M options out of the water, with near-Otus performance at half the size and weight.”
The best 50mm M mount option is BETTER than the Otus at less than 1/3 of its weight and size (970g vs 300g). The Apo Summicron 50mm paired with the Z7 is the highest resolving combination I have found to date. It is cuttingly sharp into the corners even wide open at It is a VERY small and light setup compared to anything else that comes even close.
Sorry, no, it isn’t; I’ve used one and reviewed it here. It’s a stop slower, and flare performance is poor. The Otus doesn’t have any flare issues, plus f1.4. It’s also quite a bit cheaper, and we haven’t even started to talk about quality control and sample variation on Leica products. I agree on the weight, though.
Ming, you mention QC of Leica lenses then talk about one you used. I can tell you for certain, the copy I have is near impossible to get to flare and is cuttingly sharp at f2. I sold the Otus as the ONLY advantage it had was one stop brighter. On a D850 it was like going to the gym to take it along. The Z7 + Apo 50 is far far superior for what I do.
Seems you should at least be open to the possibility that you are incorrect. Your definitive statement, “sorry, no, it isn’t” is off-putting in what should be a conversation. I can assure you, in my experience, the copy I have is most definitely better.
And, what you claimed was that the Nikon 50s 1.8S blows it out of the water. Uh….sorry, but no. Not now, and almost certainly not 25 years from now when the Leica lens is as good as it is today and the plastic and rubber of the Nikon lens has long ago rotted away.
Fine – six new 50/1.4 ASPH lenses with varying problems from detached aperture blades to decentering to asymmetry; one M9-P that corrupted SD cards at random; the IS mechanism in my Q giving double images after 15,000 shots; four samples of the 24-90 SL lens with focus shift so bad at 90mm that f8 wasn’t enough DOF to cover the shift; a 50/0.95 Noctilux with a separating rear element; and yes, two samples of the 50 APO that had serious flare issues (I’m not and wasn’t arguing about resolution). On top of that, Leica lenses are short throw and extremely sensitive to helicoid adjustment – use them enough and it seems they have a habit of loosening or going out of calibration (not an issue with mirrorless, but it IS an issue with an M rangefinder). After spending a serious amount of money, I’m not wasting anymore. How many identical samples have you tested? Have you even tried the Nikon? Never mind that it is one tenth the price of the Leica, and in 25 years – you can buy another one. As for plastic rotting, we wouldn’t have pollution problems if it did…
I have tried many many 50-ish mm that will fit on a Nikon body. The Sigma Art, Old Noct, many Zeiss variants. I am thrilled with the S lenses so far on the Z7. With a caveat. Nikon is using baked in lens corrections to make them as good as they are. The 24-70 has huge distortion without software correction. It makes me nervous.
If you care to, you might try again. If flare was the one big issue you had, that would be easy to test for with a mirrorless camera before buying. As far as QC goes, I was only saying that you tested a prototype and did so before the lens flare issue was addressed in later serial numbers. It doesn’t line up with my experience at all. I have only a WATE, 90 f4 and the APO as Leica products. None have issues. I’d like the aperture rings to be tighter, but other than that they are beautiful pieces of engineering.
Again, in my 35 years of making images, the Z7 + Apo is the all around most stunning combo I have ever used. Micro contrast by the bucket. Rich color saturation. Beautiful and natural looking fall off separation from sharp subject to background. It doesn’t scream BOKEH but instead is easier to look at. Very neutral RAW files as a starting point for editing.
The 25mm 1.4 Milvus is the other lens that gives me similar pleasure.
As for plastic rotting, it gets brittle and cracks, gets sticky, flakes, etc. Give me metal everyday. Last thing I want is a lens I love that needs to be replaced every x years. To each his own with that I guess.
I agree on the 24-70: even with the corrections there’s weirdness at the edges that look like a mix of field curvature and god knows what; I don’t see the telltale smearing on the 50, though. And without corrections the picture doesn’t change much – you’d have thought by now making a 50/1.8 with a fairly open set of options shouldn’t be that hard!
50 APO flare: all of the first generation lenses had it. There were silent corrections done during production to second batch and onwards, but it didn’t completely eliminate things (I have tried other examples later on, but not to the same extent I shot with the prototypes during development). You can check this with Lloyd Chambers and Sean Reid – we compared results in minor disbelief to make sure we didn’t have single isolated sample issues. The Otus doesn’t suffer from flare to this degree because it has a concave front element (as opposed to a convex one) that controls reflections much better, as well as edge lacquer and AR coatings on all elements to further suppress internal reflections.
