sebboh wrote:
what i find interesting is that the rokkor seems to beat pretty much all the lenses at f/1.4 (and f/1.2) at the borders, but then falls behind in the corners.

I noticed that too.

I'll reiterate someone's sentiments from before, you need to have two 50s.

For me, it's my most common focal length.

Ideally, I'd have the 50MP(or, honestly, a Zuiko 50/2) and the Rokkor 58 converted. I like native EF mount stuff.

As it stands now, I've got the Sigma 50. I'll always stand by it. I don't know that even if I could afford the 50L, I'd own it. The Sigma is 98% the L.

The 50/1.0L did much better in the center sharpness than I expected--indeed, it is among the sharpest in the entire group, for the image center. Its straight aperture blades, unfortunately, made it one of the worse performers in terms of bokeh when stopped down. The mirror box clipping of the blur discs is very severe wide open, no surprise there. Corner performance was lacking, but this too was not a surprise. Personally, I think the extreme flare of the 50/1.0L looks really neat as a special effect, but I wouldn't want to deal with it all the time.

The 50/1.2L was surprisingly good as well, but overall, the Rokkor 58/1.2 was just a bit better. Just look at the corner performance...what an amazing little lens from Minolta. No wonder it's so highly valued. Bokeh on the 58/1.2 is lovely--of course, the extra 8mm is a bit of an advantage, but the evenness is enviable. The 50/1.2L does have curved aperture blades, which the Minolta lacks. But who are we kidding--we'd shoot either of these lenses wide open 90% of the time.

The Zeiss 50/1.4 is disappointing. Sharpness is dismal wide open, and doesn't improve very much upon stopping down. Even the "lowly" EF 50/1.4 handily beats it for contrast and detail. Compared to some of the other offerings here, it doesn't offer a good value. Flare control is superb, though, owing to the T* coatings.

The Zeiss 50/2 MP does well in sharpness and is the clear winner in flare control, but this is hardly surprising for a f/2 macro. It's also the best performer in controlling chromatic aberration. Other than that, it doesn't really stand out among the other faster lenses. I think its design is a good demonstration that there's much more to a lens than being sharp.

For the price, the EF 50/1.4 does remarkably well. It's not going win awards, but it's a great bargain lens, I think still better than the Sigma overall. Shame that its build quality is so poor.

As for the 50/1.8, clearly, it's outclassed by many of the other lenses in this grouping.

I must add that I'm quite impressed by the comprehensiveness of this test, and the range of lenses being compared. If I had to choose the best value, I'd go with the Rokkor. It's just that good. But the 50/1.0L was the surprise here--I wasn't expecting it to do as well as it did. It would make a remarkable portrait lens, though its corner softness even when stopped down makes it less well suited for landscape.

plasticmotif wrote:
sebboh wrote:
what i find interesting is that the rokkor seems to beat pretty much all the lenses at f/1.4 (and f/1.2) at the borders, but then falls behind in the corners.

I noticed that too.

I'll reiterate someone's sentiments from before, you need to have two 50s.

For me, it's my most common focal length.

Ideally, I'd have the 50MP(or, honestly, a Zuiko 50/2) and the Rokkor 58 converted. I like native EF mount stuff.

As it stands now, I've got the Sigma 50. I'll always stand by it. I don't know that even if I could afford the 50L, I'd own it. The Sigma is 98% the L.

only two? that's crazy

i think this is probably the straw that breaks the camels back and finally pushes me into getting a rokkor 58/1.2. have to see whether the rokkor 50/1.4 stays after that.

Very awesome lens comparison!!

I love my rokkor 58/1.2 but it's half-impossible to hit focus without using live view. Is that just me? Need to invest in a focus confirmation chip...

Hi U.C.,

Fantastic job here, thanks so much for taking the time to do this and for sharing it here!

Just a few questions on your focusing methods...were multiple AF attempts made and, say, a best-of-three choice made (for AF lenses, obviously). Or perhaps everything was done consistently across-the-board, manually focusing using, say, live view focus? Some clarification or insight would be helpful in assessing some of the results.

And a few suggestions...I think incorporating a purple-fringing test in your battery—shooting, say, silhouetted tree branches against a bright sky—would be a relevant addition. And an equalized-FOV resolution test, using a standard test chart, should suffice for those of us looking to compare equally-framed test images.

Other than that, I really like the single-row presentation of the results, it beats the more commonly seen scrolling-up-and-down-a-page format for me!

