Lens calibrating (Page 3): Davey, it's definitely a pain to track down focus issues, especially when you aren't sure if it is a lens, camera or both that are contributing to the...

Gene51 Loc: Yonkers, NY, now in LSD (LowerSlowerDelaware)

[quote=DaveyDitzer]

Gene51 wrote:

Lens "calibration" is best left to those... (show quote)

Davey, it's definitely a pain to track down focus issues, especially when you aren't sure if it is a lens, camera or both that are contributing to the problem. I would start by making sure your camera body or bodies are correct - if not, get them fixed. Then it is easier to test individual lenses and have them fixed. There will be minor errors usually on the order of ±2 on the AF Fine Tune utility in the camera. Funny thing is, that if you take 10 pictures in a row of a static subject you will likely get more than half that are right on the money. So I wouldn't sweat it. I don't.

TriX Loc: Raleigh, NC

Gene51 wrote:

Perfection is the enemy of very good. I am fine wi... (show quote)

Gene, as I said, no one can argue with your results, and if you are happy with 95%, then God bless and carry on. But if you do not check/calibrate with good tools, then you do not have the measurement capability to know how good your lenses COULD be. I may be anal about calibration because I come from a precision background where everything is calibrated, but in engineering, design, medicine and manufacturing, everyone knows that without calibration, measurements and accuracy are meaningless, and optics, because of the very small distances and short wavelengths involved, require the highest level of measurement and calibration, so why should a precision instrument such as a $1,000 lens and a $3,000 camera be any different? In a clinical laboratory, the chem analysis machines are calibrated to a known multiple times per day. In an electronic shop, every instrument is calibrated periodically. In a manufacturing or machine shop, every measurement tool and gauge is regularly calibrated by a standards lab. Even on a military base, they have PMEL labs to calibrate everything from test equipment to gauges.

Having been a designer (and quality manager) in the aerospace industry, every drawing to manufacture a part has a tolerance on every dimension, and when a designer designs an assembly of multiple parts, one of the key things he considers is will the assembly work if all the tolerances go “the wrong way”. I can’t think of assemblies much more precise or complex than a modern lens with 10-20 elements. You then assemble that to a camera with hundreds of parts, some number of which are associated with AF. Each of those goes through QC and are measured with carefully calibrated instruments and must fall within a range of tolerances, so each assembly can vary within that range. You now take two of those complex assemblies and mate them. Each may pass final QC and be within tolerance, but they are NEVER tested together unless you, the user does so. The net-net is that you may have a body that barely fits into the negative end of the AF tolerance range and a lens that barely fits within the positive end of the tolerance range, and when you put them together, the result is that the AF is off. It may look good to your uncalibrated eye, but it could be better if the parts were matched, and that is what MFA does. It isn’t perfect, but it’s a big improvement of doing nothing. Canon and others provide one adjustment for a prime and two adjustments for a zoom, while some manufacturers with “docks” provide four or more. If you look at the link to the test I performed, you’ll see that the compensation isn’t perfect at all distances, but it is ALWAYS better than no compensation.

Now I have previously posted a comparison of a high quality uncalibrated lens and then a comparison with it after being calibrated. Everyone who responded could see the difference, and if you like I can publish the results again. But all users who have calibrated a lens and have seen the difference already know the improvement it can make - it is mentioned by multiple responders on every thread I’ve seen on MFA. In fairness, there are also pros, like yourself, who never calibrate and show great results, so maybe it would be interesting to have a thread and see how many working pros do and don’t calibrate their lenses, but for me, I already know the answer. I have seen the actual acuity measurements and side-by-side comparisons of uncalibrated vs calibrated for all my lenses, and I know it makes a difference, and for me (and many others) that $70 and 5 minutes per lens periodically is worth that 5% and knowing that my system is as sharp as it can be. I honestly think that if you actually tried it and saw the results, you might change your opinion, but in the interim, we can still be friends and agree to disagree.

Cheers,
Chris

dyximan

[quote=CHG_CANON]
Where would I go to find out the information about whether the 18 to 300 is capable of moving subjects other than just practice

TonyBrown

Maybe also increase shutter speed as 1/1000 is slow for a big shot. 1/2500+ for larger slower moving birds.

sscnxy

Nalu wrote:

One of the things I don’t miss about my now replaced Canon gear was the need to fine tune my lenses with various bodies. I always wondered whether I was getting the best out of my lenses. I don’t have to do that with my Sony gear.

Sorry, just had to say it.

Mirrorless cameras don't need AF fine tuning. But that's not the case with DSLR cameras. All their lenses should be checked for front or back focus issues and AF fine tuned if either exists. What's the point of shooting anything that is not tack sharp, unless fuzziness in your images is a deliberate, artistic choice.