I’m all for permanence and metal, but not if the cost is many times what the other alternatives are. I’m a little less pro-metal after asking a similar question to the Sigma engineers and being told that metal is elastic, plastic is binary: the metal can be bumped/shifted out of tolerance enough to affect performance but still being functional (think lenses with barrels out of round whose focusing rings do not turn smoothly) but plastic is either in tolerance and aligned, or not at all (and easy/ cheap to replace). My preference is for the haptics of metal too, but I am not so sure from a performance standpoint. Carbon fibre is also for all intents and purposes plastic…and nobody would dispute the longevity, durability or strength of that material. I would be surprised if there are no plastic components at all inside the 50 APO – there must be some nylon bearings in there at the very least…
basil, a lot of what you say is down to personal choice.
IMHO it’s not really true to suggest that the Otus’s only advantage is an extra stop. If that was the criterion we have to chase, there are new lenses that are much faster. I have two of the Otus lenses (FF/Nikon mount) and from the moment I opened the box and attached them to my camera, I was in love – absolutely entranced by them. For the first time in my life I was seeing EXACTLY the same tones and colours in the viewfinder as I was seeing with my eyes, without the aid of a camera at all. Not everyone wants that – OK , they make other choices.
And the weight – well that’s undeniable – so is the fact that a D850 weighs in heavier than a Z7 – but once again, I am in love with my D850 (don’t spread that around – some of my photography friends would scream with laughter about that, because for a long time I was very cross with Nik, for introducing it just after I bought my D810, and said I’d wreak my revenge on them by never buying one – except that as soon as I had my hands on a D500 I HAD to pair it with the D850). And since I used to do body building for 2 hours a day, 6 days a week, for decades, I couldn’t care less about the weight.
But none of that means that any of these choices is “wrong”. There are heaps of reasons for choosing different lenses, giving different effects – and different cams, for different purposes – and all of that, for different styles. God forbid that we ever start imitating each other and all end up with the same thing!
It’s always easy to cherry pick examples to construct a particular argument. I know because I’ve recently looked to the Nikon Z as a potential replacement for my Olympus m43 gear. However, once you construct a system in its totality (so from UWA through to mid tele at least) then the FF system rapidly becomes bigger. Go to long tele and of course thugs diverge a lot.
I’m not so sure about this any more. We’ll have to see. All I need are small, fast, high quality versions of lenses I want – one of the reasons Leica M has always suited me. So far, the two Nikon Z zooms are small and light. I just hope that they continue to make very high quality optics with a real eye on size and weight. This is, after all, a touted advantage of that very wide mount. It could, obviously give them an edge on the no compromise-regardless-of-size Panasonic path, which holds no interest for me. If cherry picking is easy, then why not?
I looked closely at the Nikon Z. The Z6 plus 14-30 and 24-70/4 would be pretty compact option I agree. Throw in the 35/1.8 and you’d have a very nice three lens combo that would do UWA to short tele very well, with a nice bright prime for low light. I mostly do landscapes so this would work seemingly pretty well. What’s not to like? Well, there are a few factors that still mitigate in Olympus’s favour:
– If I want to go beyond 70mm then things get bigger very quickly. Nkon’s planned Z series 70-200 is a f2.8, which will be large and expensive and not yet available. Adapted F mount options are big and heavy too. There is nothing in the Nikon system that would give me the range nor image quality that that Olympus 12-100 provides in anything like a comparable size and cost package.
– Even if I was happy with 70mm max, if I shoot as base ISO then m43 IQ is more than enough. The Nikon would certainly be better as the ISO rises, but since I mostly do landscapes I am usually at base ISO. Olympus’s exceptional stabilisation is a huge win here. I can reliably do 4s exposures at 24-50 effective focal length with the EM1.2 + 12-100 and hey a50% hit rate at 8s. That’s incredible.
– Features such as Live Time, Live Comp and liveview highlight peeking are massively useful to landscape shooting. Nikon doesn’t offer these.
– Finally, the 60fps electronic shutter inn the EM1.2 plus its frame limiter allows multi frame stacking that opens up all sorts of creative and technical opportunities – reducing noise, increasing resolution, hand held long exposure simulation (e.g 25 shots at 1s separation and stacked is pretty close to a single exposure of 25s), HDR, focus stacking etc
Net, net – I personally can make Mr Olympus great work well for me and its not an a priori fact that FF would rejoiced any better results – and in some cases it would hinder me.
300pf is a VERY light option
Excellent article. Thank you 🙂
What are your thoughts on ISO invariance, with respect to total sensor area? I have in mind the performance advantage of the Nikon D5 in low light compared to the other FX cameras – with respect to image noise and loss of dynamic range. This factor is important in genres such as action (Sport) and Wildlife.
Theoretically: same amount of light hitting the sensor = same appearance for a given output size, EXCEPT more pixels = finer tonal gradations.
Practically: the D4/5 might be cleaner than say a D850, but the D850 tends to have noticeably better color and better resolution…can’t comment on other brands/sensors as I’m not familiar.
