Thursday, April 18, 2013

Sigma 18-35mm F1.8 DC HSM Impressions (Updated)

Sigma just announced the world's first APS-C 18-35mm f/1.8 zoom.  Stop and consider what that achievement means: there aren't even any f/2.0 constant zooms for APS-C. The 18-35 f/1.8 is basically wish fulfillment for all of the DX fanboys wishing for something more interesting than full-frame handme downs.

Quick Rundown

Construction. Lots of low dispersion glass elements, with three aspherical elements. You'd need at least this much to achieve usable results for a constant aperture zoom that's this bright. We're talking molded aspheric lens elements here, which is what you would find in the mid-range  with all of the third party lens manufacturers. The professional grade lenses use ground aspheric elements, which are better, but significantly more expensive to produce. Aperture is a nine-bladed rounded diaphragm, which is what you would hope for for a lens of this quality. This thing comes in at 810g, so it's not lightweight. To put it into perspective, the Nikon D7100 is just around 750g give or take a few. Also, it's nice to see Sigma step further and further away from the drag grey of yesteryear. This lens looks quite the business in the product shots. (Dare I say that it looks a bit Sony-ish?)

Distortion. Sigma's own graphs show barrel distortion at 18mm that's in keeping with what you would see with most 17-50 f/2.8 lenses; actually it seems a little less. The down side is that there is pin cushion distortion at 35mm. You wouldn't normally expect to see this at 35mm; it's usually inthe middle of the zoom range for many lenses or you would be dealing with a prime. This is a bit disappointing considering that 35mm is the "normal" focal length of DX (rightly or wrongly... mostly wrong), but not unexpected given the compromises that they would have had to make to achieve a constant aperture zoom this fast.

Vignetting. Sigma's published data indicates about 60-70% drop off in illumination from the center at f/1.8, which is about 1/2 to 2/3 EV. It'll be noticable, but not out of line with many fast APS-C lenses, and it would certainly be correctable in post. 

Sharpness. Don't expect miracles at very wide and very open. Sigma's own MTF charts show contast and definition taking a dive when you go into the corners at 18mm. Things look more consistent across the frame at 35mm. Note: Sigma's office MTF charts don't indicate what aperture they are representing, but it does seem like the 18mm chart is showing the lens wide open. At least I hope that's wide open...


Don't get too excited just yet. Get intrigued, maybe, but I'm advising against excited. Cautiously enthusiastic if you will. The difference between Sigma and Nikon or Canon is that Sigma has the specs at a lower price, but you would want to use the Nikon all the time at the headline spec. For most Sigmas and third party lenses, it's been a case of "Yes it's nice to have f/2.8", but the difference is that with Nikon, you won't have any hesitations about using f/2.8 all the time. So that is the challenge with f/1.8; I'm skeptical that it will be anything but "nice to have", but if this lens is as sharp by f/2.8 as f/2.8 zooms are at f/4, then that would be a useful accomplishment in itself.

However, the question is... what do you want this lens for? The focal length precludes it from doing portrait work, but it looks otherwise ideal as a general-purpose photojournalism lens. Do you want f/1.8 for more shutter speed? The obvious problem is that the plane of focus will be too narrow to consistently focus on fast moving objects. If you want it for the bokeh, that might be a better proposition, as Sigma has certainly raised some eyebrows with what they can do with their 85mm f/1.4 and 35 f/1.4 full frame primes. At these focal lengths and typical shooting distances, foreground-background isolation is a challenge. If you want ultra creamy backgrounds, you still have to go to full frame... and you still won't get there without an expensive lens.

However, my first hope is that the 18-35mm f/1.8 is another decent bokeh oriented lens in Sigma's portfolio. If it turns out that that it's a sharp lens, then that would be icing on the cake. But for all-around use, the 17-50 f/2.8 OS seems like it would be a more flexible option. It will also probably be less expensive than what the 18-35mm will come up as. One thing to bear in mind: no matter how fast a lens is, the results will still look like hell if you shoot video with it; you want stabilization for that. Also, for portraits and headshots, don't underestimate 50mm DX and f/2.8. You position a subject with a 17-50mm so that that the background isn't distracting, but a 18-35mm lens is too short to get you to a visage flattering focal length.

So in other words, what Sigma has created is not the uber version of the f/2.8 normal zoom, but rather a niche specialist lens for the the wide-normal photojournalist use. And was that what were asking for all along... for more interesting lens choices other than full frame lenses kludged onto our crop cameras?

Update: April 21, 2013

The first sample pictures with a Canon EOS 600D are out here. Whoever took this set of sample images did a good job and obviously shot with some love/appreciation for the city of Seoul. There aren't a lot of technical shots, but lots of glorious scenery showing you what you situations that you would actually use the lens with. My first impressions are that the results seem to match the ambition of this lens; yes, there's no arguing that the subject impression one gets is that images produced look nice. As expected, wide and open is weak from an outright resolving standpoint, but the bokeh quality is nice. The wide angle on this lens seems to be all about the bokeh... to that end, corner resolution doesn't matter so much. Hard to say for now, but it does seem to follow in the 17-50OS's footsteps... sometimes the background blur looks a bit nervous, but it's not too hard and fringey. Look at the branches in the shots of the cherry blossoms; it's not full frame creamy, but it it's not harsh in the way that the Nikon 35mm DX would be. Wait did I just say that... a zoom  is better than a prime at the same focal length and aperture? Nah, I couldn't have....

In other words, crop frame shooters finally have a reasonable way of producing wide angle with shallow depth of field that full frame shooters have taken for granted. We're not talking about the ultra foreground/background isolation of the Nikon 24mm f/1.4. but what we are talking about is the ability to do this at all on a crop frame camera. Wide is the enemy of shallow; it takes some doing to get both to live together. Barrel distortion looks fairly well controlled at 18mm, but you can make out the pin-cushioning at 35mm.

Overall, once you stop down it's as good as the other good lenses in this range, but I still feel the same way about the usefullness of this lens. A lot of people are saying that it's the lens that finally gives crop shooters the equivalency of a 24-70 f/2.8 on full frame, but I just don't feel the same. Without the 35-70mm range, it's not really an all-around lens, but it's still general purpose to cover many situations. I still see this as a specialist wide-angle with the added benefit that it goes to 35mm. That in itself is an accomplishment, and until somebody can produce a lens that does the same thing, the Sigma 18-45 f/1.8 will be the very best at what it does.

Update: April 22, 2013

 DPR has a hands on preview here. I'm likngi what I'm seeing from Sigma's marketing department. If this was the past you would have thought that this was vaporware, or soemthing that would have taken forever to come down the pipeline. They're doing it the right way; headline specs, and then sample units quickly out into the field for maximum exposure. Also asurprise: compared to similar lenses, the Sigma is quite long. Not surprising considering how many elements it would take to make constant f/1.8 work.

No comments:

Post a Comment