Review: Sigma 50mm f/1.4 DG HSM Art

dilbert said:
Viggo said:
The Zeiss 50 f2.0 was the best 50 for Canon by a larger margin, before the Otus, which technically isn't a 50, but oh well. The Zeiss 50 f1.4 however is absolute rubbish. So if the Sigma is worse than that it would be demolished by the Otus in the same way the sf-s 55-250 gets destroyed by the 70-200 mk2.

I can't see how the Sigma can be only slightly better than the 50 L and still compete with the Otus, that just doesn't add up. I'm beginning to wonder if the 50 Art lenses tested are all preproduction with high copy variation when the results differentiate so much.. But in some cases the example shots are just done wrong.

I downloaded the Zip with the raws from the 50 Art and I wasn't impressed at all, and with the first images shot by the Otus they blew my mind, so either Sigma lied, plain and simple, or people can't shoot or it's very high degree of copy variation and or preproduction lenses. This also of course mean I have to buy one, and buy it new so I can exchange it 4-5 times until a good copy is found.

If you look at this review:
http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=941&Camera=453&Sample=0&FLI=0&API=2&LensComp=403&CameraComp=453&SampleComp=0&FLIComp=0&APIComp=4

then it is pretty clear what the difference between the 50A and 50L is. Now if you don't care about corner sharpness (such as when using a shallow DoF and the corner is not in focus anyway) then this won't matter much to you. But if you're using a 50mm prime for landscape (for example) then the Sigma just blows the Canon away.

Higher quality lenses also require a higher quality review(er). Just putting it on your camera and going outside and shooting some objects and posting centre crops doesn't cut it.

I am aware of TDP test charts, and I can't understand how someone could can shoot so bad as to get the SLRlounge results and be that close to the 50 L when Bryan CLEARY shows the opposite.

I absolutely do care about best as possible corner performance, that's why I sold three copies if the 50 L, center comp isn't my favorite always. So I do agree with you.

I also think that center performance should also be MUCH better with the Sigma than the 50 L, not just the corners. The Otus is epic and the Sigma doesn't even come close as far as I have seen. BUT that being said, none of the tests I've seen is the end all, so I won't judge it until I have tried it myself and seen what others here on CR have to show when they get it. But an Otus Killer it doesn't look like it...
 
Upvote 0
Judging by what you guys have posted here, the only lens to definitively come out substantially better than the Canon 50/1.2 is the Otus.

For the rest of us who own any copy of a 50mm, would it not make sense to perhaps wait and see what Canon is going to offer?

I'm a little skeptical about 3rd party lenses. Not for ability reasons but for performance. Seeing a buddy wage war after his Sigma lens stopped working on his Pentax after a firmware upgrade has left me ultra cautious.
 
Upvote 0
I don't get it. Bryan's test at TDP is a complete blowout; not even close. The Sigma is sharper in center and much much much sharper than the 1.2L away from center and in corners, and the CA on the 50L is bad, while nearly non-existent on the Sigma.
In comparing any two other lenses, where there is no brand loyalty or investment-justification involved, that kind of test result would simply be a clear blowout, and there would be no further discussion. Not here though. Here we see the defensive comments and a retreat to the trenches of the intangibles like bokeh (which is not clearly different in any sample shot I have seen anyone point to specifically) and creaminess, and the supposed uselessness of test charts (but only for this lens).
 
Upvote 0
... And raise your hand if you would be noticing the same supposed intangible advantages of creaminess, bokeh, etc. in the Canon shots and test chart if the results had been accidentally switched? If the Canon results had been swapped do the Sigma results. If the two results had been switched, who right now honestly would still be pointing at the blurry purple-fringed chart and claiming that it's meaningless if that had been identified as the result for Sigma instead of the one for Canon? I think what we would be seeing is a lot of people pointing at that purple blurry chart and saying look how worthless the sigma lens is compared to the incredibly sharp and clear canon one, if the charts and shots were swapped.
 
Upvote 0
ScottyP said:
I don't get it. Bryan's test at TDP is a complete blowout; not even close. The Sigma is sharper in center and much much much sharper than the 1.2L away from center and in corners, and the CA on the 50L is bad, while nearly non-existent on the Sigma.
In comparing any two other lenses, where there is no brand loyalty or investment-justification involved, that kind of test result would simply be a clear blowout, and there would be no further discussion. Not here though. Here we see the defensive comments and a retreat to the trenches of the intangibles like bokeh (which is not clearly different in any sample shot I have seen anyone point to specifically) and creaminess, and the supposed uselessness of test charts (but only for this lens).

