White Label Chinese 50mm f1.1 v Sigma 50mm f1.4 Art v Canon 50mm f1.2 L

To get straight into the test, jump here >. If you’re interested in a digressive preamble, read on.

Digressive Preamble

In the beginning, fast lenses were needed to gather light to quicken exposure times, freeing photographers to shoot spontaneously – to move, and to capture motion. Faster lenses enabled them to work in darker environments and use higher quality, less sensitive film. ‘The Look’ was secondary. In fact, shooting on larger formats, fast lenses are more nuisance than a benefit: it’s difficult to get all your subject in focus, and it’s too late to regret your mistakes after the film is developed.

In the digital era, however, we have instant feedback. And we’ve been liberated by grainless high ISO performance and image stabilisation. We don’t need fast lenses any more – but we still want them. Partly because we’ve not reached the ultimate limit of being able to freeze a hummingbird’s wings in pitch darkness without a flash – the affordable, hand-holdable 20-1200mm fixed aperture f1.0 lens is yet to reach the market. But partly because of ‘The Look’.

Not long ago, the 35mm format was sneered at by many professionals: the equivalent of a cameraphone: acceptable for snaps. Now, we call it ‘full-frame’: for some, it’s the digital Gold Standard. But the smaller the imaging area the more the camera wants to render everything in focus. This is a problem for photographers, who need precise control of rendition: a sharply definable focal plane that can be wielded like a scalpel rather than a blunt instrument.

And then there is ‘The Look’. Fast lenses are now cheap and therefore drop-focus imagery is commonplace – so much so, that many taste-makers are pushing back against it. And yet, most consumers of images in the early 2020s still perceive it as ‘expensive’. Brides pay extra for it.

Whether (and when) ultra shallow depth of field is wise is a topic for another day – in fact a book – but in summary we can say that the degree to which a subject should be isolated from its background depends completely on context and intent. But to have the capability of control – to be able to choose how much of the background to include – will always be desirable. You can’t have a lens too fast.

Similarly, a larger maximum aperture improves peak performance. Lenses tend to reach their highest Zone A resolution two stops down from wide open. Getting there at f2.8 gives you a wider range of uncompromised performance before diffraction takes hold than a lens with a maximum aperture of f2.8, which may not fully come on song until f4 or f5.6. Also, lenses are rarely free of optical vignetting and uncomfortable bokeh effects at their maximum aperture. An f1.2 lens gets all that out of its system by f2, and will usually have a sweeter rendition between f2-f2.8 than a lens that maxes out at f2.8. The principle is similar to the behaviour of a supercar at 100mph versus a standard saloon: they’re both travelling at 100mph, but the performance reserve of the more capable vehicle makes it safer and more comfortable at that speed.

The desirability goes hand in hand with cost: faster lenses are bigger, harder to make, and more expensive to buy. Even if no-one wanted them, they would still be costly – more so, because of economies of scale. As it is, a manufacturer can always guarantee a market for a fast lens. Enduringly so, when we consider how well old, fast lenses retain their value.

At the pivotal 50mm focal length we have the widest choice: thousands of f1.7-f2.8 lenses and hundreds of f1.4 options, but mere dozens at f1.2 and faster. Approaching f1.0 they become shrouded in a mystique not conferred on slower lenses. They acquire monikers: Canon’s f0.95 ‘Dream Lens’ occupies the same enshrined headspace as Noctilux and Nikon’s Noct. I christened my Canon 85/1.2 L ‘The Goldfish Bowl’ – but we digress. All these lenses are relatively expensive.

Skimming eBay, we find the bottom rung of the ladder occupied by the common Nikon 50mm and 55mm f1.2 AI and AIS lenses from around £175 in useable condition – none younger than half a century old. Olympus and Pentax made something similar. There are Porst, Reveunon and Tomioka variants for M42 – all typically fetching upwards of £200. Canon’s f1.2 L, discontinued in 20XX, is still valued at £500+. Contemporary lenses built for mirrorless, with larger diameter mounts, are uniformly superior and disproportionately more expensive – often exceeding £2000. Stirred into this mix, a number of Chinese manufacturers are remaking ‘classic’ fast 50s with modern coatings and production methods. Many are excellent. Some now cost almost £1000. Others can be had for as little as £150.

