Lens calibration - Focus Tune software
Posted 29/04/2013 - 18:01
Link
JAK wrote:
Just to go back to that software solution. If it doesn't have a dedicated Pentax output, it would still need a trial and error method to determine what actual adjustment is needed unless I'm missing something.
You are [missing something].Just to go back to that software solution. If it doesn't have a dedicated Pentax output, it would still need a trial and error method to determine what actual adjustment is needed unless I'm missing something.
All the software needs is 4 shots at each possible adjustment setting.
The software then wades through them sequentially and determines how sharp/unsharp each one is.
The net result is a graph showing the sharpness curve. You then pick the adjustment setting that is at the peak of the curve, apply it to your camera, and job done.
This seems to me to be a very technical approach to focus calibration, and one that takes us users out of the decision process. All we have to do is take the 84 images necessary for each lens by following the included instructions.
Peter E Smith - flickr Photostream
Posted 29/04/2013 - 18:15
Link
Mannesty wrote:
All the software needs is 4 shots at each possible adjustment setting.
The software then wades through them sequentially and determines how sharp/unsharp each one is.
The net result is a graph showing the sharpness curve. You then pick the adjustment setting that is at the peak of the curve, apply it to your camera, and job done.
This seems to me to be a very technical approach to focus calibration, and one that takes us users out of the decision process. All we have to do is take the 84 images necessary for each lens by following the included instructions.
How would you check the focus at infinity?JAK wrote:
Just to go back to that software solution. If it doesn't have a dedicated Pentax output, it would still need a trial and error method to determine what actual adjustment is needed unless I'm missing something.
You are [missing something].Just to go back to that software solution. If it doesn't have a dedicated Pentax output, it would still need a trial and error method to determine what actual adjustment is needed unless I'm missing something.
All the software needs is 4 shots at each possible adjustment setting.
The software then wades through them sequentially and determines how sharp/unsharp each one is.
The net result is a graph showing the sharpness curve. You then pick the adjustment setting that is at the peak of the curve, apply it to your camera, and job done.
This seems to me to be a very technical approach to focus calibration, and one that takes us users out of the decision process. All we have to do is take the 84 images necessary for each lens by following the included instructions.
84 images? Good job it isn't slide film! But all you need is two or three for the job in hand:
1. to determine if there is a front/back focussing issue in the first place.
2. adjust the AF setting to + or - 5 as appropriate to see how close it makes it.
3. having got close, add or take a bit and you should be close enough by now.
BUT - but only for that image in that lighting at that distance. Any other and it could all need doing again. It's highly likely that the adjustment needed at a couple of feet is quite different to that needed at infinity, so only use adjustment when it is necessary.
John K
John K
Posted 29/04/2013 - 18:19
Link
Mannesty wrote:
All the software needs is 4 shots at each possible adjustment setting.
The software then wades through them sequentially and determines how sharp/unsharp each one is.
The net result is a graph showing the sharpness curve. You then pick the adjustment setting that is at the peak of the curve, apply it to your camera, and job done.
This seems to me to be a very technical approach to focus calibration, and one that takes us users out of the decision process. All we have to do is take the 84 images necessary for each lens by following the included instructions.
There's no need to take 84 images ( including uploading them to computer, renaming them, etc,etc ). It's obvious that most settings are going to be way out of focus.JAK wrote:
Just to go back to that software solution. If it doesn't have a dedicated Pentax output, it would still need a trial and error method to determine what actual adjustment is needed unless I'm missing something.
You are [missing something].Just to go back to that software solution. If it doesn't have a dedicated Pentax output, it would still need a trial and error method to determine what actual adjustment is needed unless I'm missing something.
All the software needs is 4 shots at each possible adjustment setting.
The software then wades through them sequentially and determines how sharp/unsharp each one is.
The net result is a graph showing the sharpness curve. You then pick the adjustment setting that is at the peak of the curve, apply it to your camera, and job done.
This seems to me to be a very technical approach to focus calibration, and one that takes us users out of the decision process. All we have to do is take the 84 images necessary for each lens by following the included instructions.
Lets say the lens is back focussing: Even if you use the successive approximation method, you'll take one at +10, if that font focusses, set it to +5, if that still front focusses set it to +2 or +3, or +7 or +8 if it's back focussing again.
IOW you can soon narrow it down with 3 or 4 shots ( viewed in camera - use the maginfier if you need to ) ) Then when you arrive at what looks like the correct adjustment take a few more shots to verify and account for variability.
[i]Bodies: 1x K-5IIs, 2x K-5, Sony TX-5, Nokia 808
Lenses: Pentax DA 10-17mm ED(IF) Fish Eye, Pentax DA 14mm f/2.8, Sigma 17-70mm f/2.8, Pentax-A 28mm f/2.8, Sigma 30mm F1.4 EX DC, Pentax-A 50mm f/1.2, Pentax-A 50mm f/1.4, Pentax-FA 50mm f/1.4, Pentax-A 50mm f/1.7, Pentax DA* 50-135mm f/2.8, Sigma 135-400mm APO DG, and more ..
