Gene51 wrote:
Good test.
However, either the camera or the lens in your setup is bad - either the camera or the lens - possibly both. If the lens does better with other bodies, then the 5DmIII is suspect and should go in for adjustment. If your assessment that the lens is misbehaving, then it should go in for adjustment. A -10 adjustment is pretty substantial.
The 5DMIII has a two point adjustment, doesn't it?
How do other lenses do on this body?
How does this lens do on other bodies?
What about distances beyon 15ft - do you only use this lens between 5 and 15 ft?
Are you ok with the fairly substantial defocus at 3 ft?
Are your test results an average of multiple firings? If you fire the shutter 10 times, let's say at 15 ft - are the results always -11?
My experience with factory tech-trained technicians and the results they produce has always been better than anything I could achieve with the in-camera adjustments in the camera menu. When I send in a camera or lens - the precision and accuracy is better. I don't have time to fuss over this - if it is sharp to my eye, it's fine. I cannot tell the difference -1 and +1 - a spread of 2 points - especially when I stop my lens down to F5.6 or F8, so for me factory adjustment to spec is more that adequate. I am not disputing your results, btw. I think it was very illuminating - and supports both of our opinions - but I'd like to see more vigorous testing - on more lenses and bodies.
Good test. br br However, either the camera or t... (
show quote)
Hi Gene,
Thanks for taking the time to respond. So let me take your questions one at a time.
1) neither the lens or the camera is defective. I now use the lens on a 5D4 and while the correction is a little different, it’s still in the same direction and even without the correction, the lens is tack sharp. If you don’t correct your lenses, a correction of + or - 7-10 may seem like a lot - it isn’t. I have 8 Canon lenses, and 6 of then are L series and all of them require some correction to be perfect - it varies from 2-13 in either direction. Let me hasten to add that all of them are perfectly acceptable uncorrected, especially if stopped down a stop or two, but they are measurably better when corrected as I often shoot low light, wide open. And I would mention that is what many feel their lenses don’t require correction: (a) they have never seen them sharper, so they have nothing to compare to, and (b) if you typically stop down a couple of stops (and many do as a matter of course), the DOF “hides” the error.
2) yes, I do use the lens at other distances besides 5-15’, but those were the distances where I typically use the lens and I chose to demonstrate that while the correction does change (and here’s the point), it’s always better than no correction.
3) if you’ll notice the shot at the end taken at infinity, you’ll see that the lens still focuses fine, contrary to the assertion (and what I was demonstrating) that even with a substantial correction, it doesn’t prevent the lens from focusing at infinity.
4) you also asked if I mind a substantial defocus at 3’. The answer is that I rarely use a 135 at less than 3’ (I’ll have to look up the nearest focus for this lens) - I have a closer focusing true macro for those shots.
5) actually FoCal also has a seperate consistency test that defocuses and refocuses the lens multiple times and plots the error. In fact, that’s exactly what it does during the calibration - it defocuses and refocuses the lens multiple times at each correction point and all of those points are plotted on the graph, so it’s immediately apparent if a lens is inconsistent.
And that brings me to another point. The typical user that doesn’t calibrate their lenses doesn’t know (a) if their lens is focusing correctly - they have nothing to compare it to (b) when they acquire a new lens, they have no way of knowing if it’s a good copy except a subjective judgement as to whether a give image looks sharp (c) they also don’t know at what apertures their lens is sharpest (often assuming that’s a couple of stops down, which is NOT true of all lenses) (d) they don’t know how much acuity loss there is caused by diffraction at smaller apertures, and (e) they don’t know the repeatability of the AF. I immediately know all these things about all my lenses, not by subjective judgements, but by actual objective measurements. Later I will add a chart of my various lens vs correction and it will be apparent that almost all benefit from some correction. It may only increase the acuity 5-10%, but since I’m often paying $1000 for a lens or more, I want every last bit of sharpness it can deliver including and especially side open.
I come from a precision measurement and QA perspective where we calibrate all our tools because we know that every dimension and assembly has a tolerance, and if you add two complex assemblies together, the tolerances may stack up so they cancel, making everything “perfect” (which VERY rarely happens), but more likely they add to make the tolerance of the entire system worse than the tolerance of each individual part, so we calibrate the entire assembly, which is what lens calibration does. We also know that if your measurement instrument is flawed, or requires a subjective judgement, then it may lead to wrong conclusions. Case in point. Almost every week an OP posts an unsatisfactory image taken with a new lens or camera and wonders what the issue is. Is it AF, is it movement, is it inadequate DOF or SS? Did he get a “bad copy” of the lens? I never have any of those questions, because for less than $100 for ALL my lenses, present and future, I KNOW that I have a good lens or not and that I’m getting 100% of the acuity I paid for. And I don’t have to live with everything being “within factory tolerance”, but not as good as it could be. I don’t have to have my lens “matched” to a specific body, and I don’t have to ship my lenses and cameras both ways to and from the factory subjecting them to shock or damage which can undo the very reason I had them adjusted. Seems like a bargain to me.