Re “D4/5 might be cleaner than say a D850”, this goes to my earlier remark that for some other, much older (1Ds mk3 or D3X vs. resp.. 5Dc or D3/D700 –21/24 vs 12mpx (*my prior post wrongly had “D3s” vice “D3X”)) smaller-vs-larger resolution FF cameras, downrezzing for two folks making the comments got both (a) better noise cleaness & (b) sharper images. But this aspect isn’t one that I see discussed, and I surmise that most folks are content w/qualities of their high-mpx files so don’t consider down-rezzing for any improvement(s).
It’s not apples to apples, unfortunately. Sensor tech itself has moved on, so at the pixel level the 2007 D3 (for instance) isn’t as good as the 2017 Z7. Basically, we’re having our cake and eating it at the moment. Unless there are comparable generation sensors with both high resolution or large pixels – say Z6 and Z7 – then the comparison is somewhat skewed. The Z6, by the way, seems to hold an advantage over the Z7 at high ISOs that downsizing doesn’t make up for. BUT, at lower ISOs you have noticeably less dynamic range – pick your tool accordingly…
Time for me to tell the truth. I was a late starter with digital, so my first few attempts suffered from a lot of ignorance about what gear there was out there, and I’ve retreated from some of them. But I’ve now consolidated on four different cams – all chosen for a reason – and I shoot whichever one is appropriate for the shoot in hand.
They ALL work just fine. They ALL produce “good” shots. I have NO intention of buying another camera – with one possible exception, and that one’s a “proposed” camera that isn’t yet showing any signs of being manufactured & released (I just happen to have intercepted some info about the proposed manufacture of it, which has me dancing with excitement at the prospect of getting one – if and when!)
And I spent yesterday morning having a lovely time, in the company of a photo group sponsored by a guy from one of my camera shops, strolling around the local zoo, photographing the birds and the animals, and discussing our photography as we went. What made it utterly delightful was the friendly interest we each had in each other’s gear – the fact that it was ALL “different”, except for two guys, who were absolutely wrapped in the fact their cameras were identical twins, in every respect – and the complete lack of any attempt at “one-upmanship”!
This is typical of “real” photographers. This is what – for many – makes photography such a delightful hobby or occupation. Sharing knowledge and information – helping one another – being creative – and having fun.
Go swing from the high wires, everyone of you!
And most importantly, making your own and being inspired by others images…
Dear Ming, first I want to thank you very much for your honest and competent information!
I did, to my surprise, currently experienced the net results you described in my photography.
I like to take close-up portraits of cats, their faces with a little bit space around and all this indoor, one could say low light. This is quite demanding, because in M4/3 (Pana G5) I have to use at least f4, better f5.6 to get most of the face in focus (with the wonderfull Sigma ART 2,8 / 60 DN).
I try not to use higher than 1600 ISO and mostly end up with 1/40 or 1/20 sec. and several blurry results because of shake.
So I bought the fine Sony a7II for their FF low-light capabilities and used it with an MC-11 adapted Canon 2,8 / 100 Macro. You know what happened. I had to stop down to f8 minimum, use 6400 ISO for shutterspeed (same as with M4/3) and ended up with results only a little bit better than M4/3. 12800 ISO shows to many artefacts and dynamic range penalties..
The amount of blurry pictures got a bit less, because of the IBIS in the Sony, but the M4/3 was much more convenient and faster to operate, mainly because of the articulated, swifeled display.
The Sony a7II is a very fine camera, but now I think in my environment I would have been much better off with a Pana G85 / G9 and its probably more effective IBIS.
Use the best tool for the job – not the best tool for the internet 😉
A very good article Ming. I’ve been using m43 for about 6 years now, having previously used Canon FF. I’ve found m43 to be a fantastic system that delivers more than good enough image quality for nearly everything I use it for. There are times I wish it had better high ISO, but once I go looking at the implications in size and cost of larger formats, as well as generally a lack of features that I value (5+ stops of IBIS, Livetime, Livecomp, wide range of top quality glass ), I realise that there’s no free lunch that will give me all I need.
Finally, I’ve discovered that the 60fps electronic shutter plus frame limiter in the EM1.2 opens up fantastic opportunities for stacking – which allows all sorts of interesting options to work around those high ISO limitations.
There’s no free lunch anywhere; only tradeoffs that must be balanced.
But there are some techniques that are format independent – stacking is one of them – so the same relative gains apply for the larger formats too…
What about down-rezzing the large-pixel’d files to match the (mere) 20mpx of M4/3? I recall folks w/1Ds vs. 5Dc & D3s vs D700 claiming much better both resolution & noise, for downsized files. !?
