Today’s article attempts to answer a question which I’ve been asked quite a few times, both in comments and offline correspondence: what is the ‘medium format look’, and why do we find it attractive?
We must first assume that the output medium is sufficient to identify differences. Beyond the obvious very large print or Ultraprint, if you’re judging images at web sizes on a computer – or worse, a phone – sorry, you’re just not going to see it. A typical web image is less than 1% by area of a 40-50MP medium format camera. There is simply no way you can oversample that much resolution information in a meaningful way to those sizes, unless you’re heavily, heavily cropping, I suppose. How large would you have to go to see the difference? I’d say at least ~4MP (2560×1440, most 24”-30” monitors) or better yet, 4K. And that assumes the downsizing has been done in an optimal way, of course. It’s quite possible that some methods will completely throw away any resolution advantage whatsoever (line skipping, for instance).
What I’m going to attempt to do is break it down into five main categories – for digital – and please feel free to add your thoughts in the comments if you feel I’ve missed anything.
The most obvious of the bunch; there’s no question that at large print sizes and relatively near viewing distances – where you have to move your head to see the entire image – the increase in pixels per inch is going to be very noticeable. The overall impression is one of looking into a scene, or at the subject, rather than at a photograph. It’s a level of clarity that you might not otherwise see. I like to think of this quality as transparency, or immersiveness.
Until Sony decides to try to see how many of its 1.4u smartphone pixels it can cram into a 645 chip, it’s generally true for most cases that the larger the sensor, the larger the individual photosites. This in turn results in a much better signal-to-noise ratio, and higher individual pixel quality – perceived as edge definition and ‘sharpness’. This was obvious at base ISO with older CCD medium format backs, but I think it’s just as telling with the current Sony 51MP 33x44mm sensor because it has even higher quantum efficiency and better signal-to-noise ratio than the older backs, especially at high ISOs.
The optics may have something to do with it, though the previous myth of medium format lenses being better than smaller formats no longer appears to be the case; if anything, the smaller format lenses must have higher resolving power to deal with the much smaller pixel pitch – just not across as much of the frame. I’d say for the most part this is a wash with current (late 2014) product offerings – higher resolving power and smaller pixels, or lower resolving power and larger pixels.
More bits are better because it means there are more possible tonal values for each channel. It results in both higher tonal accuracy – especially in highlights, which benefit the most from the extra bits – with the penalty of larger file sizes. Until a couple of years ago, anything below medium format delivered 12-bit color, both for speed and storage considerations. Medium format captures were measured in seconds per frame, not frames per second – that would never have flown for news photographers or consumer marketing departments. Now that we have sufficient processing power, 14 bits is common across most of the board, with some inexplicable deviations (Sony’s lossy compression, for instance).
It’s worth noting however that not all bits are created equal: just because a camera outputs 16 bit files does not mean that it fills all 16 bits with information, nor does it mean that a 14 bit file is necessary inferior to a 16 bit one if it does fill all of the available tonal range well. There are Hasselblads that make 16 bit files but state very clearly they only use 15 bits; much ado was made of the Pentax 645Z only outputting 14 bit files, never mind the fact that the Phase One IQ250 uses the same sensor and also only outputs 14 bits. Or even more importantly, the 14 bits that come out of the 645Z have more latitude and equal color accuracy to the previous digital back I owned, and more dynamic range than the D800E or D810 – which are also 14 bits. The processing pipeline matters, a lot.
Tonal transitions and dynamic range
We’ve now wandered into the realm of dynamic range. This is a camera’s ability to simultaneously capture luminance variation across a large spread of brightness; the greater the difference between shadow and highlight clipping points for a sensor, the greater its dynamic range. Our eyes are capable of registering perhaps 19-20 stops; note however that perceptually, most of the very brightest stops don’t appear very different.
