Canon EF 85/1.2L II USM Lens Review

Nice review (as usual), Phil. The point about renting the lens is absolutely true. I have been able to use this lens in the past only because of rental companies and I have to hand it to them..

A plug for an excellent rental place for those in Texas (Dallas & Houston) Lighttec is an awesome resource for aspiring amateurs who can afford $25 per weekday or for the entire weekend and get to play with an 85mm f1.2L lens!

Thanks for the reviews!

"The Canon EOS lens mount is a bit larger than competitive lens mounts, which is why Canon has been able to produce crazy fast lenses such as the 50/1.0 and the 85/1.2 . . . "

So how might this exlain the Nikkor 50/1.2 with it's smaller lens mount?

I'd give the review an "A" for art, and an "F" for science.

I agree totally with your comments on the usefulness and the effects possible with this lens. But the review falls down in several places.

"Mounting an f/1.2 prime lens instead of a typical f/2.8 professional zoom will send six times as much light to the viewfinder, yielding roughly an apparent doubling of brightness."

A stock "bright screen" scatters so little light that the change in viewfinder brightness (and visible DOF) going from f2.8 to f1.2 is barely noticeable, not an "apparent doubling of brightness". The light from the outer portion of the lens's exit pupil just falls outside your eye's pupil. Even a "coarse ground" screen like Canon's Ee-S "super precision" or a KayzEye (wihtout OptiBrite) screen screen doesn't scatter light over a sufficient angle to result in an "apparent doubling". And those coarse ground screens are dimmer, to start with.

"The Canon EOS lens mount is a bit larger than competitive lens mounts, which is why Canon has been able to produce crazy fast lenses such as the 50/1.0 and the 85/1.2."

Nikon produced three quite servicable f1.2 lenses, the 50mm f1.2 (reasonable priced and pupular for its bokeh), the 55mm f1.2 (weakest of the three) and the 58mm f1.2 aspherical NOCT (discontinued, but so in demand that used ones are going for 4x what new ones did 5 years ago). The only lens that the Canon mount made possible was the "publicity stunt" 50mm f1.0. Form what I gather, there was only one production run of that lens, and it took Canon over 10 years to clear them out of the warehouse and distribution channel so that they could announce it was "discontinued".

A minor note, if you don't mind a manual focus lens, there's now a Zeiss 85mm f1.4 Planar available in Nikon F and Pentax K mounts. And the "new" Sony 85mm f1.4 is also branded "Zeiss", but a slightly different rar focusing design to permit autofocus (although they still label it a "Planar").

And, as Mark U pointed out on the discussion forums, the point about the small-sensor bodies is incorrect.

"The lens will work fine on a small-sensor body, but you will be carrying around ridiculously more glass than you need."

The size and weight of a telephoto is determined by the size of the front elements, which is determined by the aperture and focal length. As you move deeper into the lens, element size is influenced by the sensor size, but the elements are getting so much smaller in the rear that the difference is negligable.

To the folks who asked about lens mount size... I'm not sure that you're thinking about the geometry. An f-number is a ratio. The fact that Canon can fit a 50/1.0 lens on the camera does not mean that it could make a 300/1.0. The f/1 aperture on a 50mm lens is much smaller than the f/1 aperture on a 300mm lens. The fact that Nikon can get a 50/1.2 to fit on the (rather small) F-mount does not mean that they could make an 85/1.2. f/1.2 at 85mm is much larger than f/1.2 at 50mm.

Philip, I didn't actually "ask" about the lens mount; I'm pretty clear on the issue.

You're discussing the geometry of focal length and aperture. What matters in this situation is the geometry of the exit pupil and the rear element. Draw a sensor as a 43mm line (that's its diagonal). Now draw the lens rear element as a 45mm line (I'm doing this from memory, but that should be the biggest rear element that will fit through the EOS mount) 39mm from the sensor. (Canon and Nikon share a 39mm backfocus, that's as close as the rear element of the lens can come to the sensor without hitting the SLR mirror). And finally, draw a 50mm f1.0 lens as a "simple" lens (not retrofocus), a 50mm line, 50mm from the sensor line.

Notice how you can draw lines from the edges of the lens to the center of the sensor and they will totally clear the rear element of the lens. Lines to other points on the sensor will have just a tiny bit of vignetting.