And here we get into the equivalence discussion. Yes, everything else being equal, the larger files will be better. But everything else isn’t equal; you’re carrying 500g or less with M4/3 and stabilised f1.2, and 1.5kg or more with FF and slower lenses with perhaps no stabiliser. It isn’t equal at all, because downsizing ISO 3200 on a FF sensor still won’t be equal to or better than ISO 200 on M4/3. If we’re talking ISO 3200 to ISO 800 then it might be different. However, the hardware differences and practical deployment make a case for themselves often for the smaller format. If you have enough light and/or weight capacity say for a tripod or faster primes, that’s something else. But not every subject will wait for a tripod, nor can you carry one everywhere…
But, put the CV 40/1.2 or 35/1.2 on the Nikon and you have a very different camera. Now it is as ‘fast’ as the Olympus and can shoot at ISO 200 too. Now the larger sensor will shine. Yes, of course DOF will be much narrower, which in some cases could be counterproductive, but it could also be something wanted, and it can’t be obtained by the Olympus. So you can argue this by matching DOF as you do, in which case FF and M43 are similar (perhaps) or you can add in the extra versatility of the FF and its fast lenses. In lower light where subjects are moving, there are also a number of reasons why FF would be the camera I would choose. There are also many more very fast FF lenses to choose from, and often they are smaller than slower M43 lenses. An example would be the CV 50/1.2 which is smaller than slower 25/1.4 m43 lenses and two or more stops (equivalent) faster. Look how big and weighty those CV M43 f1.2 and f0.95 lenses are that you mention, and they are only (roughly) f2 and f1.4 lenses, common focal lengths in FF; many FF manual focus lenses of these speeds are smaller and lighter, including CV’s own. Is there even a 25/0.75 M43 lens, and if there were, how big would it be? So it all depends on how we look at it.
I agree, and that’s no longer equivalency – it’s one of those exceptions where you can have your cake and eat it, because there exists a faster lens or a special sensor – just not the DOF.
Regarding DOF, most of the times you can work it out, it’s not true when people say m4/3 DOF sucks comparing to other sensors sizes…just remember, like in Exposure where you have to deal with Aperture, shutter speed and ISO to get the desired exposure, with DOP you have focal length, distance to subject, the acceptable circle of confusion size, and aperture. If you play right with these, you can achieve very similar results with a m4/3 sensor, for example.
Firstly, I never said this. Secondly, if you have to rely on shallow DOF to make a ‘good’ image, you’re missing the point of photography entirely.
Excellent and practical article Ming.
I suppose it would be nice if everyone on the internet was as sensible as you are, but if that were the case, you wouldn’t stand out from the rest!
But in a strange way, it would make my life a lot easier 🙂 Though I suppose a lot of the rest of the internet would not exist: there’d be no arbitrage on sensationalism, anymore.
Funny enough, my challenge with making better images of my 3 yr old has led me to some similar thoughts. I usually shoot in shade, overcast, or indoors for softer shadows, so an abundance of light is not often available. Roughly 35mm-50mm EFL.
You need a minimum DoF suitable for movement, but also a higher shutter speed, and fast AF.
When I used a FF camera, I’d use >f5.6 SS1/60, but the AF would suffer from lack of light. Sometimes the aperture would stop down further to try and focus and not work at all. Net result = lots of out of focus images as the AF systems couldn’t always keep up; sometimes outright failure.
When I used a MFT camera, I’d use ~f2.8 SS1/60, but high ISO performance would be a problem due to the shutter speed. Stabilization, though excellent for longer exposures, was never fully realized. Net result = images with poor dynamic range (also strange and hard to correct color shifts that made skin tones very unappealing as ISO got past 1600).
When I used an APS-C camera, I’d use ~f4 SS1/60, and get unsurprisingly a kind of mix between the two. However it was a bit more manageable overall with a still effective AF and better ISO performance (more on that).
So eventually I’ve settled on a Fujifilm X-T3, but it provided a bonus that I had not expected. For some reason, the higher ISOs on the X-T3 generate very minimal color shift. As a consequence, skin tones still look relatively good in spite of the reduced dynamic range and noise (it doesn’t seem to have the odd waxy look the X-T2 did). Ceteris paribus, on test charts, APS-C should theoretically roughly be as good as MFT 1600 with 3200. But my experience with actually using the X-T3 has made me mostly comfortable going to 6400 without thinking much of it, which I am quite happy to have. I’m not sure if it’s the white balance implantation or something else, but is not a usual consideration I’ve normally heard. I don’t think this advantage is as useful for non-human subjects, or even human subjects in unusual lighting, but for my snapshots it’s been a nice extra.
Right tools for the job and all that – the Fuji files I find tough to work with (and the jpegs have strange color shifts in skies and blues) so they don’t really work for me, but I can see why nice skin tones in low light would be a huge plus for some types of work.
I for one am delighted to read a an informed opinion from someone who has real experience with all of these formats, rather than regurgitation of this equivalency pseudo-science.
If we are not shooting tripod based or in great light, shooting envelope for all formats drops pretty drastically. But its funny to see how, for reasons there than sensor size and related IQ characteristics, the smaller sensors can actually perform a bit better at the extreme ends. Counterintuitive, but makes sense when laid out that way.
But I still cant get away from full frame. Maybe it was learning on 35mm film, maybe its just habit, but the 35mm frame just feels “right to me”
Thanks for making sense.
