I did a test of measured focusing distances last year, and yes, a 4 to 6 inch difference is very possible. Here is a photo of the test setup
_MG_6565_20D-35mm f-5 FLASH ISO 800.jpg photo - Derrel photos at pbase.com
and here is the caption from that post I did:A lot of people on the web suggest the focus-and-recompose approach, using the center AF point,but that leads to inaccurate autofocusing under challenging conditions. Unfortunately, at closer distances my Canon 5D and my EOS 20D will MIS-focus a shallow depth of field lens like the 135mm f/2 L at indoor distances, and the same holds true with other cameras that have most or all of their autofocus brackets in the very center of the frame. If one wants to do a little bit of drawing with some graph paper and a straight edge, it's possible to see that the distance from the center of the frame to the focal plane of a camera is NOT the same as the distance from the top or bottom edge of a frame. If you want to prove it to yourself, take a piece of string or fishing line and tie it to a tripod-mounted camera's lens,and then physically "string out" the distances to what lies under the center AF bracket, and the far edges of the frame, on your office wall.
I just set up my 20D on a tripod at 128 inches from a wall. The measured distance to a Post-It note positioned 12 inches inside the left edge of the frame my 18-55 set to 35mm measured exactly five and one-half inches farther away than the measured distance to a Post-It note positioned under the center AF bracket. The focusing distance arrived at by using focus-and-recompse for a person's face at the left edge of the frame could/would be off by almost six inches! The distance error to the right-hand Post-It note positioned 12 inches inside the FOV of the lens was off by a measured three and one-half inches. So, maybe the wall is not square, or my eyeballing alignment of the camera was not good enough. Average the readings,and the focusing error using focusing and recompose would be about 4.5 inches at 128 inches shooting distance. That is nowhere good enough to "cover" your butt when shooting an ultra-speed lens in a real-world situation. You can NOT rely on focus-and-recompose using the center AF bracket when the subject is off-center. Try the experiment for yourselves if you want a real eye-opener.
With a wide aperture lens at close distances, depth of field at f/1.4 or f/1.8 is very shallow. Using focus and recompose is about the only way to do autofocusing on my 5D or 20D--all of the AF brackets are very much in the middle of the frame. Some of the professional cameras and semi-pro bodies have wide-area AF systems,where AF brackets cover almost the entire frame. Nikon D2x has a wide-area AF system; Nikon's D300 has a 51-point AF system with AF brackets blanketing almost the entire focusing screen. I have taken a lot of heat primarily because I do not advocate using the center AF point in situations where a lot of other people do; much of that is from my 2005-2009 use of the Nikon D2x camera, which has what Mr. Hogan referred to as "the D2x's focus anywhere capability". I hate to sound like a homer, but the ability to autofocus near the edges of the frame,both the left and right,and the top and bottom edges of the frame, is one of the reasons I still have a D2x; it has the absolute-best focusing system I have ever seen, and I have some very wide-aperture lenses, which that camera can autofocus,even out to the most-extreme AF points, even in low light. Lenses like 85/1.4, 105 and 135 f/2, and 200/2 and 300/2.8: with the D2x's professional-level, wide-area AF system, it's easy to focus any lens I have using the AF sensor, or small group of AF sensors, that will deliver the right AF off-center by quite a ways, and not be fooled by back-focusing issues or front-focusing issues. That is not my experience with the 20D and 5D and its highly centrally-located AF system.
I bought the 20D in mid-2005 I guess it was, and the 5D in mid-2006,and have noticed one thing,and have taken a lot of flack for it, but the 20D and 5D AF systems are simply NOT able to focus well using the outer focusing points, necessitating focus-and-recompose or manual focusing. When using wide-aperture lenses at close distances, the Depth of FIeld can be as little as 1 inch; focus and recompose when using the camera in tall orientation on a standing person will lead to an AF lock that might be simply NOT precise enough for the limited depth of field. The SAME problem has been around for decades even on some of the world's finest rangefinder cameras with lenses like the 75mm f/1.4 Leica lens and 85- and 90mm lenses used on various rangefinder cameras that have a short base. The absolute degree of precision needed to achieve focus when the DOF "band" is one inch or two inches deep simply can not ensure accurate enough focusing on most consumer d-slr models, and it was in fact, a HUGE PITA back in the 1980's with my 85/1.4 and 135 f/2 manual focus Nikkor lenses; today, we have a lot of web-based people who complain that the "lens focuses poorly" or "inconsistently" when using it in poor light, or on moving people in social situations, or when using focus and recompose on consumer-type cameras. The issue of tricky focusing with a 50mm f/1.4 lens in poor light using a split-image rangefinder was probably 10x worse with a Nikon F3 than it is with a Nikon D2x. And yet, despite difficulty focusing a 50mm f/1.4 lens on a Leica M3 or M6, or on any 35mm manual focus SLR, we have people commenting that they cannot place a 1- to 2-inch DOF consistently enough to suit them. The problem has been around for decades! Experienced photographers are aware of the issues involved. The leading edge is also called the bleeding edge.
![[No title]](/data/xfmg/thumbnail/31/31706-3e429b21053f11072ed2e5b37c019073.jpg?1734160380)
