Camera to LensAlign Distance (Canon 50x recommendation)

[Michael Tapes - (MichaelT)]

From Joe:

Based off Chuck Westfall's December 2008 tech tips....

He states when performing AF micro-adjustments tests, to perform the tests at 50x the focal length. (I have read multiple posts about this and it seems distance to focus chart does play a role in the adjustment number, but lets assume we go with Canon's recommendation)

Camera: Canon 1Ds Mark III
Lens: Canon EF 24-70mm f/2.8L USM
Calibrated @ 70mm. 70mm x 50 = 3500mm == 11.48 ft.

Using an online depth-of-field calculator, the total depth-of-field @ f/2.8 for 70mm on 1Ds3 is 1.37 ft. Ok. Sounds just fine. But here is the problem. The LensAlign ruler is only 9.5 inches, and the tick marks only take up 8.5 inches on the ruler itself. The total depth-of-field is greater the length of the ruler itself.

I took a couple of test shots. Even after trying "Find Edges" or emboss filter in Photoshop, it is clearly visible that the entire ruler is in focus. I had to move in quite close in order to get the DOF small enough to make any judgments about the AF.

I even briefly tried the lens I bought the LensAlign for in the first place. The Canon EF 85mm f/1.2L II. After a quick calculation, 85mm x 50 = 4250mm == 13.94 ft. DOF for the 1Ds3 @ f1.2 is 0.57 ft. This results were easier to see then the 24-70, but certainly no trivial task. Again, the DOF (0.6 ft) is almost the length of the entire ruler.

I know there is talk about the Long Ruler Kit, but the 24-70/2.8 is no monster telelphoto lens. Neither is the 85/1.2. So either the 50x distance recommended by Canon and Chuck are incorrect, or LensAlign has a ruler length issue... Or I am ignorant of something, and my entire test is invalid for some reason I am overlooking.

Any thoughts? This all seems very, very nitpicky, but I guess you have to be a nitpicker in the first place to buy such a device and test your equipment to such tolerances :)

Joe.

[Michael Tapes - (MichaelT)]

Hi Joe,

Canon Inc. officially tests and calibrates their lenses at 50x, so that is where the recommendation from Chuck came from. But one has to use some intuitive judgment as well. For example, if you normally take full face portraits using your 85mm f1.2, then you would test at a much closer distance; one that reflects the usage of the lens.

For the 70mm f2.8, it simply has a pretty wide DOF as you pointed out so unless you test at very close distances it will not show distinctively on the LA ruler display. But then again with that wide a DOF wide open, why would one be tweaking it to the nearest 1/10 of an inch with AFMA?

So here is the guide-line from me. Use practical sense. Canon's recommendation is simple a generic recommendation in absence of a specific focusing need. So by using 50x, the AFMA will be general in nature, improve focus of the lens at most if not all distances, and not screw anything up. Whereas if you adjust for a specific close distance it might not play out properly at farther distances.

Also practical and logistical issues come into play in terms of where one could and would test at long distances. So keeping all of that in mind, make a practical adjustment, and then test at longer distances by using LA as well as real world shots. Turn AFMA on and off and see if the adjusted position is better or worse at longer distances. Or get the Long Ruler Kit and do actual tests at long distances. But always take real world shots to confirm.

And never adjust at the MFD (minimum focus distance) unless that is the distance that you shoot at, like with a macro lens. In fact that is a perfect example. If you were adjusting a 100mm F2.8 macro, you would not test and adjust at 16 feet (100 x 50 = 5000mm).

Hope this helps..

[Tim Leverett - (tlever)]

To quote Pink Floyd: I gotta admit, I'm a little bit confused. You mention here and I think I saw it in other posts, that we should be testing lenses using the scenarios we shoot in with that lens. I find this a slightly absurd statement for most photographers that I know.

Sure, if you're a very highly paid pro with a bag full of 10 prime lenses and you ONLY use an 85/1.2 for close face portraits, than your theory makes sense. But what if I use it for close face portraits, mid and full body portraits, landscape shots, and some night photography??

And the theory also does not work at all when talking about a zoom. I have a 24-70 2.8L and use it for, well, everything. I couldn't possibly pin down a distance that I commonly use it at for ANY of the available zoom range. So how can I calibrate that lens with LensAlign when the calibrations can be wildly different at the opposite ends of the lens and at different physical distances?

I bring this up because none of this is mentioned in ANY of the marketing. This is not an inexpensive product. I paid more than $200 for something that is likely useless according to many of the posts I'm seeing here about distance and focal lengths measuring completely differently. Perhaps these issues should be mentioned in the marketing. Perhaps this product should only be marketed to, and should contain a caveat, people who use prime lenses in very specific singular functions?

I don't want to completely rag on your product, but this seems like a deal-breaker flaw in the product. Am I missing something?

[Michael Tapes - (MichaelT)]

Hi Tim,

Let me see if I can help to clear up your confusion. Let me try to go point by point...

1 - Testing distance. We recommend 25x the focal length of the lens. Simply thought of as 8 feet per 100mm. This distance is optimum because it allows a relatively narrow depth of field, which allows you to visually see any focus error while allowing any AF Adjustment (if your camera has it) to be done such that it is generic in nature and will be effective for most all shooting circumstances (distances).

2 - Regarding zoom lenses - here we have to compromise because at present time most camera do not allow a separate setting for multiple zoom positions.