I think it’s what you’re used to, and the aspect ratio matters a lot, too: after having gotten accustomed to M4/3 and MF – both of which are 4:3 – I keep finding FF a bit too long, or not quite long enough (16:9, preferably). I suspect if I shot with an XPAN heavily, I’d want 24:65…
I suspect that the “what I’m used to” factor is more important than it gets credit for. Subject to the maximum light circles of existing 35 mm lenses (assuming backwards compatibility), sensor/camera designers had an entirely clean slate with digital. Yet, FF = 35 mm. QED. That’s not a criticism though. Presented with a FF sensor and lens, I have no issue mentally converting a scene viewed through a given lens / shutter speed / aperture into a final print using an optical viewfinder (DSLR). I can almost do the same with an EVF (getting used to it). And I am reasonably comfortable with micro 4/3rds, I suspect because my brain copes with divide by 2 / increase depth of field without difficulty, because I am utterly hopeless with APSC. It’s just what Im’ used to.
Agreed. And it doesn’t do any harm that the EVFs give us a true DOF preview, too.
Hi, I actually think that the real sweet spot for photography is aps-c, almost for all the reasons you say but with the added advantage over m43 of the slightly larger sensor and now with the new Lr ed tool, xtrans has come of age. As a long time promoter of m43 my recent findings are that it is falling behind by some margin, especially in latitude and dr which although very good on the m43 format is lacking when put under close examination, hr brings some advantage to the platform but in real world applications is limited.
In practical application, APSC is missing the same lens variety as M4/3 or FF – this more than anything is what leaves it behind. For the companies offering larger sensors, they see APSC as the entry level budget option; M4/3 makers do not have this at all and thus do not provision for ‘something above 4/3’.
Hi Ming. Great article, as usual, thanks.
My take on it: I think we are very lucky to live in a world with some many great cameras, so many different form factors (sensor size, weight…) to choose from and absolutely fantastic new generation lenses. To some extent, technology is spoiling us beyond expectations and beyond what we need (at least for me). And probably beyond what we need to make good photographs. The other day, I was looking at a book from Raymond Depardon (great French photographer) and was thinking that, by today’s technical standards, he would probably not make the cut, but his images are really fantastic. After looking at this book, my conclusion was: don’t buy any gear for the next 5 (or 10) years…
a) Better hardware doesn’t improve clarity of a fundamentally fuzzy idea, and
b) Better hardware requires more skilled operators, but b) is still massively subservient to a) 🙂
“By today’s technical standards, he would probably not make the cut, but his images are really fantastic”
Every time I look at photo books from great past photographers, particularly in the street/documentary genre, I think the same thing.
I would actually argue that the modern obsession with technical perfection has actually made modern photographers a bit timid when it comes to executing an idea. There are so many great images form the old books with crooked/imperfect framing, blasted highlights/shadows, and half out of focus images. But they work because of the clear, concise ideas contained in them.
Its my New Years photography resolution to “loosen up” a bit. The world needs a bit less mega technical thoughtless landscapes and a bit more loose, possibly imperfect, but intelligently crafted images.
Agreed on the loosening. Ironically though – the modern hardware has definitely helped with both loosening and execution…
What you are describing is equivalence, but you’re not quite there yet. Your comment about larger pixels compensating for less intensity of light suggests you still haven’t realised that it is the larger sensor area, not the larger pixles that “compensates” for the lower intensity, and your comments about diffraction suggest that you still haven’t figured out that all systems suffer the same diffraction softening at the same DOF.
When you have the same field of view and the same aperture diameter, everything happening in front of the lens is the same, so you get the same picture made from all the same light regardless of sensor size. You get the same field of view, the same depth of field, the same diffraction, and the same amount of light. The only difference is the exposure (light intensity), because with a bigger sensor that same light is spread out over a larger area.
No, it isn’t quite the same as equivalence because there are format-specific hardware limitations to consider that affect the ideal situation implied by equivalence.
“So if you’re a shallow DOF shooter, your gains from sensor size resolution, but access to lenses.” doesn’t make grammatical sense, or actual sense, I think you’ve a typo.
Thanks, words missing: “not necessarily from sensor size or resolution” – now corrected.
Excellent article! I shall use it therapeutically for my recurring GAS-attacks … 🙂
The fact that it takes such a careful analysis to state the strengths and weaknesses of the different formats reinforces the high level of quality we have available today. Our range of subjects and how comprehensive the system is (lenses, accessories, software) I.S. probably going to determine our choice. I can certainly understand you refraining from preaching for a preferred format.
Whatever best fits the intent 🙂
Just LOVED this article. I love m43’s versatility and so it’s my main format, but it is so tiresome having to read and hear everywhere how I’m missing out on the “ultimate quality” of full frame. They can be complementary, and eventually I’d like to have a full frame kit for specific uses, but it does not take anything away from smaller formats. Rarely have I seen such a well composed piece regarding the mumbo jumbo around sensor area that marketing has spread. Kudos!!!
