Sigma 50mm F1.4 EX DG HSM - Leopold's View

Table of Contents

1. General
2. Look and Feel
3. Image Quality
   1. The Camera and Its Settings
   2. Colours
   3. Light Fall-Off
   4. Center Resolution and Contrast
   5. Border Resolution and Contrast
   6. Bokeh
   7. Auto Focus
   8. Flare Resistance and Loss of Contrast
4. Conclusions

General

Lately I've been migrating more and more towards large aperture photography, one step on that path being getting Canon's excellent EF135mm f/2L USM and shooting it almost exclusively wide open. The inadequacy of Canon's EF50mm f/1.4 USM performance wide open has buggered me for a long time, and my sad experiences with Canon's EF50mm f/1.2L USM auto focus system left me believing that there would be no 50 mm lens for my needs in the near future. Enter Sigma and its new 50mm F1.4 EX DG HSM lens, a complete new design optimized specifically for shooting at large apertures. Features of the lens include HSM (Hyper Sonic Motor, similar to Canon's USM), full-time manual focus, a non-moving front element, a weight of half a kilo, a non-rotating front element and a 77 mm filter thread. A lens hood and a soft pouch are included in the price that was 424.90 € locally here in Tampere, Finland. (For comparison: Canon's EF50mm f/1.4 USM with a lens hood but no soft pouch costs 374.80 € at the same shop.) The following are my first (and second) impressions of the combination Canon EOS 5D + Sigma 50/1.4 (picture above), and how it compares to Canon's 50 mm f/1.4 lens (picture below). This article doesn't even attempt to be a one-in-all full test of Sigma's new lens. It's just a collection of my own experiences. They may be useful, they may be not. For rigorous tests with lots of numbers, have a look at PhotoZone or some other proper lens test site.

Look and Feel

The lens feels good in the hand. Manual focus is not quite L smooth, but perhaps comparable to Canon's f/1.4. Otherwise, the construction of the Sigma is clearly superior to Canon's.

It should come as no surprise that the 5D's viewfinder is really bright and manual focusing with the Ee-S Super High Precision Matte focusing screen is no problem. By the way, I recommend this or a similar focus screen to all large aperture shooters.

People seem to have mixed opinions of Sigma's Auto Focus. To me it has felt slightly slow at times, but nevertheless accurate. Perhaps I'm one of the lucky ones, or perhaps my 5D has repaired itself after the sordid affair last year when I tested five different copies of Canon's EF50mm f/1.2L USM and none of them could focus at any aperture. Auto focus will be discussed in its own chapter further down.

Image Quality

The Camera and Its Settings

Center point auto focus was used all the time, and each photo was taken only once. While using auto focus doesn't necessarily tell how well a lens could perform at optimal conditions, it tells me which kind of pictures I can expect 90% of the time. I only use manual focus when a situation is too tricky for auto focus.

Colours

Light Fall-Off (Sigma vs Canon)

	Sigma f/1.4	Canon f/1.4
f/1.4
f/2.0
f/2.8
f/4.0

Here we can see that both lenses vignette a lot when used wide open, though the Sigma does it perhaps slightly less. However, at f/2, Sigma's border's have brightened much more than Canon's, which still brightens a lot between f/2 and f/2.8. This must be due to the large front element. Excellent!

Another thing that becomes apparent is that the Sigma has a wider field of view, making it appear slightly shorter than 50 mm. This isn't necessarily significant because lens focal lengths are only defined at infinite focus distance, and this bookshelf was not even near infinity.

Center Resolution and Contrast

	Sigma f/1.4	Canon f/1.4
f/1.4
f/2.0
f/2.8
f/4.0
f/5.6
f/8.0
f/11.0
f/16.0
f/22.0	No such aperture on the Sigma

Well, do I need to say anything? Look at f/1.4, there is just no contest. Canon's 50 mm lens has always had a reputation of being hazy wide open, and this sure shows why. Canon shows signs of longitudinal chromatic aberration (the letters are purple), so we might get a better performance with careful manual focusing, but if I would have been happy to use manual focus all the time, I might just as well have bought Canon's notorious unable-to-focus-if-its-life-depended-on-it-1.2L.

For some reason Sigma's quality decreases between f/1.4 and f/2. Perhaps auto focus has done an error, maybe it's focus shift. However, the Sigma is better at full aperture than Canon is at f/2.8, and that's good enough for me.

Both lenses start to suffer after f/11 from diffraction effects, which is hardly surprising.

Border Resolution and Contrast

	Sigma f/1.4	Canon f/1.4
f/1.4
f/2.0
f/2.8
f/4.0
f/5.6
f/8.0
f/11.0
f/16.0
f/22.0	No such aperture on the Sigma

Well well well, the tables have turned now, haven't they? Now it is Canon that has the clear advantage. At its best at f/8-f/11 Sigma reaches the level that Canon has at f/4.0, but the Sigma cannot touch the levels of detail that Canon has between f/5.6 and f/11.

Bokeh

Now, let's have a look at how Sigma's bokeh looks like.

Pictures: A demonstration photograph taken wide open at f/1.4. Focus is at the lantern at a distance of about 1.5 meters.

	Sigma f/1.4	Canon f/1.4
Full frame (downscaled)
Partial crop #1 (50% size)
Partial crop #2 (50% size)
Partial crop #3 (25% size)

There is a distinctive difference in bokeh rendering between the Canon and the Sigma. Whereas Canon's f/1.4 is busy with clearly rendered circles that sometimes seem more like rings, the Sigma has turned the background into indistinctive mush.

When the subject is further away from the camera so that the background doesn't get as much out of focus as in this example, the difference between the lenses decreases. Although I have tested this, I didn't feel it necessary to include photos of non-difference.