The Wise Man once said: “a lover of silver will not be satisfied with silver”. I don’t know any serious heaphone listeners who settle for a single set of headphones, and I don’t know any photographers who don’t like shooting at 50mm – or who settle for just one lens at this focal length. You need at least one 50mm to maximise low-light capture: something in the f1.0 to f1.2 range is about as good as it gets. But such a lens will probably not autofocus very quickly, or be good at short range, or have minimal focus breathing. It will be large, and if the lens you want is native mirrorless, it may also be out of your budget: lenses like Canon’s RF 50mm f1.2 are the wrong side of £2000. An ideal companion to a very fast 50mm is something like Panasonic’s 50mm f1.8: as swift as that system allows when shot-grabbing, light, cheap, perfectly useable at all apertures, and with minimal focus breathing. In 2024, no other 50mm offers such a favourable blend of attributes – but only for L-mount users.

However, such modern lenses might leave you feeling the need for more characterful rendition – and with so many cheap 50mm lenses in humanity’s back-catalogue, the temptation is strong to go shopping. Also, even the best of all these lenses suffer by comparison with dedicated macro lenses at 1:2 and higher magnification. For most subjects, a longer working distance is preferred, but a 50mm Macro still has its place. And so the minimum number of 50mm lenses required creeps up to three, or four.

Shooting for ‘The Look’

If you’re shooting a ‘full-frame’ 50mm lens at f2.8 or narrower apertures, you’re likely shooting in reasonable light, and a less extreme prime or a zoom lens may be preferable. For instance, stopping down some f1.2 lenses to f4 results in polygonal ‘bokeh balls’ that may be undesirable. On the other hand, stopped down to f2.8, most f1.2 lenses render consistently circular bokeh balls across the frame, whereas almost all lenses with a maximum aperture of f2.8, shot wide open, will distort bokeh balls centripetally into ‘cat’s-eye’ shapes – symptomatic of mechanical vignetting. The f1.2 lens will be equally problematic at its maximum aperture, of course.

At f2 and wider apertures, the rendition of the defocused areas becomes more influential. It’s sensible, then, to choose a fast lens whose behaviour at these apertures is desirable, without worrying too much above how it performs in the f4-f8 range – the preferred domain of a slower lens. Scaling up, you’re likely to shoot a 50mm macro lens at f8-f16 – in this use-case you may not be too concerned about wide-open sharpness or a bright maximum aperture – and indeed there are advantages to choosing a macro lens with a f4-f5.6 maximum aperture when f8-f16 is the preferred range. Also consider that when shooting at very wide apertures (f1.0-2.0), you may also be shooting at higher ISOs – the effect of which is to blur differences between the resolving power of different lenses.

What is needed, then, is a big light-gathering prime optimised for bokeh and wide-aperture rendition over sharpness, plus a well corrected prime optimised for f2.8-f8 with rapid autofocus, plus a lens with a large maximum aperture optimised for f8-f16 at close range. Most modern 50/1.4 lenses are a good combination of all these capabilities, but they won’t draw quite as sexily as a dedicated faster lens, and they won’t AF quite as quickly or perform as well at f5.6 as a slower prime, or be quite as good close up (especially at f8-f16) as a dedicated macro lens.

Before zooming in on the trio of lenses selected for this comparison, I’d like to mention two I’ve known well that narrowly missed inclusion. These were chosen for ‘The Look’. The Brightin Star 50mm f0.95 is not a particularly sharp lens at any aperture. It doesn’t compare by any technical metric to the Sigma 35/1.2, for instance. But it’s hugely cheaper, it gathers more light, and it has adequate contrast, pretty sunstars at unfeasibly wide apertures, heavily blitzed bokeh prone to an oil-painting-like blending effect, and isn’t afflicted with too much spherical aberration – ‘the glow’ is subtle and (to me) more beautful than the typical vintage aberration vibe. Many years ago I had a Minolta 58mm f1.2 – a lens whose distinctive and appealing style is not easily forgotten: wide open, all subtle glow, bubbles and cream; smooth as silk at f2 and sharp as steel at f4. Now shooting cross-platform with GFX, 35mm and M43, the better vintage Rokkor lenses are triply useful – with large enough image circles to cover 45×33, and Speed-Boostable or directly adaptable for M43. We’ll be looking at ‘the look’ of these lenses on GFX shortly, too. But what do we mean by ‘The Look’?