Thanks!
Ming!! Thank you for taking the time to provide your expert opinion!! As some else has stated now armed with the correct information and not marketing information we can get back to working on the real problem the photographer’s ability.
I am but one small voice in a very noisy environment… 🙂
Thoughtful and precient. Thanks. As someone who has not experienced the wide number of formats you have been able to gain expertise with I can still understand your reasoning and it’s very much in line with my own experience with personal and professional work as a photographer spanning 40+ years. I’m pretty active for a 63 year old and still shoot events with friends who are all friendly and will swap my Olympus E-M1.2 and 300mm f/4+1.4xTC with their Sony A9 with 400mm f/2.8+2xTC or Canon 1DXmkII and 400/2.8+2xTC and when we’re all done and finished, none of us can really tell what camera took what picture in a mixup. But personally, I can tell that I’m experiencing shooting wildlife differently with the lighter gear than they are. In low light, I suspect I’m at a slight disadvantage, but overall, my aim is always for beautiful light and when it happens, I like the m43 system very very much.
I suspect your back and shoulders can certainly tell at the end of the day! You might see something if you print larger, but since all of those are in the 20-24MP range – I don’t think there’s much spatial detail different between them. If anything, your 840mm-e might have slightly higher magnification which would offset the resolution loss. While I still birded (a long time ago) I too settled on 4/3 at the time due to the crop factor advantage – my 500/4 suddenly didn’t need TCs with the attendant loss in light and quality…
A good and timely topic. I am more interested in manufacturers creating tools that make us more creative i.e. better color and ways to manipulate it. While I no longer seek to have more camera sensor pixels (over 40MP camera) I do wish for advances in sensitivity and dynamic range…
One thing rarely discussed is the preoccupation by some to equate “good photography” with sharp, high res or even “good technique” images. I truly think this is one of those things that can make photography suspect as an art form. It limits possibilities… Imagine we judged paintings by the number of fibers the artist’s brush had?
I agree with this: more resolution is only better if it helps translate your idea (more clarity, transparency, less looking at ‘an image of’ rather than the actual object). If you’re not going for an idea that requires it, more isn’t better (and maybe it doesn’t even make any practical difference in some cases).
Only one thing to note: more resolution also equals better tonality because you have more spatial steps to describe the same transitions.
I find tonal gradation is often overlooked as an important component of IQ. Coming from many years of large format negative sheet film work with variable development to optimize tonal range (within limits), I see a lot of digitally-captured photos that lack adequate gradation in the shadows, including those taken with larger sensor formats. Thanks for another great article.
And it’s not just capture, but output – which is what we actually see and perhaps even more neglected these days…
Extremely well put, cuts to the chase. Thank you Ming!
Pleasure!
Ming, thanks great article , once again . I’m being dumb I know but I’m missing something: why is it that ISO needs to increase with larger sensor size ? There must be a component to the physics I don’t grasp.
Resolution, camera shake, and the need for higher shutter speeds to handhold and see tangible resolution gains. Not immediately intuitive at all 🙂
True words, Ming, well appreciated ! All those discussions on sensor size, DOF etc. are boring so much… Being able to squeeze capable but light gear into a small messenger bag is a relief ! For me m4/3 is the sweet spot and probably will remain for a long time to come.
In the end, it’s still all about photography and making pictures (or should be, some forget that). Interestingly it seems all formats seem to converge on an ideal ergonomic side – again not surprising given human physiology – the question now becomes how we can maximize IQ given those limits, and other physical ones like trade offs between format coverage and physical lens size, for instance.
Ming, a fascinating read. As you point out, under ideal lighting conditions MF will win hands down re IQ. But IBIS has been a game changer, IMO, enabling users to successfully hand-hold at slower shutter speeds. This is its feature that is mostly highlighted in testing, but coupled with it is the advantage of being able to shoot at lower ISO settings, thus retaining IQ, and which you point out here.
There is no longer a correlation in format between film and digital any more and that can equally be applied to digital. Format size is still relevant with film, going from 35mm to MF to LF yields noticeable improvements. From what you say here, that won’t necessarily always be the case with digital sensors.
We might eventually get there though, with new and modular designs – per photosite there’s no reason we can’t have equivalence across formats. In essence, more area increases potential IQ proportionately; other ancillary support tech considerations like IBIS (or back then, early VR) and faster glass has always been there too, even on film.
Great article, Ming. 🙂
But these comparisons are for static subjects, right? When subject movement enters in the equation, the shutter speed / low light capabilities start to have a huge impact (concert shooter here talking…).
But the most important part is what you said: probably the camera is better the photographer in most of the cases.