All this talk about bokeh, but does it really show in everyday large aperture photography? In a word: yes. It was immediately noticeable that whenever I opened up the aperture, backgrounds just melted away with the Sigma. The word often used to describe the bokeh of Canon's EF85mm f/1.2L USM, buttery, comes to mind. An example, shot at f/1.4, is presented below:

Of the lenses I've owned, I'd place this lens' bokeh right along Canon's EF135mm f/2L USM and EF100mm f/2.8L USM Macro, and above such lenses as EF24mm f/1.4L USM, EF24-105mm f/4L IS USM, EF70-300mm f/4.5-5.6 DO IS USM, EF50mm f/1.4 USM and the otherwise nice and sharp EF50mm f/1.8, the Nifty Fifty.

Auto Focus

To me my combination seems to work pretty well most of the time: about 9 pictures in 10 are tack sharp, but the remaining one photo out of ten is misfocused. This is pretty much the same experiences I get with Canon's 50/1.4, except I'd say the error ratio is perhaps higher on the Canon.

But since I was at it, I did some tests at several different distances (55 cm, 1 m, 2 m, 3 m, 8 m) at 2/3 stop steps between f/1.4-4.5 (i.e. f/1.4, f/1.8, f/2.2, f/2.8, f/3.5, f/4.5, using 5D's center focus point exclusively). Here are my finding presented graphically:

Distance: 55 cm, 50% crop
Sigma f/1.4	Sigma f/1.8
Sigma f/2.2	Sigma f/2.8
Sigma f/3.5	Sigma f/4.5

As can be seen, and contrary to my earlier bookshelf test, the flower shot at f/1.4 feels a bit mushy when looked at. However, auto-focus in itself has worked: details in front or behind the center are not sharper so there is no noticeable front or back focus. This can be confirmed from full size originals.

Distance: 1 m, 50% crop
Sigma f/1.4	Sigma f/1.8
Sigma f/2.2	Sigma f/2.8
Sigma f/3.5	Sigma f/4.5

At 1 meter, f/1.4 could perhaps be a tad front focused, but I'd put that down to individual than systematic error.

Distance: 2 m, 100% crop
Sigma f/1.4	Sigma f/1.8
Sigma f/2.2	Sigma f/2.8
Sigma f/3.5	Sigma f/4.5

Again, f/1.4 doesn't look so good, but otherwise the series is very nice.

Distance: 3 m, 100% crop
Sigma f/1.4	Sigma f/1.8
Sigma f/2.2	Sigma f/2.8
Sigma f/3.5	Sigma f/4.5

At 3 meters the series looks consistent.

Distance: 8 m, 100% crop
Sigma f/1.4	Sigma f/1.8
Sigma f/2.2	Sigma f/2.8
Sigma f/3.5	Sigma f/4.5

No problems at 8 meters, either. It feels that the largest apertures work better at larger distances, with or without auto focus.

Finally, below are two pictures where I tried whether it made any difference if I focused on a stone 7 meters away or some plants at 9 meters.

50% crop
Sigma f/1.4, AF on stone at 7 meters	Sigma f/1.4, AF on plants at 9 meters

Sigma clearly can differentiate between 7 and 9 meters.

All in all, I'm pretty happy with Sigma's auto focus system. It is not as foolproof as the best of Canon, like the EF135 f/2L USM, but I don't feel it loses to Canon's 50/1.4. With a proper focusing screen you can immediately tell whether the lens actually has reached correct focus or not, so asking the camera to refocus isn't a big problem most of the time. Of course, if you are shooting fast-moving subjects the ability to refocus after a botched once-in-a-lifetime shot probably won't make you feel any better.

Flare Resistance and Loss of Contrast

Flare, 3 lenses at f/4
Only lens hood
Sigma 50/1.4	Canon 50/1.4	Canon 24-105/4

The Sigma is clearly giving the most contrasty performance. Both Canon lenses are badly hampered by flare: their images are dull and have a distinctive lack of contrast.

Flare, 3 lenses at f/4
Only lens hood (same as previous)		Lens hood + hand
Sigma 50/1.4	Canon 50/1.4	Canon 24-105/4

When the 24-105/4L's 24 mm lens hood is helped by putting a hand just outside the 50 mm picture, contrast is increased greatly and it is now roughly on the same contrast level as Sigma's 50 mm lens, though a bit darker.

Flare, 3 lenses at f/4
Hand protects lens
Sigma 50/1.4	Canon 50/1.4	Canon 24-105/4

When a hand is used to protect all lenses from any direct light, all lenses perform better. Now even Canon's 50/1.4 manages to give a good, contrasty performance. Contrast in all three images is quite similar, but for some reason the 24-105/4L is a tad darker than the others.

All in all, I still think that the flare resistance of Canon's 50/1.4 is quite poor, particularly considering how little glass there is inside the lens. The 24-105/4L's puts in a surprisingly solid performance for a zoom. The winner is, however, the Sigma.

Conclusions

Nevertheless, a landscape shooter might not want this lens as Canon clearly has the more consistent edge-to-edge performance when stopped down between f/5.6-f/11. Then again, at f/8 the EF24-105mm f/4L IS USM is so good that using a prime for that purpose might not be such a hot idea after all, particularly considering Canon f/1.4's weak flare resistance.

In short, here are my pluses and minuses for the Sigma as compared to Canon's 1.4. In both lists the points most important to me are listed first:

+	Smoother bokeh
+	No moving front element
+	Better flare resistance
+	Better center sharpness and contrast when using large apertures
+	Less light fall-off (vignetting), particularly at f/2
+	Better build quality
-	Worse corner sharpness
-	Slightly more expensive
-	Slightly slower auto focus
-	Heavier