21mm M lenses compared and reviewed
Zeiss Biogon T* 2.8/21 ZM
Announced in 2004 and released in late 2005, the Zeiss 21mm rapidly gained a excellent reputation for performance and mechanical build. It may not be as great as the current Leica 21mm ASPH, but for half the price it's no slouch.
[The 21mm ZM] delivers at least the same overall performance as the [non ASPH] Leica Elmarit-M 21mm, but without the employment of aspherical surfaces and at a lower price. The sturdiness of the mount may be not as good as that of the Leica, but the mechanical accuracy is not affected and I did not detect any decentring.
You will also find another subscription-only review at <ReidReviews.com>.
User opinions & samples
Leica 21mm versions
In April 2002 Andy Piper published detailed comparisons between the Leitz/Schneider 21mm Super-Angulon-M ƒ3.4 (1964) and the (preASPH) 21mm Elmarit-M (1980).
Pre ASPH vs. ASPH
In May 2002 Jack Flesher had the following to say about the 21mm ASPH, Elmarit-M and 24mm ASPH super-wides:
(BTW for what follows, assume "21P" is the 21mm Elmarit-M (ie the pre-ASPH model), while the "21A" is the new 21mm ASPH (and not the 21mm Super-Angulon, usually shortened to 21SA):
Since the question of "Which wideangle for my M?" had come up several times […] and since there did not seem to be any clear-cut answers, I decided to compare these three lenses for myself.
I have read and re-read Erwin's reviews of the 21's and the 24, but had difficulty integrating his results. In my mind, his review makes any 21 prior the to Asph version seem like it is not worth having - they "didn't bear the Leica badge with full honor". He boasts on the performance of the 24A, but makes no comparison between it and the 21A. The problem for me was that my 21 Pre-asph seemed like a pretty decent performer. Not as sharp as my 24, but certainly not horrible, and notably better than any of my Nikon 20mm counterparts. In actual use however, I found myself continually choosing the 24 over the 21 to go into my bag because of its superior performance, yet I found myself more than once wishing I had brought the 21 for that little bit extra spread. So, curiosity got the better of me. Could the 21 Asph be as good as the 24? It was difficult to imagine it could, as the 24 is such a stellar performer, but I wanted to know the answer for myself.
Here is what I found out for those of you interested - But please note that I offer this with all of the usual disclaimers about my tests not being the end-all for everybody else, my giving subjective comments about objective results, non- perfect scientific testing criteria, etc:
First, a small note on ergonomics: The 21P has a quicker focus throw of about 90* from .7M to infinity, and has a slightly larger aperture ring than the 21A. The 21A matches the 24 in a longer focus throw of about 105*, and both the 21A and the 24 have similar aperture rings that are smaller than the 21P.
ƒ2.8 -- Here I got a surprise. When comparing the centers for the 21A and 21P, I found the 21P to be sharper(!) Not by much, but definitely a visible difference. I double-checked my notes to insure that I had not inadvertently swapped the slides. Indeed, I hadn't. At the corner I received another surprise. The 21P is notably better than the 21A (!!). The 24 is the overall winner at ƒ2.8, being sharper in the center, yet it essentially only equaling the 21P at the corner. It is worth noting that the 21P is performing significantly better at the corner than at its center. Perhaps an issue with under-corrected spherical aberrations or even a curvature-plus-focus issue? Yes, I was beginning to regret my 21A purchase as it looked like the 21P wasn't all that bad as I had originally thought.
ƒ4 -- An interesting phenomenon here. The 21A takes a quantum leap in image quality at the center, now almost equaling the 24, which also takes a fairly significant jump in quality. Both are clearly better than the 21P in the center whose performance here is actually a bit lower than it was at ƒ2.8. At the corner, the 21A improves a lot, but the 21P also improves a little, and in fact just edges out the 21A here; and even very slightly edges out the 24(!) So, the 21P loses the center, but hangs on to the corner by a tiny margin.
ƒ5.6 -- Here the lenses all come into their own and behave much as we might expect. The 21A is very sharp in the center, but not quite up to the 24. The 21P lags behind notably. In the corners, the 21A is now notably better than the 21P, and for all intents and purposes the 24 and the 21A are equal here. The 21P is still performing better in the corner than the center, but not by much, and is showing essentially even results across the image.
ƒ8 -- The results are essentially the same as above, with the 21A losing a little bit more ground to the 24 in the center and the corners.
ƒ11 and ƒ16 -- The 21A and 24 fall off here, but interestingly the 21P actually improves to the point where it is almost as good as the other lenses. So, for all intents and purposes in general picture taking situations, one can consider all three lenses to be equal performers at these stops.
For the field test, I shot the same image with each lens, keeping the exposure and focus-point identical for each lens. All shots were hand- held. For some of the shots I also altered my shooting position with the 24 relative to the 21 in an effort keep the main subject of uniform size for comparisons in the final image.
Field test results and "IMO" Conclusions:
Do I get (or in my case keep) the 21A or 21P? Based on the test-targets, if I was looking to have the sharpest image possible with a 21, and cost was no object, the 21A would be the answer. However in actual shooting situations, these differences became very slight, and almost unnoticeable in many images. In the sweet-spot stops - ƒ4, ƒ5.6, and ƒ8 - the 21A is clearly superior to the 21P on the test-targets. In the field these differences are mitigated to a significant degree, and while the differences are still noticeable, I'm having a tough time convincing myself they are significant enough to justify the extra cost of the 21A.