Yes they do, except: try finding a f1.2 or f0.95 lens for medium format; you have some options for M4/3…so even though ISO 6400 might be jellybeans, you gain on the lens side (and similarly, lose on large formats). There aren’t as many rest lunches as you might think…
Well said. And minimum dof requirement for something moving might require stopping down this 1kg f1.0 or so FF bazooka to at least 2.8.
Well written and well said. The best and most accurate article on this matte I have read. You even reminded me of a few photographic basics I had forgotten.
Thanks!
If we need to use medium format to their full potential focus stacking ,as described by your friend Lloyd Chamber, probably is one of the option. However it only solves the DOF part but not if the scene is changing fast. For me m4/3 is the most handy one but I also enjoy the experience of using nikon DSLR (I still prefer optical view finder personally…….I also enjoy using D750 + 45 2.8P the balance is so lovely…..so I have no plan to buy z6/z7). For X1D, I can only see the gain in image quality only in very very limited situation……………although I do appreciate the build of the camera and the balance/ergonomic of the X system camera/lens very much.
Probably it is time for photographer to spend more on their creation rather than chasing after new models……………..as you mentioned the limiting factor now mostly belongs to the photographer 🙂
The limiting factor has always belonged to the photographer. More capability = requirement for higher skills to fully utilise it and show the difference…
One very little point, regarding choice of tripods. Traditionally – and I think it is still valid even as tripod technology has improved with the use of carbon fibre – one could get away with a lighter tripod with a heavier camera, as the heavier camera weight tended to take up slack in the tripod, embedded the feet in the ground, and reduced the effect of wind disturbance. And to this day people hang their bags under the tripod for the same reason.
Actually, you raise a good point about tripod technology: it really hasn’t improved that much compared to camera hardware. I’ve tried lighter tripods, and to some degree hanging a bag off or using a bungee can help wind – but it doesn’t help precise framing and camera sag, especially with longer lenses. No substitute to size yet. That said: I find myself using any form of tripod a lot less these days given the efficacy of stabilisation systems…
I’ve gone back to heavier tripods, Ming. If I need one, then it’s one that’s solid. Top of the range now, I guess, would be one of the RRS carbon fibre ones. The way carbon fibre ones flex has put me off. I now use one or other of two Linhof pro tripods. I’m still working on wind – that’s a disaster with long telephoto shots – I need some assistants, holding a wind baffle or something, some days.
And in other situations, you’re right (as always! – LOL) – the stabilisers these days are astonishingly good – took a hand held shot a couple of years ago from the top of l’Arc de Triomphe, looking towards l’Arche de la Defense, several kilometres away, at 1/100th, and was astounded by the result. The only thing affecting the ability to read the signs on the buildings at that distance was the size of the pixels! There was NO camera movement. Actually I don’t rely entirely on stabilisation systems – having spent decades shooting with a Zeiss Contarex, I like the feel of a camera that has a bit of weight in it – because the weight of the camera contributes to stabilisation, simply through its inertia. I know the lighter cams are easier to lug around, and lighter to hold – but all these things have trade offs, and we have to choose between them – there’s no “one size fits all”, no “right or wrong”, in choosing these solutions.
I hear you!
My Benbo #1 is still my “go to” tripod. I find the infinitely-variable leg configuration means it can be stable anywhere, regardless of the local topology.
But I try not to carry it very far. 🙂
I have a Gitzo 5-series systematic with a long feared (aluminum) column, Arca Cube and focusing rails…that’s a 11kg+ monster I don’t carry at all! :p
Yea, I’m afraid of long columns, too! 🙂
(I assume you got auto-corrected… trying for “flared,” perhaps?
Autocorrected yes…but actually was trying for ‘geared!’ 😛
The weight is fearful though, but at least it’s solid.
Ah, another Benbo user. Although no longer used, I still have my Benbo 2 which I acquired nearly 40 years ago. Big and heavy, but extremely versatile. When I look back, I’m surprised that I used to carry this thing around with a 5×4 technical camera!
It’s the only thing I put my Linhof Super Technika on… which doesn’t see much film, these days.
I bring it to schools where I sometimes give talks on photography. The kids love getting under the dark cloth and seeing each other upside-down.
Well, on the plus side – I remember my LF lenses were much lighter and smaller than anything else save perhaps Leica M and M4/3, so the total weight would be comparable to FF (at least for my usual loadouts). Cheers to optically symmetric longer lenses…
I bought my first Linhof tripod to support my 5×4 Technika, too! I still have it – and I’ve bought another one since, for my macro work – and for my astronomy & other long distance stuff (massive telephoto range!)
Y’mean, like this? :-}
I tend to use myself as the wind blocker – with an open jacket if required!
Resonance in CF tripods tends to be better if they’re thicker section – not just leg diameter but also the walls themselves too.
It means that I can throw away my X1D !! Seems that Olympus is taking a strong hold in the M3/4 market even though not a dominating one.
Not if you print large, or need a lot of dynamic range, or color accuracy – those last two items apply to smaller output sizes, too…but if you don’t need any of these, then you actually may well be served better by a smaller format.
