Why do people say they have a very sharp copy of a lens?

[martyn]

Why am I always reading that people have a 'very sharp copy' of a lens, especially when its something like a 24-105 f4 L? In Bryan's review of this lens he says:


... the Canon EF 24-105mm f/4 L IS USM Lens is very sharp wide open and shows only very slight improvement when stopped down.


So what does this mean then? Is there really that much play in the manufacture of the L lenses such that the one reviewed was very sharp but others might not be? Also, how do these people know they have a sharp copy vs that lens just being sharp, period. I see people making this comment *way* too often for the answer to be that they've owned multiple copies of that lens.


Or am I missing something?

[Ken Schwarz]

My theory about this is that every lens and body need a little bit of calibration to work together optimally. This isn't because they are broken or otherwise defective, it's just that they need the adjustment to focus as accurately as possible. The latest Canon bodies (the 1D3, 1DS3, 50D, 5D2) provide user control to make this adjustment. Without this calibration, a lens that tests sharp on one body will seem soft on another. Same lens, different result! I recently upgraded from a 30D to the 5D2. One reason for choosing the 5D2 over the less expensive 5D (which would have been satisfactory in many ways) was the ability to make this adjustment myself.


The irony of this is that fast lenses, which tend to be the most expensive, are also the ones that most likely need adjustment. This is because their depth of field is so shallow that even the slightest miscalibration will show up. Lenses intended for consumers (cheaper kit zooms and tele zooms) have enough depth of field that you can't tell if they're off.


I did some initial calibration tests with my 5D2 and found that the f/4 lenses were calibrated well enough that without spending a whole lot more time testing I didn't think I could confidently improve focusing accuracy. An 85/1.2 lens seemed to improve with a bit of adjustment, but there was enough inconsistency in focus from test shot to test shot that I couldn't tell if it was the adjustment or the inconsistency that I was seeing. The lens that needed it the most was the 50/1.4, which was significantly off and much improved with adjustment. Hooray, it worked!


I don't have any f/2.8 zooms, but here I think adjustment would make a difference. A lot of people have complained about the 24-70/2.8 zoom being soft out of the box and going through several returns to get a good one. I'm pretty certain that the problem was that the calibration was off and that any of them would have been fine. It's always possible to get a lemon, but to get three in a row is really unlikely, and I have lost enough money at slots to be very confident in saying so.


In a final note, I think that most people expect "pro" equipment to be more reliable and better adjusted out of the box than "consumer" equipment. From a manufacturer's perspective, the opposite is true! If you ship 1000 units at a 1% defect rate to pros who have local pro service centers, it's entirely managable. If you ship 1,000,000 units to consumers at that same defect rate, you have a nightmare! I'm not saying that Canon neglects testing and manufacturing controls on their pro equipment, but I don't think that it necessarily gets "special" treatment either.

[portej05]

Both cameras and lenses have manufacturing tolerances, and if you got a camera at one end, and a lens at the other, you may find that you don't get a sharp image out of the box. On the other hand, you may get both camera and lens with good calibration straight out of the box.


I'm sure that Canon has fairly good tolerances on most of its equipment, and so this range doesn't produce a huge problem with the majority of equipment.

[Tim]

The L series lenses are more tightly monitored, therefore its safer to assume you will have a sharp lens out of the box. This is the reason why I only had one go at the tamron 17-50. I initially had gotten this lens and it was soft at allapertures, I had read that many people had to return multiple copies until they were satisfied, so I returned it and purchased the 17-40 instead, and it seems to be as sharp as can be.

[Stefan Stuart Fletcher]