NMY

Gene51 Loc: Yonkers, NY, now in LSD (LowerSlowerDelaware)

TriX wrote:

Gene, as I said, no one can argue with your results, and if you are happy with 95%, then God bless and carry on. But if you do not check/calibrate with good tools, then you do not have the measurement capability to know how good your lenses COULD be. I may be anal about calibration because I come from a precision background where everything is calibrated, but in engineering, design, medicine and manufacturing, everyone knows that without calibration, measurements and accuracy are meaningless, and optics, because of the very small distances and short wavelengths involved, require the highest level of measurement and calibration, so why should a precision instrument such as a $1,000 lens and a $3,000 camera be any different? In a clinical laboratory, the chem analysis machines are calibrated to a known multiple times per day. In an electronic shop, every instrument is calibrated periodically. In a manufacturing or machine shop, every measurement tool and gauge is regularly calibrated by a standards lab. Even on a military base, they have PMEL labs to calibrate everything from test equipment to gauges.

Having been a designer (and quality manager) in the aerospace industry, every drawing to manufacture a part has a tolerance on every dimension, and when a designer designs an assembly of multiple parts, one of the key things he considers is will the assembly work if all the tolerances go “the wrong way”. I can’t think of assemblies much more precise or complex than a modern lens with 10-20 elements. You then assemble that to a camera with hundreds of parts, some number of which are associated with AF. Each of those goes through QC and are measured with carefully calibrated instruments and must fall within a range of tolerances, so each assembly can vary within that range. You now take two of those complex assemblies and mate them. Each may pass final QC and be within tolerance, but they are NEVER tested together unless you, the user does so. The net-net is that you may have a body that barely fits into the negative end of the AF tolerance range and a lens that barely fits within the positive end of the tolerance range, and when you put them together, the result is that the AF is off. It may look good to your uncalibrated eye, but it could be better if the parts were matched, and that is what MFA does. It isn’t perfect, but it’s a big improvement of doing nothing. Canon and others provide one adjustment for a prime and two adjustments for a zoom, while some manufacturers with “docks” provide four or more. If you look at the link to the test I performed, you’ll see that the compensation isn’t perfect at all distances, but it is ALWAYS better than no compensation.

Now I have previously posted a comparison of a high quality uncalibrated lens and then a comparison with it after being calibrated. Everyone who responded could see the difference, and if you like I can publish the results again. But all users who have calibrated a lens and have seen the difference already know the improvement it can make - it is mentioned by multiple responders on every thread I’ve seen on MFA. In fairness, there are also pros, like yourself, who never calibrate and show great results, so maybe it would be interesting to have a thread and see how many working pros do and don’t calibrate their lenses, but for me, I already know the answer. I have seen the actual acuity measurements and side-by-side comparisons of uncalibrated vs calibrated for all my lenses, and I know it makes a difference, and for me (and many others) that $70 and 5 minutes per lens periodically is worth that 5% and knowing that my system is as sharp as it can be. I honestly think that if you actually tried it and saw the results, you might change your opinion, but in the interim, we can still be friends and agree to disagree.

Cheers,
Chris

Gene, as I said, no one can argue with your result... (show quote)

I don't think you are understanding my point, Chris. I am not arguing that there is no difference. What I am simply stating is that due to tolerances in the lens and in the camera, if you do MFA or even if you send a camera and lens in for calibration - something I did with my 600mmF4 and my D800, you are never going to see "perfect" focus 100% of the time anyway. With your background I am sure you understand the difference between accuracy and precision. I am presenting the notion that even if you adjust everything for accuracy, as the gear ages and tolerances increase, there is a loss of precision, and what once was a lens/camera that was adjusted to produce an accurate focus result 90% or even 95% of the time, after it's been in use for a time it will likely require readjusting, and the level of precision that once possible is no longer due to mechanical wear.

I see the MFA as a quick and dirty solution that can get you good results - but it is not that panacea that many proponents make it out to be - at least not in practice. I have much more faith in the lenses that offer a dock simply because you can adjust for multiple focus distances (and multiple focal lengths if you have a zoom). The calibration procedure is a dead giveaway - in that for the best results with a zoom you need to pick your favorite focal length and working distance. In my case, I will use my 150-600 to shoot anything from 15 ft to infinity - an insect, the moon, and everything in between. I have tried the Focal system in the past, which was good for what it did, but I found it pretty useless for zooms, so I returned it. As I relayed earlier - I did have a bad D800 that could not focus correctly with my lens - a 600mm F4. It went to Nikon with the the lens (actually carried it there since I only lived an hour away), and was clear that the lens was fine with my other cameras, so they were not to touch the lens other than to check it's focus for accuracy. All the adjustment had to be done in the camera. When I got it back it was perfect - and exactly the same as my untuned D800 - it wasn't better. I used to have a favorite target - a street sign that was about 50 ft away that had a reflective material with a relatively small pattern of fibers on it. Only a lens that was perfectly focused could capture it with a 36mp camera, and it was all but invisible with a 12 mp DX camera. I could see no difference between the cameras. It was at that point that I decided that having everything within spec was sufficient, and that accepting that a particular combination might be out by a percentage point or two was not important. I also weighed the time cost of calibrating everything with MFA and decided I wanted better, more consistent focus across multiple distances and focal lengths than that process could provide.