There is a LOT of red-ring disease in this forum. Understandable I guess?
 
Upvote 0
ScottyP said:
I don't get it. Bryan's test at TDP is a complete blowout; not even close. The Sigma is sharper in center and much much much sharper than the 1.2L away from center and in corners, and the CA on the 50L is bad, while nearly non-existent on the Sigma.
In comparing any two other lenses, where there is no brand loyalty or investment-justification involved, that kind of test result would simply be a clear blowout, and there would be no further discussion. Not here though. Here we see the defensive comments and a retreat to the trenches of the intangibles like bokeh (which is not clearly different in any sample shot I have seen anyone point to specifically) and creaminess, and the supposed uselessness of test charts (but only for this lens).

What don't you get? I'm sure people that have the 50L love the fact that some are saying their lens sucks and that the Sigma is the bee's knees. I think it's more the attitude of some posters that are getting people to be defensive. The 50L doesn't get any worse now that the Sigma is coming out, and for many that have the 50L, the decision is not as clear cut, especially if they're waiting for the 50L II to come out before making a decision to switch the S50A.

I've tried Canon's 50 f/1.4 and have the 50L and my experience is that the 50L is better than the 50 f/1.4 wide open to about f/2.8, which is consistent with the results of LensRentals 50mm shootout while TDP shows the 50 f/1.4 to be a better performer. But the 50L AFs a lot better and can be used wide open. The 50 f/1.4's AF was not accurate from f/1.4-f/2.

With Sigma's success with the S35, I'd expect the S50 to perform well AF-wise. Over time, I'd suspect that many would trade the 50L for the S50A, but it is a bit premature when it still is not available to everyone. And if one has the 24-70 II, then the 50L would be used for portraiture/low light only, so edge/corner performance is not as important. The S50 is even better than the 24-70 II at 50mm. Is everyone now going to say that the 24-70 II sucks too for landscape and that the S50 A should be used instead?
 
Upvote 0
http://www.amazon.com/gp/aw/d/B000RXW0AI/ref=mp_s_a_1_1?qid=1397654437&sr=8-1&pi=SY200_QL40

Dylan, mac... I was thinking abit a b&w nd and using a step down filter. 77 will fit the 70-200 and the 24-105... but sure about the 100l, but I know it is smaller so I can use a step down ring there as well.

Having said that... I've heard horror stories about the screw grooves being wrong and people having to cut the ring off their filter and lens... so is there a better brand of step down?
 
Upvote 0
JD, I've bought step-down rings from many different vendors and brands, but they all appear to be the same, and I've never had any issues with them. Keep in mind that the hood won't work if you use, them, though. I ended up buying a 72mm ND filter for my 50L / 85L II. One shoot in the sun with the "1/8000s" flashing in the viewfinder until you hit f/2.8 or f/4 was enough for me :)


Random Orbits said:
ScottyP said:
I don't get it. Bryan's test at TDP is a complete blowout; not even close. The Sigma is sharper in center and much much much sharper than the 1.2L away from center and in corners, and the CA on the 50L is bad, while nearly non-existent on the Sigma.
In comparing any two other lenses, where there is no brand loyalty or investment-justification involved, that kind of test result would simply be a clear blowout, and there would be no further discussion. Not here though. Here we see the defensive comments and a retreat to the trenches of the intangibles like bokeh (which is not clearly different in any sample shot I have seen anyone point to specifically) and creaminess, and the supposed uselessness of test charts (but only for this lens).

What don't you get? I'm sure people that have the 50L love the fact that some are saying their lens sucks and that the Sigma is the bee's knees. I think it's more the attitude of some posters that are getting people to be defensive. The 50L doesn't get any worse now that the Sigma is coming out, and for many that have the 50L, the decision is not as clear cut, especially if they're waiting for the 50L II to come out before making a decision to switch the S50A.

I've tried Canon's 50 f/1.4 and have the 50L and my experience is that the 50L is better than the 50 f/1.4 wide open to about f/2.8, which is consistent with the results of LensRentals 50mm shootout while TDP shows the 50 f/1.4 to be a better performer. But the 50L AFs a lot better and can be used wide open. The 50 f/1.4's AF was not accurate from f/1.4-f/2.

