I’ve had a hankering for a long time to get faster, better glass for my DSLR. The all-encompassing Nikon 18-200mm f/3.5-5.6, while convenient, was not pushing my happy buttons. Building a lens with a 10x zoom range necessitates some compromises — namely, aperture size and image quality.
After a second vacation with the 18-200mm, where I wound up happy with a few photos but mostly “blah” (or downright angry) about many in terms of IQ, I decided it was time to find something better. I’m not crazy about having to swap lenses in the field, but I’m even less crazy about soft images. I’ll be the first to say it’s really the photographer, not the gear, that makes the photos, but even when I thought I did everything else to the best of my ability — appropriate settings, composition, making best use of light, etc — I often expected better out of the gear than what I got.
My other criteria was a lens that was better for indoor use: I wanted the better low-light performance that you can only get from a wide aperture. So I started looking around at fast-aperture standard zooms that had a focal range somewhere around 17-50mm (this is approximately the DX equivalent to the “standard” zooms on FX, 24-70mm). I knew I didn’t want to pay Nikon pro-glass prices ($1400 for their 17-55mm f/2.8), so I went on the hunt in the third-party market. The offerings and quality there have improved greatly in recent years. Here’s what I honed in on during my first pass:
- Tamron AF 17-50mm F/2.8 SP XR Di II VC (say that five times fast) – $650
- Sigma 17-50mm f/2.8 EX DC OS HSM FLD (almost as bad) – $600
- Sigma 17-70 2.8-4 DC Macro OS HSM “Contemporary” (there’s 3 versions of this lens; this is the newest) – $500
- Tokina 16-50mm f/2.8 AT-X 165 PRO DX – $950
All of these come in between $500 and $650, except for the Tokina, which I quickly discarded because it was by far the most expensive (almost $1000), didn’t appear to be in production anymore, and some of the reviews I found weren’t so complimentary. I was tempted by the Sigma 17-70mm, especially the new Contemporary version which several reviewers spoke highly of. It has good range and the build quality is supposed to be superb. But it’s not constant aperture (note the f/2.8-f/4), and you pretty much lose the f/2.8 as soon as you zoom in past 20mm or so. By the time you get to the long end at 70mm, where you probably really want the big aperture more often than you would at the wide end, you’re limited to f/4. So, that was also a no-go.
Which leaves the Tamron and Sigma 17-50mm f/2.8 lenses. Both have the same range, the same constant f/2.8 aperture, and both have active image stabilization (Tamron calls it “VC” for Vibration Control; Sigma is “OS” for Optical Stabilization). They are as closely matched, feature-wise, as possible. Some people debate the benefit of active stabilization in wide or standard zooms. However, it is a huge advantage when handholding in low light conditions, and I knew I wanted it.
Reviews of both lenses were mostly positive with some caveats; there were a handful of comparisons from people who have used both, with most of those giving a slight edge to the Sigma — although there are plenty of fans of the Tamron. One thing I found mentioned repeatedly was that there can be significant sample variation with third-party lenses; sometimes you need to send them in to get calibrated right out of the box to address backfocusing, decentering, or other issues. Given the enormous savings over the similar Nikon lens, most hobbyists feel it’s worth the risk.
Since I couldn’t decide purely from reading user reviews, I decided to order both. I would spend some quality time with each, gather my impressions, and then return the one I didn’t want (hooray for Amazon prime for free 2 day shipping and easy returns). Here’s my thoughts and some comparison images. Or just jump straight to the conclusion.
Build and feel
Both lenses exhibit mostly solid build quality, well above that of your typical $100 18-55mm kit lens, but not up to the same tank-like level as Nikon or Canon pro lenses (from what I can tell; my time with such lenses is limited), or even my Tokina 11-16mm f/2.8 ultrawide, which is very impressively built.. They have appropriate heft and tight seams. The rubberized zoom and focus rings are pleasing to the touch, but the action on the Tamron is a little stiff for my liking. Not a deal breaker by any means. The finish on the Sigma seems nicer. The lens hoods both lock in securely. One annoyance with the Sigma is that its zoom ring rotates opposite that of Nikon lenses (so it matches the Canon system), although I can’t really deduct points for that — Canon users would have the same beef about Tamron.
I like the overall feel and look of the Sigma slightly more, but the lens makes a disturbing rattle when you shake it — from what I’ve read this is normal (and might be the OS system), and it certainly doesn’t seem to be hurting anything, it just doesn’t sound good. The rattling seems to go away once the lens is in use on a camera. If it weren’t for this rattling, I’d give this category to Sigma, but as it stands there isn’t a clear winner here.
Build and feel winner: Tie
Here’s where these lenses have some significant differences. The Tamron’s autofocus system is a bit dated; it uses an older style motor, and as a result is a bit slower and a lot louder than other more modern lenses. Tamron’s 70-300mm VC, which I also recently purchased, has the newer system, which Tamron calls USD (Ultra Sonic Drive). More importantly for this comparison, the Sigma 17-50 employs that company’s version, HSM (Hyper Sonic Motor). Whatever you call it, the net result is that the Sigma focuses more quickly and quietly than the Tamron. The autofocus noise of the Tamron is distinctly audible in a quiet environment, and may be inappropriate depending on what you shoot. To me, it makes the lens sound a bit cheap.