The exception being Olympus who does allow separate settings for the short and long end of the zoom range. I have not yet tested the Olympus cameras so cannot comment on the effectiveness of their Adjustment functionality, however. Working on a camera to test this.

So the technique for a zoom lens, assuming that a generic setting is desired, is to first test at the long end of the zoom, and determine the AF Adjustment needed. Then do an additional test at the short (and potentially intermediate) position, setting the proper camera to LensAlign distance for each zoom setting. If the AF adjustment works well at both zoom positions, then you are done. If they are different then a compromise setting may be in order. Alternatively, you may decide to have the camera calibrated by the factory or service center to resolve the issue. That is up to each photographer.

Most of our customers use LensAlign to test their gear, and use the AF Adjust to fine tune their body/lens combinations. But some of our customers simply use LensAlign to test their systems, and then use the test photos as documentation for the factory or test center. This makes it very clear to the service facility as to the problem, and allows the customer to confirm that the problem was properly corrected when the gear returns.

3 - I admit and apologize (as I have before) for the lack of complete documentation for the LensAlign system. We are working hard to correct this. It has been our experience that most customers do fine with the documentation and videos we have provided, as well as the many discussions of LensAlign on Blogs and reviews, such as the Luminous-Landscape review and video. A complete list of reviews is here, and our documentation page is here. Those that have questions ask them and they are promptly answered. These have been via email, but I am encouraging everyone to post any questions here at the forum so that all can share in the discussion. (and I thank you for posting here rather than emailing).

4 - Regarding your 24-70. There is no need to test at all distances. Test at 5.5 feet for the 70mm end and then check at 2 feet for the 24 end. The generic test distances should allow the AF Adjust to improve things in almost all general shooting.

5 - Maybe this analogy will help (but maybe not)....

A CF card does not make pictures. Even after you install it into your camera. It's job is to perform the storage function, and the quality of the pictures will be dependent on many factors, but not the CF card. It's job is to store the data.

Likewise the LensAlign will not correct any focus errors in for your camera and lenses. It's job is to allow the accurate and repeatable testing of the Focus system for each body/lens combination, and determine the health (performance) of the focus system, for each combination and in general.

6 - The AF Adjust system now featured in every (to the best of my knowledge) new mid to high end DSLR is direct evidence that the camera companies acknowledge that the AF systems do not always perform up to expectation in many cases. So they now provide an adjustment for the user to help to improve the AF performance. LensAlign provides the missing link between the AF Adjust feature and the ability to have it be useful to the customer.

Having said all of the above, the AF Adjust mechanism is not the complete answer to AF performance. It is a help, and in a large number of cases it can provide an additional margin of "safety" against soft shots caused by AF system errors. Through our own extensive testing and customer feedback, as well as discussion with camera makers, lens makers, and repair facilities, we know that the AF problems are real, and can be minimized by the AF Adjustment feature in most situations. Some errors detected by LensAlign can be quite severe and require factory service. And in many casituations LensAlign will simply confirm that the AF system for a given lens and body is working properly, which is valuable information to most photographers.

My 70-200 Nikkor on my D700 requires a -7 adjustment. Today a friend brought over his D300 and 70-200 and it tested perfect. In both cases LensAlign has done it's intended job to allow a user to test and asses the performance of the AF (or manual) focus system in their kit.

I hope this helps to clear up your confusion. If you have further questions please post, or if you are having a specific issue better handled by email or phone we can do that as well. My apologies for any confusion previously caused.

Thanks..

[Olivier Hericord - (hangon)]

Originally Posted by MichaelT

Hi Joe,

And never adjust at the MFD (minimum focus distance) .

why? at that distance the DOF is razor fin... it would help to detect a little front or back focus???

[Michael Tapes - (MichaelT)]

Originally Posted by hangon

why? at that distance the DOF is razor fin... it would help to detect a little front or back focus???

Yes. And to test at the MFD is fine, but it can be "dangerous" to base the AF Adjustment on the data from the MFD because the Adjustment as determined by MFD might adversely affect the AF accuracy at longer distances. That is the only reason we recommend a minimum AF Adjust distance of 25xFL.

However, as an alternative method, you can also test and adjust at MFD, and then check it at 25xFL (or longer) to see if the adjustment is valid. It is clearly easier to test and adjust at MFD, so maybe this alternate method will be faster in the end. I have not really worked that way but will experiment with it next chance I get. I believe that the results will vary from lens to lens as to whether the MFD adjustment will be valid for longer distances.

Hope this helps...

[Olivier Hericord - (hangon)]

Originally Posted by MichaelT

I believe that the results will vary from lens to lens as to whether the MFD adjustment will be valid.

Hope this helps...

perhaps it would be cool to maintain a list on the site for that

[Michael Tapes - (MichaelT)]

Originally Posted by hangon

perhaps it would be cool to maintain a list on the site for that

There are sometimes behind the scenes differences between lens of the same make and model, so hard to say that "this model" lens performs in a certain way. Because a lens is on the market for so long, there are production changes that go on with no notification or markings so anyone can tell. Sometimes these changes are to fix problems, sometimes to reduce the cost of manufacture.

We will be creating a database of our internal lens tests and AF adjustment findings, and while they will not apply to any other "serial #s" except the ones we test, I am hoping that it will give you an idea of the range of native performance and AF Adjustments that we find in the equipment that we own and borrow for test.