With Sigma's success with the S35, I'd expect the S50 to perform well AF-wise. Over time, I'd suspect that many would trade the 50L for the S50A, but it is a bit premature when it still is not available to everyone. And if one has the 24-70 II, then the 50L would be used for portraiture/low light only, so edge/corner performance is not as important. The S50 is even better than the 24-70 II at 50mm. Is everyone now going to say that the 24-70 II sucks too for landscape and that the S50 A should be used instead?
Well said, Random, and I'm sure it's just a matter of time before we see the comparisons between the S50A at f/11 and the 24-70 II at f/11. The S50A appears to be sharper. I fall into the portraiture category and just as my 24-70 Mk I photos don't suddenly suck because I have the 24-70 II, the same goes here.

Back to ScottyP, I think what many of us are trying to say is that sharpness and a flat field were clearly not Canon's highest priority in designing the 50L. If you read the Press Release, you'll see that it was designed for, "[W]edding and portrait photographers, as well as professional photojournalists." For these subjects, it is an excellent lens and the bokeh, color, and contrast are what makes it great. Look at these photos on 500px if you think I'm crazy.

Yes, the Sigma is better at test charts, and yes, it looks like it will ultimately prove to be the better lens in terms of IQ. No, it's not the only lens that does poorly on test charts but works better in actual use. The 24-70 Mk I also had horrible field curvature, but produced excellent photos. The 85 f/1.2 II is also softer on the test charts (outer frame) than it is in use. The 180L macro is another interesting one - some sites list at as the sharpest lens they've tested, while many show it to be only above average. My copy is just short of my 300 f/2.8 IS II.

Also, I think what many of us are saying is that the other factors (bokeh, color, contrast, flare resistance) ARE more important for portraits than sharpness. That's not to say that the Sigma won't beat the Canon on those factors as well. Most of us who are saying this have a bag full of sharp lenses that we use for everything else, but sometimes we want a unique look that can't be achieved by the 24-70 II or other lenses.

For me, one of the reasons I like the 50L is because it's small and relatively light compared to my other lenses. It's much less conspicuous to walk around with it on my 5DIII than my 24-70 II or 85 II.

Finally, I think Canon 50L (soft overpriced crap lens?) and Sigma fans (poor QA, crap AF?) alike are going to be defensive on this one, but it's no surprise given what Roger over at LensRentals says about the 50L:

"Well, we could talk about this for hours: this is one of the most controversial, irritating, and spectacular lenses in the lineup. I won’t pretend to know what you’ll think of it, but our customers are evenly divided with “I love it” and “I hate it.” Here are a couple of pointers:

  • If you’re not used to working with very narrow depth of field (f/1.2 close up is NARROW) this lens will take some practice. Use one autofocus point only or it will focus where it pleases, not where you want.
  • The plane of focus is not flat; it’s slightly curved. Focus-recomposing will guarantee you an out of focus shot. Just don’t do it.
  • The 50 f/1.2 L exhibits focus-shift for near distance shots, meaning that shooting objects a few feet away between f/2 and f/4 the lens will probably not autofocus accurately. At f/1.2 it’s accurate, and by f/5.6 the depth of field is wide enough that you won’t see the effect.
  • The 50 f/1.2 L is camera specific: a copy that is wonderful on one camera may backfocus on another. It’s best used, for that reason, on cameras with focus adjustment like the 1D series or the 5D mkIII. If you don’t use autofocus adjustment, your images are likely going to be back or front focused.

The bottom line: when this lens is right, the shots are spectacular and the background blur is awesome, just like the 85 f/1.2. But it’s more finicky and more difficult to get those shots with this lens.
 
Upvote 0
dilbert said:
I'd love to see someone demonstrate a photographic shoot where the 50L produces images that the S50A can't touch.
I don't think that will happen or is possible :). The Sigma is an excellent lens and appears to be equal or better than the 50L in all areas other than size and weight. The only exception might be bokeh, but from what I've seen so far, the bokeh differences are slight outer ringing and very minor minor onion rings that are only going to be visible at 100% magnification and won't show up in normal-sized prints. The 50L bokeh may be a hair better in smoothness and artifacts because it (theoretically at least) has a larger physical diaphragm. Unless it has some fatal flaw with point light sources or something, I don't think the 50L will beat it by a big margin at anything.

I think what we have seen and will continue to see is the Sigma smoking the 50L on sharpness outside of the center of the frame.

Also, the one thing that most impresses me most about the Sigma is the lack of distortion. We all take that for granted with PS, LR, and DxO, but lenses with distortion that low are a rare breed.
 