As for accuracy, the Tamron was more accurate out of the box when it locked on correctly. There were a few shots where it acted like it focused (and allowed the shutter to fire), but upon later examination it had actually back-focused. It didn’t do this all the time though; my focus tests showed it to be spot-on. I did some focus fine tuning with the Sigma and set -10 on my D7000. Once I had the Sigma dialed in, its accuracy seems better, with one caveat: if you place the focus point on the edge of something (in my case, the edge of my wife’s pregnant belly in profile), such that the left half of the focus box contains your subject and the right half contains the background, it tends to focus on the background instead of your subject. This appears to only happen when the background is to the right, like there’s a very slight bias in that direction. I don’t know if this is just my copy that does this, but it’s possible. In practice, it hasn’t often been a problem.
In low light situations, the Sigma has a slight edge — it will lock on faster with less hunting than the Tamron.
Autofocus winner: Sigma
Both systems worked reasonably well. The Tamron’s was much more noticeable — it was audible, first of all, although not what I would call “loud” like the autofocus system. More importantly, it was really obvious when looking through the viewfinder that the VC was working. The image really “froze” once you half-depressed the shutter, and I could tell how much it was helping me out.
The Sigma’s OS system was nearly silent, but not quite as effective. It definitely helped me in low light, but not quite as much as the Tamron. If I had to quantify it, I’d say it was about 75% as effective as the Tamron.
Image stabilization winner: Tamron
Here’s the big one. We’re going to break it down into several categories: sharpness, bokeh, vingetting, contrast/color. I made some test shots out on my back patio using a tripod and a remote trigger, so the positioning of each shot was exactly the same and there was no chance of my hand on the camera introducing shake. Optical stabilization was off for these tests.
The holy grail of many photographers is finding a lens that will be as sharp as possible. As you probably know, the weakest point for most lenses in terms of sharpness is when their apertures are wide open — in the case of these two, that’s f/2.8.
In general, I found the Sigma to be sharper at all apertures, but especially wide open at f/2.8 and at 50mm. The two lenses weren’t all that far apart in the center — in fact, in some cases there was no clear winner when looking just at the center at f/2.8 — but the Tamron become quite a bit less sharp as you moved away from center. I should note that sometimes, under certain conditions (a bright, well-lit foreground subject against a dark background at f/2.8) there was a little bit of “glow” and edge softness introduced at the transition point between subject and background. However, I was not able to always recreate this. I didn’t notice it initially until I imported my first batch of photos for editing, and when I went to take the next set of images I couldn’t get it to happen. Since it was not repeatable, I’m not giving it a lot of weight. I will continue to investigate, but it wasn’t terribly noticeable. Zoomed all the way out to 17mm, the Sigma still wins at sharpness but it’s not as stark as at 50mm.
Stopped down to f/4, the Sigma still has the advantage, although it’s much closer. Above that, there isn’t much difference between the two.
Sharpness winner: Sigma
The quality of bokeh is generally judged by the smoothness of the background blur. It can be subjective, but I find most people’s interpretations of bokeh quality are pretty close. Better bokeh helps isolate your subject from the background, and makes for a better-looking photo.
To judge bokeh on these two lenses, I looked at the background of my various f/2.8 test shots. While not an enormous difference, I found the bokeh of the Tamron to be almost universally more pleasing. Particularly obvious when there were busy backgrounds (containing lots of trees, leaves, etc like in my test shots), the Tamron produced smoother transitions whereas the Sigma at the same settings would have a slightly “noisier” background.
Neither of these lenses are specifically designed for fantastic bokeh, and I’d say they’re both around the middle of the pack.
Bokeh winner: Tamron
As with sharpness, the weak spot for vignetting is when the aperture is wide open. While they were similar at 50mm, the Tamron exhibited fairly strong vignetting when zoomed out to 17mm at f/2.8, which improved quite a bit by f/4 (but was still noticeable). Conversely, the Sigma had impressively little vignetting at f/2.8 at all focal lengths, and virtually none by f/4.
Vignetting winner: Sigma
Contrast and Color
There’s probably a way to judge contrast and color rendition objectively, but I’m just going to go by observation. I had heard Tamron lenses were known for having good contrast and producing “punchy” images. I was looking forward to this, as I tend to play with contrast quite a bit in post (possibly because the Nikon 18-200mm had poor contrast and tended to produce flat-looking images in all but the most perfect conditions).
But it wasn’t quite what I expected. I found that the Tamron generally did produce very contrasty images, but not always in what I would consider the most pleasing way. The best way I can describe it is that while they were contrasty, sometimes they lacked color richness and depth. Often the colors seemed a touch oversaturated, which contributed to a bright but washed-out feel. In contrast (heh), the Sigma usually produced what I would call more “balanced” images with greater color depth. The effect was subtle, but noticeable.
When shooting something without bright colors in the frame, I slightly preferred the contrast and rendition of the Tamron, but not enough to overcome the issues in other conditions.
Contrast and color rendition winner: Sigma
Both of these lenses have pros and cons. The degree to which one is better than the other is important, as is how meaningful that area of performance is to your style of shooting.
Sharpness is probably the single most important quality to me here, and the Sigma takes that. Given the other categories it also won (autofocus, contrast/color, vignetting, and very close on the build quality), I’m going to be keeping the Sigma and returning the Tamron. If I could take the bokeh and image stabilization from the Tamron and put them into the Sigma (and reverse the zoom ring direction!), it would nearly be the perfect lens for what I need.
That said, the Tamron is not a bad lens, and it’s entirely possible that my copy was particularly soft at f/2.8 and another copy would be better. It’s also worth noting that if you don’t want/care about the Vibration Control, the non-VC version of this lens has reportedly better image quality and is a bit less expensive. You’d still have that somewhat loud autofocus, though.