<p style="margin: 0cm 0cm 10pt;" class="MsoNormal"]<span style="font-size: small; font-family: Calibri;"]I believe this has to do with four factors:
<p style="margin: 0cm 0cm 0pt 36pt; text-indent: -18pt; mso-list: l0 level1 lfo1;" class="MsoListParagraphCxSpFirst"]<span style="mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"]<span style="mso-list: Ignore;"]<span style="font-size: small; font-family: Calibri;"]1)<span style="font: 7pt 'Times New Roman';"] <span style="font-size: small; font-family: Calibri;"]Resolving power, which is where those bizarre test sheets with horizontal and vertical lines come into play. Simply put, this is the ability of a microscope, telescope or lens to measure the angular separation of images that are close together. As cameras become increasingly powerful light sensors, the importance of resolving power grows. L lenses are supposed to have a higher resolving power, but not always.
<p style="margin: 0cm 0cm 0pt 36pt;" class="MsoListParagraphCxSpMiddle"]<span style="font-size: small; font-family: Calibri;"]Some lenses simply can&rsquo;t resolve the image exposed on the sensor sharply enough, especially cheaper and, in some cases, third-party lenses and particularly those specifically designed for cropped sensors (APS-C, etc.).
<p style="margin: 0cm 0cm 0pt 36pt;" class="MsoListParagraphCxSpMiddle"]<span style="font-size: small; font-family: Calibri;"]Magnify an image you know to be sharp. Look at how details, especially verticals and horizontals, can blur. I test the sharpness of my lenses by photographing a wall-to-wall bookcase and see whether I can read the book titles at various settings.
<p style="margin: 0cm 0cm 0pt 36pt; text-indent: -18pt; mso-list: l0 level1 lfo1;" class="MsoListParagraphCxSpMiddle"]<span style="mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"]<span style="mso-list: Ignore;"]<span style="font-size: small; font-family: Calibri;"]2)<span style="font: 7pt 'Times New Roman';"] <span style="font-size: small; font-family: Calibri;"]Centre and corner sharpness. The optical elements in a lens will handle light differently in the centre compared to those refracted at the edges, especially at different apertures and on different cameras. Modern full-frame bodies are less forgiving than cropped-sensor cameras, if only because the image surface area is smaller (edges are excluded). Those highly technical reviews Bryan and other authorities like dpreview conduct are way over my head, but I know for example that my 17-40 fill focus sharply in the centre and/or corners only at specific apertures.
<p style="margin: 0cm 0cm 0pt 36pt; text-indent: -18pt; mso-list: l0 level1 lfo1;" class="MsoListParagraphCxSpMiddle"]<span style="mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"]<span style="mso-list: Ignore;"]<span style="font-size: small; font-family: Calibri;"]3)<span style="font: 7pt 'Times New Roman';"] <span style="font-size: small; font-family: Calibri;"]Back focusing issues: under specific conditions or with slightly &lsquo;defective&rsquo; lenses, the image is not focused at the point where the sensor lies but slightly in front of it, resulting in a slightly out-of-focus image even though you&rsquo;re sure your AF is working fine and you&rsquo;ve taken every precaution to ensure image stability. In my experience, this happens more often with cheaper, third-party and zoom lenses than L lenses and primes.
<p style="margin: 0cm 0cm 10pt 36pt; text-indent: -18pt; mso-list: l0 level1 lfo1;" class="MsoListParagraphCxSpLast"]<span style="mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"]<span style="mso-list: Ignore;"]<span style="font-size: small; font-family: Calibri;"]4)<span style="font: 7pt 'Times New Roman';"] <span style="font-size: small; font-family: Calibri;"]That&rsquo;s life! We&rsquo;re not perfect and Canon certainly isn&rsquo;t. Sometimes, whatever magic that&rsquo;s in an expensive, highly praised lens just doesn&rsquo;t work. Poor QC, production, who knows? I find Ken Schwarz&rsquo;s final remark very appropriate and all-too likely. I think it is possible to have several copies of a lens and see differences, which is why I prefer to spend more money in a shop where I know I can return a defective lens for a replacement than always looking for the cheapest retailer. I once returned four copies of an L until the shop gave me a decent copy. Moral is, always check out an expensive lens for vignetting, CA and other visible defects before parting with your hard-earned cash. Reputable stores let you test their wares first.
<p style="margin: 0cm 0cm 10pt;" class="MsoNormal"]<span style="font-size: small; font-family: Calibri;"]Hope that helps. As you can see from my shaky definitions, I&rsquo;m not much into the technical side.

[martyn]

Cheers guys, thats interesting. It explains why there are variances in sharpness, but perhaps not why *everyone* seems to have a sharp copy ;-) It seems this is being touted as an exception to the rule when a) it sounds like its not, a duff one is the exception and b) from what you guys have said it might very well only be sharp on that person's body anyway.

I have to say I find all the second point a bit of a worry. Are we saying that if you have a 70-200 2.8 and say a 5D, 20D, and a 1D MK ii, then you can only calibrate the lens to work 'perfectly' with one of the bodies and its going to be slightly off on the others? Whatever the situation, it sounds like I need to establish if my lens is as sharp as it could be.

Is this post purchase calibration something that everyone does and I've just missed this?