Upvote 0
ScottyP said:
I don't get it. Bryan's test at TDP is a complete blowout; not even close. The Sigma is sharper in center and much much much sharper than the 1.2L away from center and in corners, and the CA on the 50L is bad, while nearly non-existent on the Sigma.
In comparing any two other lenses, where there is no brand loyalty or investment-justification involved, that kind of test result would simply be a clear blowout, and there would be no further discussion. Not here though. Here we see the defensive comments and a retreat to the trenches of the intangibles like bokeh (which is not clearly different in any sample shot I have seen anyone point to specifically) and creaminess, and the supposed uselessness of test charts (but only for this lens).

ScottyP said:
... And raise your hand if you would be noticing the same supposed intangible advantages of creaminess, bokeh, etc. in the Canon shots and test chart if the results had been accidentally switched? If the Canon results had been swapped do the Sigma results. If the two results had been switched, who right now honestly would still be pointing at the blurry purple-fringed chart and claiming that it's meaningless if that had been identified as the result for Sigma instead of the one for Canon? I think what we would be seeing is a lot of people pointing at that purple blurry chart and saying look how worthless the sigma lens is compared to the incredibly sharp and clear canon one, if the charts and shots were swapped.

Thanks. This is really not a subtle difference. The 50mm f/1.4 Art is among the best lenses you can buy for any camera. It has performance at f/1.4 that equals the 135mm f/2.0L. At f/2.0 it equals the Canon 200mm f/2.0 IS which is regarded as the best autofocusing lens money can buy. While the 50L is the softest lens Canon currently makes in terms of average resolution.

Here's a comparison that I found very revealing, taken from one of the better SLR lounge raws, with my own sharpening and contrast preset applied to both in lightroom, and color corrected to match. Open both in separate tabs and flip back and forth and see if you can tell the difference:

http://www5.picturepush.com/photo/a/14119258/img/Picture-Box/SLR-IMG-3113.jpg

http://www1.picturepush.com/photo/a/14119259/img/Picture-Box/SLR-IMG-3119.jpg
 
Upvote 0
dilbert said:
Higher quality lenses also require a higher quality review(er). Just putting it on your camera and going outside and shooting some objects and posting centre crops doesn't cut it.

That's true of all lenses, of course. But I don't think test chart comparisons are terribly helpful either. What I would rather see - and which Brian will surely do when his review shows up; it's one of the most useful aspects of his reviews - is comparisons of photos of actual subjects of various sorts: portraits, buildings, landscapes, etc.
I suspect the Sigma will prove to be a superb lens, and wouldn't be at all surprised if in may ways it beats the 50L - though if I used a 50mm lens to take portraits I would be far more interested in which has better bokeh than which shows the most clinical detail.

(Personally, I have no stake in this particular debate - if I want crazy sharpness from a 50mm-ish lens, I suspect that my Sony/Zeiss 55mm 1.8/Sony A7r combination is at least as sharp as the Sigma would be on any current Canon body, while weighing less than the Sigma lens alone, and if I want a blur-fest from a 50mm-ish lens I'll use my manual Canon 55mm 1.2 on the Sony.)
 
Upvote 0
The EF 50 f/1.2L and EF 85 f/1.2 excel at portraiture, where bokeh is king, and excessive sharpness just gets smoothed away in post production. People don't want portraits of their pores, they want portraits of their faces.
 
Upvote 0
ScottyP said:
Here we see the defensive comments and a retreat to the trenches of the intangibles like bokeh (which is not clearly different in any sample shot I have seen anyone point to specifically) and creaminess [...].

Your overall point re pro-Canon bias may have more than a hint of truth about it, but I disagree with your implication that "intangibles like bokeh" don't matter (at least, I assume that's what you mean by "retreat to the trenches"). Of course, bokeh/blur is a tricky issue for reviewers because it can't be measured in the same way that resolution/sharpness can, and while there's presumably no debate whether one image is sharper than another, there may well be disagreement over which has better bokeh (and, of course, whether it matters in the first place). But the fact that it can't be measured doesn't mean it's "intangible" - you see blur just as you see sharpness - and whether one image is "clearly different" from another is also subjective (in the 1.8 vs 1.4 comparison someone posted earlier to show that there was almost no difference, the difference seemed quite obvious to me and I had no difficulty at all in forming a preference). And for some people a lens's bokeh/blur properties may trump other considerations.

That said, for all I know the Sigma may have better bokeh than the Canon (unlikely though that may seem), just as the Sigma 35 1.4 seems to have better bokeh than the Canon 35 1.4 (at least as per Brian's comparison at the digital picture.).
 