Flash: AF-540FGZ, Vivitar 283
Lenses: Pentax DA 10-17mm ED(IF) Fish Eye, Pentax DA 14mm f/2.8, Sigma 17-70mm f/2.8, Pentax-A 28mm f/2.8, Sigma 30mm F1.4 EX DC, Pentax-A 50mm f/1.2, Pentax-A 50mm f/1.4, Pentax-FA 50mm f/1.4, Pentax-A 50mm f/1.7, Pentax DA* 50-135mm f/2.8, Sigma 135-400mm APO DG, and more ..
Flash: AF-540FGZ, Vivitar 283
Posted 29/04/2013 - 18:28
Link
I posted a link sometime back to a very simple way of
doing the AF ADJ on Ricehigh's site. I can't find it
at present, but anyone with half a brain could easily
figure it out for themselves
'
doing the AF ADJ on Ricehigh's site. I can't find it
at present, but anyone with half a brain could easily
figure it out for themselves
'
Half Man... Half Pentax ... Half Cucumber
Pentax K-1 + K-5 and some other stuff
Algi
Pentax K-1 + K-5 and some other stuff
Algi
Posted 29/04/2013 - 18:43
Link
Smeggypants wrote:
There's no need to take 84 images ( including uploading them to computer, renaming them, etc,etc ). It's obvious that most settings are going to be way out of focus.
I know that there was no need to take so many images, but I wanted to, partly influenced by the torrential rain we had yesterday and I had nothing much else to do.There's no need to take 84 images ( including uploading them to computer, renaming them, etc,etc ). It's obvious that most settings are going to be way out of focus.
I know all of the other methods used to adjust the AF focus and I've probably used all of them at one time or another. This 'different' method deserved a try out, so I did try it and considered that the more sample images it had to digest, the better.
There is absolutely no need to rename all of the images, all you need to make note of is the adjustment setting of image #40 (if that's the peak sharpness point in the curve). So in my case, images 1-4 where at -10, images 5-8 at -9 etc, etc.
Once I'd got into the swing of taking the images it took about 5 minutes to take the 84 images that I wanted to take. The lens that took the longest was my Sigma 150-500mm because my tripod proved to be not stiff enough to stop it vibrating when the mirror went up.
I imagine that the software works along similar lines to those used for stacking images where only the sharpest part of the image is chosen to be included in the final output.
Of course, once I've taken a few shots with these newly calibrated lenses I may well be in for generous helpings of humble pie if they turn out to be pants.
Peter E Smith - flickr Photostream
Posted 29/04/2013 - 18:45
Link
JAK wrote:
How would you check the focus at infinity?
Not possible since infinity does not exist How would you check the focus at infinity?
But photographically speaking you'd need a very large target to focus on and placed several miles from your camera, so probably not worth the trouble.
Peter E Smith - flickr Photostream
Posted 29/04/2013 - 18:54
Link
Mannesty wrote:
I know all of the other methods used to adjust the AF focus and I've probably used all of them at one time or another. This 'different' method deserved a try out, so I did try it and considered that the more sample images it had to digest, the better.
There is absolutely no need to rename all of the images, all you need to make note of is the adjustment setting of image #40 (if that's the peak sharpness point in the curve). So in my case, images 1-4 where at -10, images 5-8 at -9 etc, etc.
Once I'd got into the swing of taking the images it took about 5 minutes to take the 84 images that I wanted to take. The lens that took the longest was my Sigma 150-500mm because my tripod proved to be not stiff enough to stop it vibrating when the mirror went up.
I imagine that the software works along similar lines to those used for stacking images where only the sharpest part of the image is chosen to be included in the final output.
Of course, once I've taken a few shots with these newly calibrated lenses I may well be in for generous helpings of humble pie if they turn out to be pants.
I think that sums it up really. If you've really nothing else to do then use this software for adjusting the AF on your images.
Smeggypants wrote:
There's no need to take 84 images( including uploading them to computer, renaming them, etc,etc ). It's obvious that most settings are going to be way out of focus.
I know that there was no need to take so many images, but I wanted to, partly influenced by the torrential rain we had yesterday and I had nothing much else to do.There's no need to take 84 images( including uploading them to computer, renaming them, etc,etc ). It's obvious that most settings are going to be way out of focus.
I know all of the other methods used to adjust the AF focus and I've probably used all of them at one time or another. This 'different' method deserved a try out, so I did try it and considered that the more sample images it had to digest, the better.
There is absolutely no need to rename all of the images, all you need to make note of is the adjustment setting of image #40 (if that's the peak sharpness point in the curve). So in my case, images 1-4 where at -10, images 5-8 at -9 etc, etc.
