In October of 2013 I bought one of these lenses on Ebay for $199, and then bought the Ed Mika converstion kit for about $140. So the cost of the converted lens is about $339, which compares nicely to the $1600 price tag for the 50mm f/1.2 L lens. What you get when you pay the big money is a rugged autofocus lens.
One fellow says:
I am more than impressed with this lens, I found it sharper wide open than the C/Y Zeiss 50mm f1.4 with better contrast and comparable colors.
The Ed Mika kit is excellent. For this lens, the conversion process took roughly 5 minutes and was well documented. You could, no doubt, do clumsy things and screw it up -- but this would take some effort.
After the conversion the lens mounts on an EF mount camera (such as a 1D). There are problems with full frame 5D cameras at infinity focus. The report is that the camera mirror collides with the rear lens element. Reportedly this does not cause damage - even after 50 crashes, but I intend to avoid this combination. The mirror frame in the 1D is metal (not plastic as in the 5D) and is thin enough that it causes no problems.
If you have a 5D and are crazy you can look into the mirror shaving business. You are actually not shaving the mirror, but the plastic frame that holds the mirror. Apparently most people who try to do this themselves break the mirror mechanism, which leads to an expensive repair. It is delicate and the mirror provides a lot of leverage and bad things happen. The recommended way to do this is to purchase a replacement mirror, have it shaved, then have it installed. This will cost well over $500. After learning these things I decided it was a path that I was not going to follow.
The converted lens includes a focus confirm chip, which turns out to be a very nice feature. It is next to impossible to judge focus using the viewfinder, but the chip causes the camera to report proper focus (even though it cannot adjust it). It does this using a visual indicator in the viewfinder along with an audible "beep" if the beep is enabled. I find the beep to be the perfect thing.
This lens is amazingly tricky to use, especially wide open, due to the shallow depth of field. In fact I began to suspect that either the lens itself was unsharp or that the reported focus had an offset. As you will see, the depth of focus is amazingly shallow, especially at close distances and if you or your subject move at all, everything has changed. In fact if you do as I have long done and use the central focus point then recompose you will loose sharp focus.
The way to lay questions about sharpness and proper focus to rest is to do some experiments, put the camera on a tripod, and take some photos under controlled conditions. Here are two photos of my test subjects at the closest focus distance (about 20 inches for this lens). [As an aside, the 24-70 f/2.8 L lens will focus on objects 6 inches from the front element -- which given the length of the lens is about 13 inches from the sensor plane, an amazingly versatile lens!] The first photo below is at f/1.2 and the second at f/2.8. Notice the change in depth of field.
The following are crops from the above images. Notice that I focused on the threads of the tap, which was directly under the 6 inch mark on the ruler. (Parallax makes it look otherwise, but trust me.) Take note of how shallow the depth of field is, especially in the f/1.2 image on the left. My impression is that best focus is about 3/8 of an inch behind the plane I focused on.
As a comparison, I also shot the same setup with my 24-70 f/2.8 L lens, set to a focal length of 60mm (according the EXIF data -- I tried for 55mm) and at a f/2.8 aperture. The following images were both taken at f/2.8, the 55mm on the left, the 24-70 on the right. It looks to me like the 24-70 focuses right on or maybe 1/8 back. Comparing the sharpness of the threads on the tap confirm that the FL 55mm lens has not achieved proper focus.
As another comparison, I shot the following pair of photos both at f/5.6. On the left is the image from the FL 55mm; on the right from the 24-70 set to 50mm.
Both were focused using live-view and shot from a tripod. The raw files were given identical processing in Lightroom. I am satisfied there are no significant image quality or sharpness issues with this 55mm f/1.2 lens. Some crops that will display at full resolution (1:1) follow, first from the 55 f/1.2, then from the 24-70. The 24-70 looks like the sharper of the two lenses.
To put the chip in program mode you put the camera in "M" mode, set the exposure time to 13 seconds. (Any exposure time greater than 11 seconds will do). It is also recommended that you verify that the camera is set for 1/3 stop EV increments, and that you put the camera into AI Servo mode (so it won't refuse to make an out of focus exposure). Then attach the lens. If the lens is already attached, just release the mount and rotate it out of registration and back again. Then within 2 seconds take an exposure.
Now change the camera mode to Av. You write decimal digits to the adapter by taking exposures with various aperture settings. But first you take an exposure to tell the chip what setting you want to modify:
The digit codes are as follows:
Tom's Digital Photography Info / tom@mmto.org