You guys have done a bang-up job...keep up the good work!


Jimmy G

quicksilver33 wrote:
Very awesome lens comparison!!

I love my rokkor 58/1.2 but it's half-impossible to hit focus without using live view. Is that just me? Need to invest in a focus confirmation chip...

i wouldn't count on the focus confirm chip being much better, they tend not to be super accurate on f/1.2 lenses either. do you have a precision focus screen?

sebboh wrote:
plasticmotif wrote:
sebboh wrote:
what i find interesting is that the rokkor seems to beat pretty much all the lenses at f/1.4 (and f/1.2) at the borders, but then falls behind in the corners.

I noticed that too.

I'll reiterate someone's sentiments from before, you need to have two 50s.

For me, it's my most common focal length.

Ideally, I'd have the 50MP(or, honestly, a Zuiko 50/2) and the Rokkor 58 converted. I like native EF mount stuff.

As it stands now, I've got the Sigma 50. I'll always stand by it. I don't know that even if I could afford the 50L, I'd own it. The Sigma is 98% the L.

only two? that's crazy

i think this is probably the straw that breaks the camels back and finally pushes me into getting a rokkor 58/1.2. have to see whether the rokkor 50/1.4 stays after that.


Oh, if money were no object...I'd have like thirty different 50s.

Helios 44/2, Zuiko 50 1.4, Zuiko 2 Macro, Zuiko 50 1.2,Zuiko 55 1.2, Sigma 50, Rokkor 58, Noctilux, Summilux R, C/y 1.4, C/y 1.7, Zeiss 50 1.4, ZE 50 MP, Ai-S 1.2, Ai-1.4, Ai-S 1.4, Ai-S 55 1.2, 55/2 Ai-S Macro, 50/3.5 Ai-S, Canon FL 55, Canon 50L, SMC 50/1.4 & 1.7 ...

It's late can't think of any others.

oh yeah, 50/2.8 Tessar, Konica 50/2, Canon 50/2.5, R- Summicron,

Ok, I'm done for real now.

If you're wondering why I want to have all these lenses, it's because I've never collected anything. And besides cycling, photography is my only hobby. I don't have the room for bikes....so, lenses make sense.

Thx for this impressiv test. I have also a big collections of 50th. My favorits are the Zuiko Auto-Macro 2/50 for best overall performance and sharpness. The Summilux 1.4/50 for Biotar like bokeh. The Rokkor 1.2/58 for the fattest bokeh and the Rokkor 1.2/50 for it's Noctilux like bokeh.

As I said before - great test. I see however that some people seem to over-interpreting the results a bit.

First regarding sharpness - it's a test at a fixed and relatively close distance. For instance the Canon 50/1.4 is rather good at closer distances while awful near/at infinity (extreme field curvature). A corner crop at f/4 of a shot at infinity would reveal that for all practical purposes it had no corners at that distance. The Zeiss 50/1.4 is the exact opposite - it's designed as a portrait lens with high spherical aberrations at portrait distances (i.e close up) while at infinity stopped down a bit it would probably be the sharpest 50 in the test. There are exceptions of course - the Zeiss 50 MP is really good at all distances and apertures. It is however an exception. Most 50's have a distance that they are optimized for.

Equally the color test should not be overinterpreted. The test image has bright primary colors and the difference will perhaps tell you if you need to do some color adjustments in PP. Any weaknesses there are however very easily fixable. The big thing that you can't fix in post and where there is great variation between the lenses is the color separation i.e how good a lens is at separating the finer nuances in a color.

So my general advice is to take the tests very literally and don't generalize too much from them.

To answer the comments/questions about focussing: we've used live-view MF in all the tests. From experience only, I know that the 50/1.0L, 50/1.2L, Sigma 50 and Rokkor 58 shift much.

^^^ What denoir says.

An infinity (or at least long focus distance) test is the only thing I'm missing. Conversely, the brick wall test isn't really useful for my main shooting purposes of a 50mm.

Another conclusion that can be easily made from your test is that the Sigma really is a ~46mm lens (which was already well known).

Anyway, thanks for the effort, and I agree the presentation is very good.

What doesn't make sense to me is why you would want a portrait lens to be soft wide open in the center of the image at close focusing distances, yet sharp in the corners stopped down at infinity. That doesn't add up. If you called it a landscape lens or street lens, that's an assessment I could agree with.

