Calibrating the lens

[victor_ng2]

<p>Hi,<br>

I'd like to ask you about calibrating the camera lens. I've read some articles about it but I'm not quite understand, please help me out.<br>

First and foremost, what is calibrating the lens? Do we really need to do that on new, quite advanced technology nowaday lens? What's the benefit of doing that vs. not doing that? Is there a "clear" different result between doing that and not doing that? If yes, please show me. I've never done this before, please show me in detail. Thanks!</p>

[bobatkins]

<p>http://photographylife.com/how-to-calibrate-lenses</p>

[ellis_vener_photography]

If you mean autofocus micro-adjustment , you aren't doing anything to the lens. You are adjusting an individual camera's

autofocus system to work better with that individual lens.

 

I use the LensAlign/FocusTune target/software combination for this. Http://www.lensalign.com

[q.g._de_bakker]

Good question, Victor. Focus being controlled by a computer, making it possible to bias that computer at any time later on, is possibly used to skimp on making sure a lens is properly adjusted to begin with. Not really a benefit. Advanced technology is not used to achieve advanced and precise products.<br>If there really is a clear before-after difference, that would show that they indeed use it as an excuse to produce ill-adjusted lenses. If there isn't such a difference, it is a pointless feature. I rather hope that it is the latter, though i fear it is not.

[Ed_Ingold]

<p>Sensors used for focusing are located in the prism housing of a DSLR, and may not correspond exactly with the focal plane. Furthermore phase detection focusing achieves high speed by predicting the direction and distance the lens is to be turned, then reverts to contrast sensing for fine-tuning. A chip in the lens is used for predictive focusing, and may be out of calibration. Contrast focusing is based on sensing edge clarity, but is much slower and tends to hunt.</p>

<p>Accuracy becomes important mainly when the lens has an extremely shallow depth of field, typical of long telephotos and fast normal lenses. The lens chip can be reprogrammed to improve predictive accuracy, and many cameras allow a software calibration. this must be done if the focusing mechanism is replaced, or to accommodate normal manufacturing tolerances, damage or wear.</p>

<p>Manual lenses usually have a hard infinity stop. True infinity focus must be accurate to a very tight tolerance (< 0.001"). Sometimes shims are used to position the lens in the focusing body. Other times, the mechanical stop itself may be adjusted.</p>

[q.g._de_bakker]

Indeed.<br>And all those adjustments should be part of the manufacturing proces. No lens or camera should leave the factory still needing to be adjusted.

[Ed_Ingold]

<p>Cameras are like any other precision instrument, designed and built to standards which include reasonable tolerances. Leicas, medium format cameras and some others are individually fine-tuned at the factory, especially normalizing the position of the sensor. However even these adjustments are subject to tolerances. It's part of life and the real world. Design and maintenance of the process are paramount. It is an established principle that you can't inspect quality into a product, it must be built in (are you listening, GM?). In fairness to Leica and Hasselblad, the "quality" of design includes adjustable but stable key components. AFIK, once done, they don't have to be repeated. (You can't cast an housing (+/- 0.005") or cut (+/- 0.001") a mounting pad to the precision needed to focus an image whose DOF at the film plane is smaller than manufacturing tolerances. Something has to be adjustable at some point in time, or you need an entirely different process, like grinding or lapping*.)</p>

<p>Lenses? Like any other mechanical device, cameras and lenses wear out and tolerances increase. If the position of the mirror is off by 0.001", so will be the focusing in a DSLR. Mirrorless cameras have a distinct advantage in that the focus sensors are embedded in the image plane itself, with no intervening moving parts. The most rudimentary example of this philosophy is live view.</p>

<p>* A surface grinder can be set up to spark a piece of metal, write on the ground surface with a pencil, then erase the pencil mark on a second pass without sparking. A pencil mark is about 0.0001" thick. A really good surface grinder can be offset then restored to the original setting and do the same thing with the pencil mark. Lenses are lapped to an accuracy of 1/4 wavelength of light - a few millionths of an inch.</p>

[Robin Smith]

<p>My experience is that you probably will not need to do any "calibrating" - you only will need to do this if you notice that something is wrong.</p>

[aplumpton]

<p>I have confidence in my own capacity to focus and like to selectively focus so I do not use autofocus optics or even technolgy that assures accurate focus. That leaves me open to possible (and even likely) errors caused by using different lenses on different bodies and use of adapters when the mounts differ. </p>

<p>Before autofocus and autoexposure made errors less likely and digital recording allowed quick correction of exposure and focus we had to calibrate the aperture of each lens and the camera mechanical shutter speeds as the variance from the indicated values was sometimes quite important, especially for those using slide films of more limited latitude. If one uses manual aperture and speed selections on modern cameras the values can still vary from those indicated, such as speeds at higher settings being slower than indicated.</p>

<p>The main calibration (or information) I need with my lenses is which f stops are optimum for resolution and contrast and what are the trade-offs between central and corner resolutions. </p>

<p> </p>

[q.g._de_bakker]

I too find not much wrong with manual focus. Works great, though not always as fast as autofocus. The error rate however is much, much higher using autofocus And false focus makes using autofocus often a lot slower than manual focus.<br><br>As long as each camera is adjusted properly, such that optical paths from lens mount to film and to viewfinder are the same, there is no risk of errors due to using different lenses on different bodies, Arthur. What you see is what you get.<br><br>In my experience, f-stops were/are fairly accurate. (Shutterspeeds usually were not quite at the very shortest speed end of the range.) The thing that needed attention was that f-stops are correct at infinity, and that it isn't when you get near 2:1 or thereabouts before they are just nominal values and you have to use the calculated effective value. Built-in meters removed that worry.<br>The f-stop to use for highest or most even resolution and/or least aberrations is indeed something we have to find out and remember, per lens.