Furthermore, if you generally use your non-aspheric 21 at ƒ11 or ƒ16 for maximum DOF in the typical sweeping WA shot, there is probably no reason to upgrade to the 21A, as I doubt you will notice any differences at all. In a few of the field tests images I did detect some crispness in the 21A and 24 images that wasn't present in the 21P images, but to be truthful it was a slight difference and probably wouldn't bother me in most cases.
So, I would not recommend one upgrade to the 21A unless they actually see softness in their images that bothers them. While the E55 filter size on the 21A presents another plus for me, as I don't have to carry any E60 filters for the 21P and have other E55 lenses, it is certainly not a compelling enough reason alone to justify the additional cost of the 21A. As for size/weight, all three lenses are essentially the same profile so the differences in my mind are not significant either. While the 21A is a superior performer to the 21P, I have to be truthful and regretfully admit that I feel I've wasted my money on its purchase for the limited gain I am detecting. :-(
Between the 24 and 21A, I'd have to say we are comparing an ƒ2.8 lens in the 24 to an ƒ4 lens in the 21A, which could become an issue while shooting in low light or building interiors - however hand-held shots at slow speeds probably have bigger limitations than the difference in optical performance here. In the field tests, most of the shots between these two lenses appeared almost identical in resolution, with the occasional edge given to the 24. Based on the test-target results, if you need the sharpest super-wide period, the 24 is the superior performer. However, from ƒ4.0 on these two lenses are pretty darn close - close enough that most of the differences will not be noticed when using these lenses hand-held, so perspective will probably become the dominant criteria for lens choice.
Some final comments
Flare appears very well controlled in all three lenses, and none was detected. Color-cast is Leica-neutral in all three lenses, and contrast appears identical in all three lenses. For those of you interested, I will state that IMO (totally subjective) the bokeh on the 21A and the 24 is smoother than that of the 21P.
In answer to the original question, knowing what I now know, I will probably sell both 21's and keep my 24… Anybody want a 21A or 21P? :-)
(A.N. note: The new 21mm ASPH has of course a max aperture of ƒ2.8 - identical to that of the 24mm ASPH. What I presume Mr Flesher is talking about is that when compared to the 24mm, the 21A only begins to peform well at ƒ4.)
In June 2002, "Jay" had this to add on the 21mm Super-Angulon-M ƒ3.4:
I used the 21/3.4 S/A-M (first a chrome and then a late black one) extensively. It suffers from noticeable vignetting or light falloff, which doesn't completely disappear until ƒ8. In my experience it's about a 2.5-stop difference between center and corners at ƒ3.4 and it should have come with a center filter like the Hologon. Unless your subject is in the very center of the image and the surroundings are unimportant or much brighter, you'll need to add about 1 stop. I used my S/A as a landscape lens and never shot it wider than ƒ8.
The 21mm ƒ3.4 was also discussed online in detail in March 2002 at <Photo.net: #002jD5>.
Finally, can you hack the 21mm SA so you can use it on a M6 with the built-in meter? Yes - basically you grind a deep notch into the rear metal lens barrel, after dissasembling the glass elements(!) Take a deep breath and work through the detailed instructions at:
The Voigtländer 21mm alternative
Wondering how the Voigtländer Color Skopar 21mm ƒ4 stacks up against the Leica crew? Jack Flesher again:
Well, as it just so happens... I had an opportunity to try out one of these little gems. Simply stated, in terms of resolution it performed at least as well as the 21P if not a bit better, and as such is not really very far behind the 21A. Its big faults were more significant falloff in the corners (maybe 2 1/2 stops?), more visible distortion in the corners, and the fact it is more flare-prone than either Leica 21.
BUT, then again, we are talking sub $US 400 for the lens INCLUDING a finder -- a finder that puts Leica's BL finder to shame! (FWIW, the VC 21 finder appears to be their 15 finder with framelines.) So, for somebody who only needs a 21mm occasionally, or someone who wants a really tiny compact ultra-wide lens for their M, I think this one is the buy of the century!
21/4 review and discussion links
General review with sample photos by Jim Tardio
The new "over-sized" Voigtländer metallic rectangular hood on Stephen Gandy's cameraquest site
- <LUG - v27/msg06704.html>
In May 2004, Kyle Cassidy posted a note about meeting Lee Friedlander at a graduation ceremony and how much he liked his Voigtländer 21/4(!)
- <Photo.net: #008IVl>
Online discussion - May 2004
- <Greenspun.com: #0098Kd>
Online discussion - May 2002
- <Photo.net: #001S6D>
Online discussion - Jun 2001
Kobalux 21mm ƒ2.8
Although now discontinued, the Kobalux 21mm ƒ2.8 by the Y.K. Optical Company, Yokohama Japan, often gets a thumbs-up from a ultra-wide shooters.
So if you absolutely need a ƒ2.8 lens and cannot bring yourself to cough up the multiple thousands of dollars required for the Leica version(s), then it may be worth your time tracking one of these babies down. BTW, the lens is only known as the "Kobalux" in the USA - in Japan it is known as the "Avenon". Same lens, different brand name.