Now draw a simple telephoto as a 300mm line, 300mm from the sensor. You can draw the same lines from edges of lens to the sensor, and get the same result. Ray cones targeting the center of the sensor have a totally clear aperture, but those targeting the edges of the sensor are slightly occluded because they go outside the area bounded by the rear element.

Even if you go to more realistic lens designs, retrofocus wides and slightly retrofocus normals, or conventional Galilean telephotos, you're still dealing with moving the exit pupil and scaling it, so you're always dealing with the f number as a ratio of pupil size to distance, and you always construct the clearance to the rear element the same way.

Hope this clears it up for you.

Joseph: Your command of theory is impressive. We are left, however, with the fact that Nikon makes 50mm lenses as fast as f/1.2. Canon makes f/1.0. Nikon makes 85mm lenses as fast as f/1.4. Canon makes the 85/1.2. Canon seems to be able to get an extra 1/2 f-stop out of their max performance lenses. Perhaps they have accomplished this via some means other than their (indisputably) larger lens mount.

You have a very compelling theory about viewfinder brightness and say "going from f2.8 to f1.2 is barely noticeable". I set the 85/1.2L to f/2.8, push the depth of field preview button and have no trouble telling that the screen has become darker.

Philip:

Canon made a 50/1 lens. Canon made a 200mm f/1.8 lens. Leica make a 50mm f/1 lens.

Your assumption on the Nikon F mount being "small" for fast lenses is incorrect.

I do have/use a 9.8cm f/1 lens (made for an aerial camera) on Nikon bodies. Yes, it does focus to infinity and the front elements are huge while the rear element is quite small.

Sample shot taken with a Nikon D70 at f/1:

Edit: In response to the later post of Joseph- Front element (not filter thread size) is 78mm, rear element 18mm (yeah!), conveniently arrived in a T-mount. So, the rear element is well clear of the F mount registry. Sydney Ray's book (ISBN: 0 240 51540 4) discusses various possibilities (not this particular lens).

Canon had an 85/1.2 L in the FD series even though the lens mount was a bit narrower. And Canon had a 58/1.2 Super Canomatic R lens on the Canonflex in 1962 which was carried foward into the FL line a couple years later. The FD, FL and R lenses all had the same bayonet. The differences between them were the location and number of pins that projected from the back of each lens series. And remember that the Nikon F and Canon R-FL-FD mounts were designed about 30 years before the EF mount arrived.

I have had the EF 85/1.2L for over 10 years and it's truly an excellent portrait lens and has other interesting possibilities as well. I also have the Canonflex lens mentioned above and all the others up to 400/4.5. along with lots of other stuff to go with them.

Vivek, that's both a beautiful shot and a fascinating concept. Can you tell me more about the lens? What is the rear element diameter, and how far is the rear element from the image plane at infinity focus?

Philip, as Joe Richards pointed out, Canon also made the 85mm f1.2 in FD mount. On a thread about your review in the Canon forum, Fred C mentions that Zeiss also made a 50mm f1.2 in C/Y mount. Those mounts have the same restrictions on the size of the rear element and the allowable back focus as the Nikon F mount does. So, there are no issues with the F mount prohibiting an 85mm f1.2.

I don't consider f1.0 an issue, at all. As I pointed out, there was one production run of the Canon 50mm f1.0, and it took a decade to sell off that one run. If it were any more of a flop Canon would have had to melt them back down for their glass. If Nikon wanted to make an f1.0 lens, there are optical designs to reduce the rear element size (basically, a positive Galilaean system after the main lens). It's a pretty trivial expense, considering the cost of the front elements of the 50mm f1.0 (large, deeply curved, and every time you go up a stop, you double the precision requied in the lens grinding to maintain resolution). Topcon actually used this design in their 58mm f1.4 "UV Topcor" to reduce vignetting (it's a really weird lens, as I recall, the rear element is only about 22mm wide, when 35mm would have easily fit through their lens mount. I'll have to look at my specimine tonight).