First, there is no ‘The’ look. There are several (but not infinitely many) ‘Looks’. Most obviously this centres on bokeh and field curvature. The best modern lenses have comparably flat field rendition and bokeh free of distracting artefacts, and there is relatively little variation in presentation as the image circle is traversed. Vintage (and poorly corrected) lenses vary significantly across the image circle: creating rings – sometimes patches – of relative defocus; stretching and deforming shapes outside the focal plane, creating swirly bokeh. There is also considerable variation in how foreground and background elements appear to be blended: some lenses draw outlines around bright edges: making ‘soap-bubbles’ of specular highlights. These lenses appear to stack defocused elements on discrete layers; other lenses stir them together liked mixed oil paint – sometimes roughly, as with brush-strokes; sometimes smoothly, as with water. Bokeh can subjectively feel ‘edgy’ or ‘calm’. As alluded to above, the shapes and properties of ‘bokeh balls’ is a topic in its own right: varying as to shape across the frame, how strongly they are bordered or blurred, and what appear inside them – including onion rings, spots and other artefacts.

Contrast is highly variable, too – a joint function of coating design and interior treatment of lens housings: Average modern lenses tended have high macro- and micro-contrast; older lenses tend to have one or the either – or neither. The combination of contrast and bokeh smoothness gives rise to the phenomenon of ‘3D pop’. Also connected with coating and mechnical construction are flare properties, which vary majestically from design to design. Colour rendition is another factor: varying in intensity of saturation and amount of cast.

‘The Look’ is influenced strongly by the number of lens elements and design scheme. Modern lenses tend to have many (typically seven or more) elements – essential for the high degree of aberration correction we demand from state-of-the-art optics. However, there’s a growing forum sentiment that simpler lenses (four or five elements) have a more appealing, or natural, rendition.

It’s interesting to compare the reputation – and valuation – of cine lenses in this regard. Given the relative ease with which 4K sensors can be resolved, DP’s often focus on ‘The Look’ above properties that may seem fundamental requirements to photographers – for instance, happily living with a high degree of barrel distortion that would be offensive to stills shooters. The lenses most often rehoused tend to be relatively simple designs from the 1960-1980 period. Lenses accorded cult status, or are still regarded as a benchmark of a natural, cinematic rendition, have more in common with simple enlarger lenses than a G-Master or Sigma Art prime. Perhaps best epitomising this is marketing by the new owners of the Taylor Hobson lineage who trademarked ‘The Cooke Look’ as a way of defining a certain (proprietary, of course) smooth/sharp combination that has powerfully influenced the visual style of 20th Century cinema.

Why These Three?

Comparably cross-platform compatible, a Chinese-made 50mm f1.1 available in Leica M-mount demonstrably has the spec and ‘The Look’: minimal glow and very smooth bokeh. The same lens is known as Mr Ding v1.0 (somewhat evolved into later versions, but not fundamentally altered), Artalabs, Syoptic, Vlogmagic, DI Optical or even unlabeled.

This trio is distinguished by being a cheap way to get an expensive look: the 50/1.1 is £160 new; the Minolta 58/1.2 is around £250 used, and the Brightin Star 50/0.95 is £395. Why these three of all available options? The 50/1.1 makes its case by simply being cheaper than the alternatives: it has a 50% head start for value. Of all the vintage options (many of which I’ve tried or owned), the Konica Hexanon 57/1.2 and Minolta 58/1.2 stand tallest: the Konica has nicer bokeh, but not such rich colours or pretty sunstars – and it doesn’t cover GFX. Of the better Chinese options, the Brightin Star is the most complex design at this price point and has proven to be useful. Going upmarket, I would be looking at the Voigtlanders, Laowa Argus 45, TT Artisan f0.95 and a used Canon L – but that’s a test for another day. Separated by as wide a gulf again would be a test of current mirrorless 50/1.2 lenses – much better optically; much more expensive.

White Label ‘Mr Ding’ 50mm f1.1

Construction: 6 Groups / sometimes given as 7 or 8 elements
Diaphragm: Sometimes given as 10, 11 or 12 blades. This sample X
Mount: Leica M (adapted to GFX and Nikon Z)
Weight:
Coatings:
Manufacture: current
Price: £160

Minolta 58mm f1.2

Construction: 7-Element / 5-Group
Diaphragm: 8-Blade (Curved)
Mount: Minolta
Weight:
Coatings:
Manufacture: MC Rokkor PG (Knurled metal ring, Thoriated: 1996-1972 [#2004860-2576349) / MC Rokkor PG (Rubber ring [#2587080-2600885 ] / MC Rokkor (not marked PG): [#2747410-2770060]

Brightin Star 50mm f0.95

Construction: 10-Element / 9-Group
Diaphragm: 10-Blade (Straight, with subtle convex curve)
Mount: Panasonic L, Nikkor Z
Weight:
Coatings:
Manufacture: current
Price: £395