The other side of this is what happens when you get your picture from the camera to print or projection. All that acuity and fine detail is lost anyway when all is said and done.

I like Canon's system for MFA since it does try to address zoom lenses by providing an adjustment for wide and another for tele - and it does address individual lenses and being able to provide an offset adjustment for multiple lenses at one time, but they still recommend picking your favorite working distance and focal length and adjusting for that. I am intrigued by the AutoFineTune utility in the D500 and the D850, but not enough to warrant the purchase of either. I still use my D800 and D810 as my main cameras and am still getting great results with them.

I suppose I will try to test a couple of lenses and cameras at some point since I do have an open mind about stuff, despite my strong opinions. But as I said earlier, none of this is enough to lose sleep over. Different strokes. I am happy that you were able to achieve the results you were looking for with MFA, btw.

(Download)

TriX Loc: Raleigh, NC

Gene51 wrote:

I don't think you are understanding my point, Chris. I am not arguing that there is no difference. What I am simply stating is that due to tolerances in the lens and in the camera, if you do MFA or even if you send a camera and lens in for calibration - something I did with my 600mmF4 and my D800, you are never going to see "perfect" focus 100% of the time anyway. With your background I am sure you understand the difference between accuracy and precision. I am presenting the notion that even if you adjust everything for accuracy, as the gear ages and tolerances increase, there is a loss of precision, and what once was a lens/camera that was adjusted to produce an accurate focus result 90% or even 95% of the time, after it's been in use for a time it will likely require readjusting, and the level of precision that once possible is no longer due to mechanical wear.

I see the MFA as a quick and dirty solution that can get you good results - but it is not that panacea that many proponents make it out to be - at least not in practice. I have much more faith in the lenses that offer a dock simply because you can adjust for multiple focus distances (and multiple focal lengths if you have a zoom). The calibration procedure is a dead giveaway - in that for the best results with a zoom you need to pick your favorite focal length and working distance. In my case, I will use my 150-600 to shoot anything from 15 ft to infinity - an insect, the moon, and everything in between. I have tried the Focal system in the past, which was good for what it did, but I found it pretty useless for zooms, so I returned it. As I relayed earlier - I did have a bad D800 that could not focus correctly with my lens - a 600mm F4. It went to Nikon with the the lens (actually carried it there since I only lived an hour away), and was clear that the lens was fine with my other cameras, so they were not to touch the lens other than to check it's focus for accuracy. All the adjustment had to be done in the camera. When I got it back it was perfect - and exactly the same as my untuned D800 - it wasn't better. I used to have a favorite target - a street sign that was about 50 ft away that had a reflective material with a relatively small pattern of fibers on it. Only a lens that was perfectly focused could capture it with a 36mp camera, and it was all but invisible with a 12 mp DX camera. I could see no difference between the cameras. It was at that point that I decided that having everything within spec was sufficient, and that accepting that a particular combination might be out by a percentage point or two was not important. I also weighed the time cost of calibrating everything with MFA and decided I wanted better, more consistent focus across multiple distances and focal lengths than that process could provide.

The other side of this is what happens when you get your picture from the camera to print or projection. All that acuity and fine detail is lost anyway when all is said and done.

I like Canon's system for MFA since it does try to address zoom lenses by providing an adjustment for wide and another for tele - and it does address individual lenses and being able to provide an offset adjustment for multiple lenses at one time, but they still recommend picking your favorite working distance and focal length and adjusting for that. I am intrigued by the AutoFineTune utility in the D500 and the D850, but not enough to warrant the purchase of either. I still use my D800 and D810 as my main cameras and am still getting great results with them.

I suppose I will try to test a couple of lenses and cameras at some point since I do have an open mind about stuff, despite my strong opinions. But as I said earlier, none of this is enough to lose sleep over. Different strokes. I am happy that you were able to achieve the results you were looking for with MFA, btw.

I don't think you are understanding my point, Chri... (show quote)

Thanks Gene for the thorough and thoughtful response. The quality of your work shows that whatever you do certainly works for you. I think that whether MFA makes a visible difference largely depends on what you shoot and may account for some of the difference in user experience. If one shoots long teles or macros or alway wide open, where the DOF may be an inch or less, then inaccurate AF will be a lot more noticeable than if you’re shooting stopped down to f8 with a wide angle lens.

Cheers,
Chris