After all, the 50/1.2L has undercorrected spherical aberration, and even it looks better at f/1.4 in the center than the Zeiss--a result I was NOT expecting. You may claim that the bokeh is the reason why you might accept softness wide open, but it isn't borne out by the results shown in the test.

Furthermore, the Zeiss wide open is just washed out--there's very low contrast at close focus, which again is undesirable, even for portrait work.

Bottom line is that I don't want any lens to be soft in the center of the plane of sharpest focus at any focusing distance. If what denoir says is true about the Zeiss at infinity, I'd use it for landscape work, not portrait work.

Um, soft wide open in the center is the hallmark of a classic portrait lens. The idea is to produce kind and forgiving portraits with a smooth skin rendering. The perhaps most famous classic portrait lens (designed in the 1930's) is the Zeiss 50/1.5 Sonnar and it's really completely uncorrected with even heavier SA than the Planar. If you have ever read any of the Zeiss white papers on optics, it's the Sonnar they show as what they consider to be the perfect portrait lens and it's softer in the center than the Planar and suffers from a rather weird field curvature.

It's true that the Zeiss 50 Planar is superb for landscape work. While the 50 MP has stronger contrast in the finer detail (40 lp/mm) the Planar shows higher contrast at lower spatial frequencies (10-20 lp/mm) which gives a really special look. However, Zeiss considers that to be a bonus - their design intention with both the 50 & 85 Planars is for them to be used as portrait lenses. And if used correctly they are really fantastic for that purpose. I'd pick the 50 Planar over the 50L and the 85 Planar over the 85L any day for portraits.

Here's a portrait with the Sonnar - which is the softest fast 50 I've ever used - softer than the Planar:

quicksilver33 wrote:
Very awesome lens comparison!!

I love my rokkor 58/1.2 but it's half-impossible to hit focus without using live view. Is that just me? Need to invest in a focus confirmation chip...


Great test, especially for those who have a soft spot for this particular focal length. I have recently gotten more interested in this focal length and fast MF lenses in particular. I have just started trying out a Rokkor and the Canon 55f1.2 both modified for EOS mount.

Initially I had a lot of issues getting focus right. I tried the EG-S screen on my 5D2, then liveview. I have just started using the Zacuto Z-finder with liveview and I'm finally getting good results. Has anyone else tried this method? Note: the responsiveness, especially the blackout time, seems like an eternity sometimes.

Portrait doesn't necessarily mean close up portrait or headshot. Some, including myself, find the DoF of the 50 1.0 too thin for close portraits anyway.

The beauty of ultra fast normal FL lenses IMO, is the ability to separate background in environmental portraits, but still include it with a normal field of view (not compressed like with tele lenses), at distances where skin flaws will not be resolved anyway. Sharpness wide open can only help here. I don't have any good examples myself, as I have had the 50 1.0 only for a short time and not really got to use it yet, but here is an example of the style i talk about, shot with the canon 50 1.2:

http://www.fredmiranda.com/forum/topic/909751/0#8578532

True, but the 50/1.4 Planar that we were discussing is only very soft wide open at close distances. At medium distances like the example you showed it's sharpness is comparable to what you get from the 50L. It has a more detail and a more SA "glow" so the look is different but the detail is there. The far field bokeh of the Planar is nothing to write home about but the close and medium distance blur is really nice.

denoir wrote:
The far field bokeh of the Planar is nothing to write home about but the close and medium distance blur is really nice.

From the samples from Makten I remember that it was the opposite. Maybe I don't really understand what you're saying here. Waiting for your samples to show what you mean.

denoir wrote:
True, but the 50/1.4 Planar that we were discussing is only very soft wide open at close distances.

The brick wall test shown here is medium distance, an image ratio of at least 50:1, that is comparable to environmental portraits, and the 50/1.4 Planar is still soft.

Edit:
And if you look at Lloyd Chamber's church mosaic at even greater distance ("comparison" chapter), it is still that soft and makes the new Distagon 35 1.4 look like a sharp beast wide open.

You should look at that review a bit more carefully

"Field shots at f/1.4 show the usual “halo” effect seen with other f/1.4 designs. It’s not that the resolving power isn’t high—it is very high (see below)—it’s that the halo effect diminishes the apparent sharpness by lowering the contrast."

I thought that you would think that I read things carefully. We are talking about aberrations reducing micro contrast, not resolving power, that was always a given to me.