For static subjects, the 80mp hi res raw files have as much dr as FF. Color accuracy is also improved. Great article btw.
Resolution yes, though not 80MP of actual equivalence. Oversampling has to make up for anywhere up to three stops less DR/color information; my recollection of testing is that it didn’t quite manage.
It’ll be interesting to hear your thoughts on the Olympus Hand Held High Resolution mode – it looks like higher ISOs can be used and you get very clean output.
To be honest, it doesn’t interest me enough to get a camera to try it…there are still the same limitations around moving subjects etc. which is where I’d want to deploy the portability of M4/3…
Most helpful indeed! (A link to this post would sit well in the Recommended Gear List).
Sadly I don’t think that will deter the people who need it most; just like the “I DO NOT GIVE OUT RECOMMENDATIONS OR ANSWER ‘WHAT SHOULD I BUY?’ TYPE QUESTIONS” text makes no difference either 😦
Yay Ming!
I’ve put a link to this in my persistent paste buffer. Now, whenever the “equivalency crowd” rear their ugly heads, it will only cost two key strokes to refute them!
One thing you only touched upon: legacy glass with focal reducers. For those of us who are comfortable with manual focus, a fast legacy lens plus a quality focal reducer gives µ4/3rds a real edge in many situations. I’m having a gas with my fast OM Zuiko glass (known for high quality wide open) with a Metabones Speedbooster Ultra. An Olympus OM Zuiko 55mm ƒ/1.2 becomes a 38mm ƒ0.8!
I guess links don’t work here? https://www.mu-43.com/attachments/_a268387-2048-jpg-jpg-jpg.698887
Links work, along with HTML and embedded images. 🙂
I first formatted it as an HTML anchor tag around some of my text. That didn’t seem to come through, so I just pasted the link, and that did come through.
I’m guessing you have WordPress to strip embedded links for security reasons? Or perhaps I just botched the anchor tag syntax…
Nope, no changes to stripping links – probably the anchor tag missing a character somewhere.
It does, but not every lens plays nice with reducers…I’ve yet to find a combination that doesn’t seem to have some visible compromises (though in theory, it shouldn’t – since you are essentially oversampling data/signal…)
So true!
I have three: Kipon Baveyes Ultra, Metabones Speedbooster Ultra, and Viltrox EF-M2 (with a thin EF —> OM adapter). (I used to have a horrible, ~$60 one, but sold it and used the proceeds to buy a lifetime supply of Vasoline™ to smear on the front element — similar results.)
99.99% of the time, the Metabones comes out on top. But especially with longer lenses, the Kipon does almost as good.
I’m still evaluating the Viltrox; I had high hopes that it was the OEM version of the Metabones, as the optical configuration, flange width, and even the packaging are essentially the same. But it seems to have horrible amounts of flare, resulting in washed-out images when compared with the Metabones. So perhaps it is a fairly careful knock-off, without the Metabones coating technology.
In general, I find that the longer the focal length and the smaller the aperture, the smaller the difference between the various focal reducers. It’s when you get wider than about 35mm or ƒ/2 where the Metabones really shines.
I used to scoff at these things. But the more I use them, the more impressed I am, especially combined with high-pixel over-sampling: I can easily tell the difference in resolution on 50Mpx images that are not so obvious on 20Mpx images.
Not surprising as wide angle optics are the most sensitive to changes – especially around the edges, and especially with optical formulae that might not originally be telecentric to begin with. Trying to condense very ‘spread’ lenses can result in very interesting results.
The higher the output resolution, the faster lenses like the Otuses pull away…
@ Jan Steinman: “Now, whenever the “equivalency crowd” rear their ugly heads, it will only cost two key strokes to refute them!”
Unfortunately that won’t work because what Ming is describing IS equivalence, he just hasn’t put all the pieces together yet.
Nick: I’m glad you mentioned this because, though I’m not a technical expert, I read Ming’s article as essentially supporting equivalence by providing precise reasons why understanding equivalence is useful. Perhaps Ming could comment on this.
Again: it’s not the technical equivalence through oversampling etc. that most other sources have been talking about. I take a view that’s altered by practical application, both in what kind of hardware is available/ useful in a given situation and how you choose to output it. It is NOT the same thing, because there remains a clear visible difference at times even under some of these circumstances…
Thank Jesus (or whoever) for Ming Thein!
Can we now, for the rest of my life, get back to enjoying our photography? And ignore these seemingly never ending arguments who has the biggest one?
For sages, this stuff has been frying my brain. Now, at last, an expert photographer with unquestionable credentials has stepped up to the plate with the most sensible discussion of these issues that I’ve ever come across.
And I can get back to playing with my toys in peace!
Thanks. I keep reiterating this: composition doesn’t change. Output matters, but you need to be fully in control for the sensor size to matter (and then you need to have the shot discipline). That is all…