But seeing the blazing success Canon had with their 50mm f1.0, and the fact that Nikon's own 50mm f1.2 sales lagged to the point where they discontinued the lens, I'm not expecting to see Nikon try for any f1.0 lenses. As I said, it was a publicity stunt, and these days, both Nikon and Canon are trying to make sure their publicity stunts are more "sellable" (Canon diffractive optics, Nikon's 200mm f2.0 and 14-24mm f2.8, for example).

As far as I'm concerned, taking a 1Ds and mounting both the Canon 85mm f1.2 and the Nikon 85mm f1.4 (with adapter), Nikon chose very wisely in making the 85mm f1.4 instead of a 1.2. Dismiss it as sample variation, if you like, but the only thing the Canon could do that the Nikon couldn't was open up to f1.2. The Nikon was the clear winner in terms of bokeh, sharpness, and contrast.

The thing that amazed me, until I read George Parris post, was how much faster the Nikon 85mm f1.4 on a Nikon D2X focuses than a Canon 85mm f1.2 on a 1D II. George mentions something I did not know, that the 85mm f1.2 is "unit focusing", the whole dang thing moves forward and back to focus, just like the older Canon FD version and the Zeiss C/Y version (and the older manual focus version of the Nikon 85mm f1.4). The AF versions of the Nikon, Sony, and Contax N 85mm f1.4 are all rear or internal focusing. Even with Nikon and Sony "screwdriver" AF, they move. Now, I don't know if making the Canon 85mm an f1.2 made the elements so thick that there isn't enough space between them for interal focusing, but it's a theory worth exploration...

These superfast lenses are usually inferior stopped down to their slower siblings. I am not sure if these lenses are made for rational reasons ;-) They're not a real alternative to the slower lenses for general purpose photography. When photographing wedding ceremonies in dim rooms, I think it is usually preferable to increase the sensitivity of the camera than use f/1.2. On the Nikon D3, an f/1.2 lens can be focused reliably accurately only by using Live View with zoom, in which case good results are possible provided that nobody moves. I cannot see any difference in the viewfinder image when stopping down from f/1.2 to f/2 but the difference between f/2 and f/2.8 is perceptible. Perhaps the Canon EOS 1 series focusing screens are different in this respect. When shooting at longer than close-up distances I think the maxium aperture is more useful and manageable, and autofocus can also be used effectively. I'm not against fast lenses by any means, but there is a sweet spot here and it's not f/1.2.

Ilkka: Notice that I said "Canon has been able to produce crazy fast lenses" not "crazy good lenses". My goal in showing the images at different apertures was to underline the fact that f/1.2 requires such high precision focus as to be impractical for most real-life situations.

I hadn't tried your f/1.2 versus f/2 viewfinder experiment until just now. On the 1Ds Mark III, my experience is similar to yours with the Nikon D3. The difference is negligible between wide open and f/2. f/2.5 is the widest aperture where the viewfinder becomes significantly darker.

"So how might this exlain the Nikkor 50/1.2 with it's smaller lens mount?"

A 50/1.2 is not a 50/1.0.

Ikka read some more reviews of this lens. For example, from Castleman: "At apertures wider than f/2.8, the f/1.8 (85mm) lens has slightly lower contrast and sharpness than the f/1.2L lens."

So, what you say may be true for most crazy fast lenses, but not this one. f/1.2 may not be the sweet spot, but for this lens it is pretty damn sweet. Until you have tried it you won't believe it. This lens is awesome at full aperture and stopped down.

The pictures in the article do a great job showing what this lens is for. The difference between f/1.2 and f/1.8 or even f/1.4 is obvious, and is the main reason to buy it. The contrast is incredible, even wide open, lending an apparent sharpness even though little is actually in focus. It's a special effect, and one that you don't see all that frequently because it's so expensive and specialized. I think that this lens is a great reason to buy into the Canon system.

Agree with Ken. There is a HUGE difference between an f/1.4 and an f/1.2 lens in terms of image rendition and the ease/difficulty of using it for picture taking.

also agree with ken regarding the special nature of this lens, but i have noticed significant chromatic aberration shooting from f1.2 to about f2.8 with the 85L. this is not so troubling with portrait use but it often is an issue for outdoor flower shots, for example. i enjoy my 85L but it's not without faults that seem inconsistent with its price.

