Third-Party Lenses are Different in Different Mounts
This one is going to cause me to get emails, I can tell you that right now. It’s one of those things that is so obvious nobody sees it, and fanboys are going to argue with it, because fanboys don’t let facts stand in the way of a good argument. For starters, the flange-to-sensor distance is different on a Canon camera compared to a Nikon camera compared to a Sony, etc. So theoretically there could be some optical differences in the angle that the image rays impact the microlenses on the sensor. But lets forget about that (because I have no idea if it really makes a difference). Lets focus on focusing.
Every camera manufacturer has a set of electronic algorithms built in to their cameras that talk back and forth with electronic chips in the lens when it autofocuses. When Tamron or Sigma or Tokina make a lens, they have to reverse engineer these algorithms and then make electronic chips. Certain chips (or firmware on the chips) are for talking to Canon cameras, different ones for Nikon cameras, etc. It’s not only possible that the electronics in the lens may work better with one mount than another, it is an absolute fact.
For example, a few years ago (2003 or so) existing Sigma lenses stopped talking to newly released Canon cameras, but they kept working fine with other brands. When the Nikon D3x was released the Sigma 120-300 wouldn’t autofocus with properly with it, although other Sigma lenses worked fine. More recently, the Tamron 17-50 VC and 60 Macro lenses were found to only be able to use the center autofocus point on Canon 50D and 60D cameras, and to not use any cross-type sensors. The reason in this latter case was apparently that Tamron used a “lens ID number” that told the camera it was actually an old, out-of-production Canon lens.
We see it all the time: certain third party lenses seem great on one mount, and have all kinds of focusing issues on another mount. It’s not that one brand of lenses is better than another overall. My point is, when you want to ask someone what they think about a third party lens you’re about to purchase, make sure you ask someone who shoots it on your camera. Otherwise their advice is meaningless.
in my experience, Canon cameras seem to have more trouble with third party lenses than other camera brands do. That’s an impression formed with thousands of lenses and hundreds of camera bodies to test them on, but still, its just my impression and I haven’t seen anyone else say it is so.
Aperture and Focal Length Numbers Aren’t Totally Accurate
Most of you probably know this. But the focal lengths printed on the lenes are approximations. Prime lenses are usually pretty accurate, but zooms may off by at least 5%. So a 50-500mm zoom may actually be 53-475mm. Its not a big deal but it is noticeable sometimes.
What might be important to a lot of people is close-focus, focal-length change that occurs particularly with lenses having rear focusing groups. (Rear focusing groups are all the rage these days because you’re only moving the little elements at the rear of the lens to focus, hence faster autofocus with small autofocus motors.) But when you focus a zoom lens on a close object, the movement of the elements actually shortens the focal length.
So a 70-200 zoom may be 200 fully zoomed when focusing at infinity, but only 150mm when focusing on something nearby. This effect is particularly noticeable on the new Canon and Nikon 70-200 f2.8 zooms and the Canon 70-300 IS L zoom.
Similarly the transmission of light by a lens is never quite up to what the f/stop says it is, the f/# is a theoretical calculation that doesn’t take into account the number of elements, coatings, and other things that might affect light transmission. Just looking at a number of 85 f1.4 lenses, for example, DXO Mark found they were all actually T/1.6 (T being the actual, not theoretical, light transmission). Again, probably not something that really matters to people.
Circular Apertures Aren't Always Circles
I’ve yet to see a lens whose marketers say “straight aperture blades making a pentagram resulting in harsh out-of-focus highlights”. They either have a blurb that says “circular aperture for pleasing out-of-focus highlights” or they say nothing at all. But even when they say circular, they lie a LOT. Take a look at the apertures of your lens stopped down just a bit: some really are circular, many others are octagons, heptagons, whatever.
Not picking on Canon (once again they are just willing to make more information readily available), everyone does it. But the apertures above are both described as “circular” in the brochure. Not so much.
By the way, on cameras that don’t have mechanical levers to let you see the aperture with the lens dismounted, you can see them by mounting the lens, setting the aperture to f/5.6 or so, hitting the depth-of-field preview button on the camera, and the removing the lens without releasing the depth-of-field button or turning the camera off.