About 16:9

It was grim in 2002: there was one full-frame DSLR and no-one made wide angle lenses good enough to withstand the scrutiny of its eleven megapixels. There were kludgy, low-res, high-rent, bolt-on backs for 35mm film camera bodies and depressingly expensive, meh-res backs for medium-format bodies. Neither had useable ultrawides. And only three terrestrial TV channels, by gum.

On the other hand, the future landed: at last there was a full frame DSLR on the market rivaling scanned film and megabucks MF backs, but with portability, AF, high ISO and system support. With elite macro lenses on our Canon 1Ds, we never had it so good.

Happy days?

Sadly, Canon’s original 16-35/2.8 and then-new 17-40L were demonstrated in our first lens tests to be failures, spawning a meandering quest for excellence through the byways and dimly-lit corners of the marketplace: from 2004-2010 we trialed (and sometimes fabricated) combinations of lenses and cameras that hadn’t previously been combined, in search of aberration-free performance on this new digital frontier.

The Fred Miranda forum was a hotbed of alt-lens creativity in this period: from it, we engineered the first Nikon G > Canon adaptors (now made by Novoflex). Conorus made the first Contax N > Canon adaptors, acquiring expertise that paved the way for the successful Metabones product range.

Exciting times: invention sprang from necessity.

Our gonzo reviews and articles from this period were rooted solidly in practical image-making and serious in their intent, if not their tone. There have always been good technical resources for evaluating  lens characteristics – but shooting charts isn’t the only form of controlled test: it’s possible to learn a lens inside-out simply by using it and making careful observation. Further, it’s the only way to answer the inevitable (prime) question raised by a MTF chart or lp/mm diagram: “Yes, but what does that look like?” Another Fred Miranda contributor from that time, Andrew Lloyd, built his recommendable pay-per-view web resource on a similar results-led principle.

Since then,  many excellent lens resources have grown: Optical Limits (previously Photozone), Brian Carnarthen’s Digital Picture Comparometer tools, LensTip, and DXO today lead the pack for data-led evaluation. But all these relegate actual pictures to ad-hoc endnotes, not directly comparable in the way – for instance – Imaging Resource’s library of test images are.  However, they keep going, accumulating reviews monthly – whereas 16:9’s reviews petered out . . .

I felt they were no longer needed, so cracked on with taking pictures.

From 2007, Zeiss tweaked and re-released their Contax and Yashica-mount designs in native Nikon ZF and Canon ZE formats – obviating the need for adaptors. In 2012, Nikon’s game-changing D800 enabled the game-changing Nikkor 14-24mm f2.8 to be natively deployed. All our hopes and dreams came true.

The good news kept coming: Canon struck back with the 5DSr in late 2015, ushering in a new (and until recently, unchallenged) level of 50MP fussiness. This time, they learned from their mistakes – upgrading the lens range prior to release of a demanding camera. Since 2015, the combination of Canon’s 11-24mm and 5DSr gives performance undreamt of a decade earlier. What’s not to like? The Mark II 24mm T/S and a slew of lens upgrades gave Canon’s system unassailable dominance in the Canikon wars for several years. It remains to be seen whether the Nikon D850 is enough to redress the balance, or whether the company’s market share will shrink to a minority of loyalists, because the game has changed again . . .

Sony’s rise, predicted by this website in 2009, looks unstoppable in 2018 – and their top-flight offerings have few faults. In 2013, Sigma debuted the first 35/1.4 ART lens. Who’d have predicted that Sigma would become the brand of choice for many demanding pros purely on quality criteria? And then there’s Micro 4/3 – a genuinely new force in the marketplace: many of the benefits of traditional SLRs, and soooo cute already! Zeiss press on with ever more exotic  designs at ever increasing prices. Truly – when it comes to our kit, we photographers today have nothing to complain about. Almost everything is eminently, professionally useable – even our cameraphones.

Where’s the story in that?

However . . . we have been tempted out of retirement by new developments, and perhaps some new problems. Frankly, it seems sybaritic to complain, but one could argue that nothing’s happened in the last five years to materially improve image quality. Yes, we’ve seen large strides made to improve AF performance – and baby steps toward expanded DR – but a well-crafted capture from a five year-old Nikon D800E isn’t far away from today’s state-of-the-art. And that rankles just a tiny bit. Sensor resolution is trending toward big lazy pixels, seeking grainless high-ISO and enhanced dynamic range. Canon’s 5DSr feels like an athlete trained to win a competition no-one else entered.

The last decade witnessed droves of stills photographers nibble toward film-making, and (not coincidentally) get fond of the rapidly expanding Micro 4/3 ecosystem. With pro/am appeal, there’s something about it – it has kitten factor. I’m a late convert. A Panasonic GH4 gathered dust in my kit cupboard for about eighteen months before I shot with it at all: murky, overprocessed files with no DR from a 16MP sensor with a thick AA filter, the wrong default ISO, uninspiring ergonomics, lousy AF tracking, an uphill struggle to defocus anything and lenses whose performance I couldn’t reconcile to the hype surrounding them. Emperor’s new clothes, anyone?

Here is a lens system whose guiding design maxim seems to be: “fix it in post.” Stop worrying about gross barrel distortion: squish the geometry in software; no-one’s interested in the corners anyway. Never mind the quality, feel the weight. When will the novelty of cheap, miniature lenses wear off, I wondered? Real problems remain with wide lenses on M43 – issues may be baked into the choice of such a short flange distance that are not amenable to the retrofocus solutions that worked for rangefinder film cameras. I’d settle for serviceable corners a stop or two past diffraction, but we’re not there yet . . .

However, M43 has real perks, apart from its bijou form factor and natural predilection for long focal lengths: the latest Olympus bodies, and the Lumix G9, finally render credibly clean, unfiltered 20MP images. AF, though distinctly inferior to phase detect DSLRs for rapid tracking, is actually superior in low light, and pegs level for shooting static or gently wafting subjects. It’s a dream platform for adapted lenses: 35mm optics support shift movements. Olympus and Leica put serious work into their reassuringly expensive top-tier offerings. But the killer app is in-body stabilisation. Not (just) because it’s great for people with wobbly hands, but because the technology made possible precise pixel-shifted multiple exposures.

The technique isn’t new: image stacking has been a staple of forensic analysis and astro-photography for decades; early MF backs used it to boost resolution and lower noise. Given the lag between the first sensor-stabilised camera bodies and the appearance of the first pixel-shifting bodies from Pentax, Olympus, Sony and Panasonic, it’s hard to shake the impression that image stacking was an unintended bonus, not a design objective. However, it’s here, and for static shooters (architecture, landscapes, products, etc) it’s news.

In the articles section, check the surprising comparison between a £4500 DSLR and a pixel shifting £1500 M43:. Not only is a Lumix G9 is capable of producing a better image than a Nikon D850 or Sony A99 II, the margin of superiority is as great as the difference between today’s cutting edge 42MP cameras and, say, a Canon 5D Mark II from 2008.

In theory, you don’t now need ultra high-resolving optics to achieve ultra high-res images. Which is just as well. However, the technique mercilessly amplifies  aberration: the ideal lenses may not yet exist; it’s all a bit new and sparsely documented; we may have to go off-piste for hardware and software fixes – but the results promise to be worth the effort. A few people are quite excited about it. In fact, it feels a bit like 2002 again . . . .