From a digital capture standpoint, I find that absolute dynamic range doesn’t matter so much as how the sensor handles clipping – very linear sensors with large dynamic range that clip abruptly at either end (especially the highlights) tend to look more unnatural than ones with less dynamic range but smooth transitions to overexposure. The reason why we want more dynamic range is because it makes it easier to use a curve afterwards in postprocessing to manage that rolloff. One of the reasons black and white film has such a distinct look is because it has an enormous amount of highlight dynamic range – the rolloff to overexposure is gradual and very closely mimicks the way our eyes work.
Angle of view, depth of field and depth of field transition
I’ve left the most contentious bit for last: personally, I think the distinctive signature of larger formats is the way different angles of view render. A different focal length on a given format changes magnification. Depth of field remains a function of distance, subject distance and background distance. What changes between formats is the fact that a larger format requires a longer focal length to maintain the same angle of view compared to a smaller one. To cover a normal angle of view – say 50mm-e in 135 format – you need 25mm on 4/3rds, 35mm on APS-C, 50mm on full frame, 70mm on 33×44, 80mm on 645 and ~85-90mm on 6×6/6×7/6×8 (aspect ratios differ).
Clearly, given the same subject and background distances, you’re going to have to use different apertures to get the same depth of field for each format: f1 on 4/3, f1.4 on APS-C, f2 on 135, F2.8 on 645 and around f3.5-4 on 6×6/6×7/6×8. But this doesn’t tell the entire story: the abruptness of the out of focus transition is very different for all of these focal lengths, partially because of the focal length, and partially because it’s very difficult to design a fast wide that’s also a) distortion free and b) sufficiently aberration free that it can still split the scene into planes with a sharp transition.
There are also optical design considerations: to make a say 19mm-e lens for 33x44mm (25mm real focal length) is a different challenge to making another 19mm-e for 4/3rds that has a real focal length of 9.5mm – and both have different coverage requirements. This results in optics that have very different ‘looks’. Subjectively, this means that a lens that covers a very wide angle of view on a larger format just feels less ‘wide’ (usually characterized by geometric distortion, think edge stretching) than the same angle of view on a smaller format. I cannot explain why in scientific terms; I welcome somebody better versed in the technicalities to explain it better and more accurately than I can. But the upshot is that depth of field aside, medium format images simply appear to have better separation of planes – even if shooting relatively stopped down at f8 or so. Personally, I like this because it means I can have my subject in sharp focus, a clearly identifiable background, but still have obvious separation between the two.
Commentary on film vs. digital at larger format sizes
The last thing I’d like to touch on is that we often associate much larger formats (think 4×5”, 8×10” and up) with being almost aperspective: partially because these cameras almost always offer movements to correct for keystoning and perspective distortion but also because they offer full control over depth of field, and most images shot with large format tend to be panfocal. These images in themselves offer a typically distinctive look anyway because they’re almost always shot on film which in itself has distinctive visual charachteristics; the only digital solutions for large formats that retain the whole capture size involve scanning backs, tethering, and are generally impossible to use on anything other than static subjects in constant light – architecture, product and landscape, mainly.
Translating the look of larger formats to smaller ones has been something I’ve personally been trying to work on for some time; I’ve come to the conclusion that it’s not easy. If you operate inside the limitations of output size, sensor and optics, it’s possible to make images that have no identifiable format signature, but not deliberately transfer the ‘large sensor look’ downwards. Sensor technology has been constantly improving, so that dynamic range remains less of a challenge than previously. A bigger problem is the optics: even if you avoid extremely wide lenses and work with everything mostly in focus (or lenses that have sharp depth of field transitions such as the Zeiss Otuses), there’s still a difference, especially at larger sizes. It’s of course very important to remember that despite the obvious (or not so obvious) technical differences, none of the mean a thing if they are not in support of the image and your output lacks the ability to show it…MT
All of the images in this article were of course shot with medium format cameras – either digital or film. For more medium format work, see this set on flickr.
Take your photography to the next level: 2015 Masterclasses now open for booking in Prague (9-14 Mar 2015) and Lucerne (17-22 Mar 2015)
Limited edition Ultraprints of these images and others are available from mingthein.gallery
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