Yes, this lens (and many others in the Canon system--even "L" lenses) have chromatic aberration in out-of-focus high-contrast transitions. It's easy to see: just shoot some newsprint at f/1.2 and look at a 100% magnification. Text in the plane of focus will be black-and-white, as expected, but text in front and behind the plane of focus will have a strong green and magenta cast to the blur. You can see it in the article's shots if you look for it in the hi-res downloads. At normal magnification, it's usually not a problem, but it can be quite noticeable and even objectionable for high-contrast, outdoor shots, as Michael says.

Working with a Digital Rebel:

Well Phil, the lens doesn't change optically, so the perspective remains the same, it's just that we get a crop from an image that would have been a full-frame image, so everything else remains the same, except that the crop shows a "smaller portion - hence crop" of the would-have-been-full-frame image.

Also using a full-frame lens on a crop sensor doesn't magically multiply the focal length of the lens, unless the lens is specifically designed for a crop-sensor (eg. Canon's EFs 10-22mm f/3.5-4.5 on a Rebel, which then does behave like a full frame 16-35mm on a full-frame.

I use the Canon EF 85mm f/1.8 USM on My D-Rebel and for the money I get alot of bang for my buck. Now if I had one of the full frame sensor Digitals then the 85/1.2 L glass lens would be the perferred lens.

"Alternatives If your wallet and shoulder can't support the price and weight of the 85/1.2L lens, consider the Canon EF 85mm f/1.8 USM, $315 (review). You lose one f-stop of speed, which means that you will need twice as much light, but the 85/1.8 is an excellent lens and balances better on the lighter digital bodies, such as the Rebel series."

Very technical, most of which I do not understand, and don't want to. I use this lens as standard on one body, the other is permanently fitted with the 100-400mm. I use all sorts of other lenses but I do like the 85mm. True it's heavy, slow to focus and a bundle of money - but the results are really rather good, to the extent that my local pro-lab have commented on it recently! I've tried the Tamron 90mm DI macro, which feels too light and small - especially on a 1VHS - and is very noisy when focusing but produces good images, but you just can't get away from the build quality and the general smooth action of this thing.

I have recent experience with the 5D and various focusing screens.

The stock screen is etched so that it is equivalent to f/2.5. Pattern in the screen actually "stops down" the image. Any wider makes 0 brightness difference. (and 0 focus difference, hence the need for the precision focus version.) With other screens, (I.E. the 'S' precision focusing screen, or also a true ground glass screen), wider apertures do make a difference. Look through the front of the lens with the stock screen and observe the circle that is close to where the aperture blades are at f/2.8.

1D and 1Ds stock screens may vary.

I understand the 85MM 1.2 is sometimes bought specifically for astrophotography.

I'd like to hear responses from anyone who has used it for such.

I've done astrophotography with the Sigma 20MM 1.8 and my Canon 50MM 1.8.

Wide open, both of those show flaring of stars (points of light turning to blobs of light) especially in the corners.

Hey!

It is a nice review. However I would add that the lens has some back-front focusing issues. I have a it on a Canon 5D and sometimes pictures are totally out of focus. My Canon 28-70 or 17-40 never had this kind of problem...Any suggestion? I have a few examles at www.usafromabove.com Furthermore, it is a heavy lens!!! :):):)

I would like to correct my previous post. Actually, the focusing issue was my fault. It took me several weeks until I finally learned how to use this lens. Disregard my previous comments. I was an idiot! It is a killing lens!!!! You can see the pictures it produces on my site. http://www.haringphotography.com Look for the blond girl and the guy on the beach in the engagement session. The wide lens was the 35L but the others belong to the 85L 1.2 I hope it helps.

"On the plus side, the equivalent focal length will be closer to 130mm, which provides more telephoto compression of features and a greater working distance from the subject. "
Not true. It just crops the image. If anything telephoto compression is reduced as you would have to back up to gain the same FOV as a FF sensor.

A nice review, but you say the following : "On the plus side, the equivalent focal length will be closer to 130mm, which provides more telephoto compression of features and a greater working distance from the subject."
Greater working distance OK, but more telephoto compression is false. A crop camera crops, that is the only difference. It gives you more DOF, the crop factor X the F value. An F2.8 on 1.6 crop-factor camera gives you an DOF F2.8 x 1.6= F4,48. But it does not compress the 85 mm into a 130 mm.