Near-Field Resolution (35cm Distance)

f1.1-f1.2

Comparing the Canon f1.2 L at f1.2 to the Ding 50mm at f1.1, we see neither perform at a high level – certainly a long way behind designed-for-mirrorless 50/1.2 lenses such as the Sigma Art, Sony G, Canon R and Nikon Z. Spherical aberration is particularly intrusive with the Chinese lens, which isn’t the best near-distance performer at any aperture, least of all wide open. Zone C resolution falls a long way behind Zone A with both lenses.

f1.4

The Sigma 50mm f1.4 is very much sharper wide open at f1.4 than the Canon L or Ding 50 stopped down to the same aperture. The difference is clearly noticeable in Zone A and striking in Zone C. The Sigma Art shows almost no degradation in performance at its minimal focal distance. The Canon is still resolving better than the Ding across the frame.

f2

Naturally all three lenses step up sharply in resolution at f2. We find a similar framewide performance delta between the Sigma Art and Canon L, and the Canon L and the Ding – which lags a fair way behind both.

f2.8

At f2.8 the Ding performs closer to the level of the Canon L, but both are still outpaced by the Sigma Art, which is close to its resolution peak in Zone A. On a 33MP Sony A7 IV, the Art lens seems to have parity between Zone A and Zone C rendition – an illusion exposed by cameras with more demanding sensors. By contrast the Canon and Ding still show noticeable softening as the image circle is traversed – the Ding slightly more obviously, though not to the same degree as at wide apertures: it resolves quite similarly to the Canon L in the corner. For comparison, I included the Sony GM II 24-70mm at 50mm: in Zone A (wide open) the Sony zoom cannot match the Sigma Art prime, but it is very close to the performance of the Canon L, and resolves better than the Ding. The reference Sony zoom improves on the Zone C offered by both the Canon L and Ding primes, but again doesn’t reach the level of the Sigma Art.

f4

The Art lens now appears to resolve the sensor in all Zones: it’s not possible to tell any difference with 33MP between the frame centre and corner. At f4 the Zone C rendition of the Ding improves greatly and is now slightly sharper than the Canon L. In Zone A the Canon retains a slight superiority over the Ding, but both are still notably inferior to the Sigma Art. The Sony GM II 24-70mm continues to outperform the Canon L and Ding primes by a small margin in Zone A, and by a slightly larger margin in Zone C. The zoom is almost as sharp in the corners at f4 as the Sigma Art prime, but doesn’t quite match its resolution until f5.6-f8.

f5.6-f11

The only meaningful question to ask about 50mm primes in this aperture range is when (or whether) the corners catch up with the centre. I’ve never seen such a lens fail to deliver acceptable sharpness in Zone A at f5.6 – at least not one made in the last half century. Using the relatively blunt instrument of a 33MP sensor, there’s only a small difference in perceived sharpness between this trio. What difference there is still favours the Sigma Art, but all three resolve well enough – the Canon L and Ding continued to have lower contrast rendition. However at f5.6 there also continues to be a disparity between Zones A and C. The Ding delivers slightly better corners, but at f8 the difference is minimal. By f11 there’s little difference between all three in terms of framewide resolution.

Far-Distance Resolution

In summary, the picture is similar to the above, with one exception: the Ding is poorly optimised for close-up work; at working distances of 5-50m, we see a repetition of the Sigma’s superiority to both Canon L and Ding primes, but the gap between the runners-up narrows – reaching parity at f4. The spherical aberration blighting the Ding’s performance wide open at 35cm diminishes almost to an acceptable level at working distances of 1-5m, but returns somewhat approaching infinity.

Summary of Sharpness Properties

In descending order of technical excellence: the Sigma Art 50mm f1.4 has better coatings and resolves better at every aperture – more notably when inspected by sensors with a pixel pitch below 5um. The combination of higher contrast and increased resolution gives very high perceived sharpness. The Canon L 50mm f1.2 doesn’t match the Sigma Art until at least f8, below which aberrations affect resolution, especially in Zone C: corners don’t reach acceptable sharpness until f5.6. Contrast, plasticity and pellucidity are both a notch behind the Sigma Art. The Ding’s coatings and optical design create much lower contrast. Spherical aberration rises prominently at close range – creating a distinct impression of softness between f1.1 and f2.8. In terms of resolution, the Ding reaches parity with the Canon L from f4-f11 – perhaps a little earlier if compared at a 2-5m distance.

Of the three, only the Sigma 50mm f1.4 Art reaches the standard of excellence expected of a fast prime in 2024. The DG HSM version of the Sigma Art is outstanding value at present prices.