Upvote 0
jdramirez said:
http://www.amazon.com/gp/aw/d/B000RXW0AI/ref=mp_s_a_1_1?qid=1397654437&sr=8-1&pi=SY200_QL40

Dylan, mac... I was thinking abit a b&w nd and using a step down filter. 77 will fit the 70-200 and the 24-105... but sure about the 100l, but I know it is smaller so I can use a step down ring there as well.

Having said that... I've heard horror stories about the screw grooves being wrong and people having to cut the ring off their filter and lens... so is there a better brand of step down?

I have never own any step down ring. So no comment. I bought the BW 72mm 3stop ND filter, since my 50L, 85L, and 135L are 72mm

This is the one I have: http://www.bhphotovideo.com/c/product/752859-REG/B_W_1066146_72mm_103_Solid_Neutral.html
 
Upvote 0
One key issue with the 50L is the "focus shift" that occurs if the lens is used near minimum focus distance at apertures between 2 and 4. Since this is quite a common setting in portrait shots, I wonder why so many people recommend the 50L so much as a portrait lens.

And the second thing which wonders me even more, is that none of these reviews checked whether the 50A suffers from this same focus shift issue. Since the issue appears to come from spherical aberrations which are less present in the 50A there is a good chance that the 50A is less affected, but the many tests and comparisons between 50L and 50A might as well take a closer look.
 
Upvote 0
200px-White_flag_waving.svg.png


Truce?

Also, Dylan, that's the same filter I have.
 
Upvote 0
Rudeofus said:
One key issue with the 50L is the "focus shift" that occurs if the lens is used near minimum focus distance at apertures between 2 and 4. Since this is quite a common setting in portrait shots, I wonder why so many people recommend the 50L so much as a portrait lens.

And the second thing which wonders me even more, is that none of these reviews checked whether the 50A suffers from this same focus shift issue. Since the issue appears to come from spherical aberrations which are less present in the 50A there is a good chance that the 50A is less affected, but the many tests and comparisons between 50L and 50A might as well take a closer look.

The 50L's focus shift occurs at min focus distance (very close). So for head and shoulder portraits, which is usually shot over 1m, then it's a minor issue and very slight. The focus shift only occurs when stopping down, as the aperture gets to f2.8 it's at it's worse. Most people are using this lens either wide open (not an issue) or stopped down to f8/f11 in the studio under flash light, so it's also not an issue.
If one uses this lens for it's intended purpose, its fine. Sure it's Canon's softest L lens (they still make other non L lenses which are softer), but it's not a lens designed to be critically sharp. It's about charector and bokeh...but unfortunately so many amateur photographers get totally hung up on sharpness and lens charts.

If you are best served with the Sigma, knock your self out and get one. If you fancy a 50L, likewise. But don't think that either lens will bring peace and happiness....it's just a lens at the end of the day.
 
Upvote 0
Why does some people think that "portrait shooting" automatically means CENTER focus and HEADSHOT?

The 50 L sucks for a 2/3 portrait or a full body shot of a kid when the edge points are used to focus. The 85 is WAY different and to mention those two lenses in the same sentence for the same things is just not right.

I have always felt the 50 L is one of the coolest lenses Canon has made, I LOVE the size and how it handles, and unlike the 135 L, 85 L and 35 L it's fully weather sealed which is a big deal for me. And for center composed shots I can't think of a lens I would rather use for portraits, BUT it's just so extremely limited.

So I both love and hate it. If the Sigma does the nice bokeh, like it seems, is sharper or as sharp in the center and much sharper in the corners and it's resistant to flare and have similar contrast and color rendition plus the already proved (almost) zero distortion, it would be the better choice no matter if you own the L or not, imo.
 
Upvote 0
Viggo said:
Why does some people think that "portrait shooting" automatically means CENTER focus and HEADSHOT?
I'm not sure, but that's what the 85, 135 and 70-200 are for :) I use the 85 for everything from full length to head shots and my favorite "portrait" lens for broader shots is actually the 24 1.4 II, but the 50mm focal length has its place, too.
 
Upvote 0
mackguyver said:
Viggo said:
Why does some people think that "portrait shooting" automatically means CENTER focus and HEADSHOT?
I'm not sure, but that's what the 85, 135 and 70-200 are for :) I use the 85 for everything from full length to head shots and my favorite "portrait" lens for broader shots is actually the 24 1.4 II, but the 50mm focal length has its place, too.

Do you center focus with full length?
 
Upvote 0