Once I'd got into the swing of taking the images it took about 5 minutes to take the 84 images that I wanted to take. The lens that took the longest was my Sigma 150-500mm because my tripod proved to be not stiff enough to stop it vibrating when the mirror went up.
I imagine that the software works along similar lines to those used for stacking images where only the sharpest part of the image is chosen to be included in the final output.
Of course, once I've taken a few shots with these newly calibrated lenses I may well be in for generous helpings of humble pie if they turn out to be pants.
[i]Bodies: 1x K-5IIs, 2x K-5, Sony TX-5, Nokia 808
Lenses: Pentax DA 10-17mm ED(IF) Fish Eye, Pentax DA 14mm f/2.8, Sigma 17-70mm f/2.8, Pentax-A 28mm f/2.8, Sigma 30mm F1.4 EX DC, Pentax-A 50mm f/1.2, Pentax-A 50mm f/1.4, Pentax-FA 50mm f/1.4, Pentax-A 50mm f/1.7, Pentax DA* 50-135mm f/2.8, Sigma 135-400mm APO DG, and more ..
Flash: AF-540FGZ, Vivitar 283
Lenses: Pentax DA 10-17mm ED(IF) Fish Eye, Pentax DA 14mm f/2.8, Sigma 17-70mm f/2.8, Pentax-A 28mm f/2.8, Sigma 30mm F1.4 EX DC, Pentax-A 50mm f/1.2, Pentax-A 50mm f/1.4, Pentax-FA 50mm f/1.4, Pentax-A 50mm f/1.7, Pentax DA* 50-135mm f/2.8, Sigma 135-400mm APO DG, and more ..
Flash: AF-540FGZ, Vivitar 283
Posted 30/04/2013 - 11:52
Link
Smeggypants wrote:
The AF adjustment is nothing to do with the Image sensor
The AF sensor is mechanically adjustable via a screw(s) under the baseplate. however I think, and I'm not 100% sure on this, that the AF parameter in the custom menu doesn't physically move the AF sensor but rather applies a relative adjustment to the signal picked up by the sensor. This would make sense as the AF sensor is phase based and can, unlike the contrast method used by the image sensor, tell whether the image is front or back focussed.
I'll eat my humble pie!! I do remember reading or hearing that this was the way it worked but on doing some overnight research I bow to your (and DPM and kh1234567890) superior knowledge, and apologies for my brain fart, Patent US 8363154 B2 (BTW this was granted to Samsung) describes the function perfectly. The AF adjustment is nothing to do with the Image sensor
The AF sensor is mechanically adjustable via a screw(s) under the baseplate. however I think, and I'm not 100% sure on this, that the AF parameter in the custom menu doesn't physically move the AF sensor but rather applies a relative adjustment to the signal picked up by the sensor. This would make sense as the AF sensor is phase based and can, unlike the contrast method used by the image sensor, tell whether the image is front or back focussed.
Although I was completely wrong in the operational workings I was correct in saying that the adjustment is between the lens focal point and the sensor plane therefore IMO it does have something to do with the image sensor
Quote:
and if a software solution helps to give me confidence that I've input the correct adjustment I'll go for it over my eyes any day.
I'm not saying the software is the be all and end all of focus adjustment but it may be a useful aid to me, having one useless eye and one well below par and wearing varifocal glasses does nothing to help me see if and image is at it's sharpest. and if a software solution helps to give me confidence that I've input the correct adjustment I'll go for it over my eyes any day.
Quote:
Do you also use software to view the pictures you take?
See above.Do you also use software to view the pictures you take?
Quote:
Not at all. I can just see when a lens is front or back focussing. Like in the example JAK posted on page 2 of this thread.
I wasn't having a dig, I've always been jealous of folk with good eysight, you should see me trying to hit a tennis ball, comedy gold
Not at all. I can just see when a lens is front or back focussing. Like in the example JAK posted on page 2 of this thread.
Cheers
Brian.
LBA is good for you, a Lens a day helps you work, rest and play.
Brian.
LBA is good for you, a Lens a day helps you work, rest and play.
Posted 02/05/2013 - 17:48
Link
I've had a quick scan through this thread and nowhere have I seen mention of one of the biggest factors I've found contributing to erratic focussing on my K-5, namely whether focus is approached from the long end or the short end. On my K-5 the difference is about 4 steps of AF adjustment.
I find any shot-to-shot variance to be minor compared to this factor. Therefore my method is to check and adjust AF by taking just two shots, one approaching focus from the long end, the other approaching from the short end. Compare the two shots and when your target is equally just off focus you've found the nominally correct AF adjustment. This will result in being about 2 steps back focussed when approaching focus from the long end and about 2 steps front focussed when approaching from the short end. I find 2 steps to be the equivalent of about 1/3rd of the standard depth of field wide open. If you want improved auto focus accuracy you might want to always approach focus from the short end and calibrate for that. If you do this check that the lens doesn't hit the buffers when attempting to focus infinity! Personally I don't find it very practical to have to think about pre-focussing in front of my subject, so I go for the centre AF adjust and accept it's going to be about 2 steps out.
Another factor I've found is how the required AF adjust varies according to focus distance. I think this may vary between lenses, with the high range zooms being more susceptible. As an engineer I too can't understand why the phase difference system requires tweaking for each lens and why it should vary according to focus distance. It sounds like a weakness of the feedback system.
Whilst I agree that a firm tripod would be ideal I've not found any difference between the results I obtained on a tripod and hand held when testing a focal length of about 80mm at distances of about 15m and infinity. After all, you should be testing wide open in good lighting conditions so shutter speeds are going to be pretty high. However, I would advise using a tripod for consistent close up tests.
Finally, I think it's easy to become obsessed with focus accuracy. In the real world most of the things we photograph are not stationary flat plane subjects, they have depth and they move. If I had to choose between having a bit of front focus and a bit of back focus I think I'd go for the latter as I usually want depth of field behind my subject rather than in front of it.
P.S. Forget the fancy software. It's pretty pointless as if the AF adjustment varies with focus distance you still have to make a decision as to the distance you want to optimise for. This method is good enough to be able to see the effect of every AF adjustment step.
I find any shot-to-shot variance to be minor compared to this factor. Therefore my method is to check and adjust AF by taking just two shots, one approaching focus from the long end, the other approaching from the short end. Compare the two shots and when your target is equally just off focus you've found the nominally correct AF adjustment. This will result in being about 2 steps back focussed when approaching focus from the long end and about 2 steps front focussed when approaching from the short end. I find 2 steps to be the equivalent of about 1/3rd of the standard depth of field wide open. If you want improved auto focus accuracy you might want to always approach focus from the short end and calibrate for that. If you do this check that the lens doesn't hit the buffers when attempting to focus infinity! Personally I don't find it very practical to have to think about pre-focussing in front of my subject, so I go for the centre AF adjust and accept it's going to be about 2 steps out.
Another factor I've found is how the required AF adjust varies according to focus distance. I think this may vary between lenses, with the high range zooms being more susceptible. As an engineer I too can't understand why the phase difference system requires tweaking for each lens and why it should vary according to focus distance. It sounds like a weakness of the feedback system.
Whilst I agree that a firm tripod would be ideal I've not found any difference between the results I obtained on a tripod and hand held when testing a focal length of about 80mm at distances of about 15m and infinity. After all, you should be testing wide open in good lighting conditions so shutter speeds are going to be pretty high. However, I would advise using a tripod for consistent close up tests.
Finally, I think it's easy to become obsessed with focus accuracy. In the real world most of the things we photograph are not stationary flat plane subjects, they have depth and they move. If I had to choose between having a bit of front focus and a bit of back focus I think I'd go for the latter as I usually want depth of field behind my subject rather than in front of it.
P.S. Forget the fancy software. It's pretty pointless as if the AF adjustment varies with focus distance you still have to make a decision as to the distance you want to optimise for. This method is good enough to be able to see the effect of every AF adjustment step.
Posted 07/11/2013 - 10:56
Link
Retested the 100mm macro on the K-3 last night, that lens that was back focussing on the K-5, using the same cornflake packet with the word search centered on the middle O! This time the focusing was bang on; no need for any adjustment this time. Sorry if that isn't a very scientific test but it works for me.

Hope that might reassure others who weren't happy with the K-5's focussing issues.
John K

Hope that might reassure others who weren't happy with the K-5's focussing issues.
John K
John K
Posted 02/12/2013 - 12:46
Link
Interesting. I've still got focussing issues on my lenses with the K5-IIs.
The main problem I have is with zoom lenses because the required micro focus adjustment can vary with focal length. The worst example is the DA 18-250mm which needs about +10 to compensate for back focus at 250mm but needs a slight negative setting to compensate for front focus at 18mm. Is there any service adjustment that can be made to the lens for this, so the compensation required is much the same at both ends of the zoom range?
The other problem is how the auto focus varies according to whether it is approached from the near end or the far end. If approaching from the far end the focus obtained will be further back than if approaching from the near end. It's very easy to spot if you pixel peep. The difference is equaivalent to about 2/3rds of depth of field wide open. I was disappointed that this didn't seem to improve much with the K5-II. I assume this has been improved with the K3?
The main problem I have is with zoom lenses because the required micro focus adjustment can vary with focal length. The worst example is the DA 18-250mm which needs about +10 to compensate for back focus at 250mm but needs a slight negative setting to compensate for front focus at 18mm. Is there any service adjustment that can be made to the lens for this, so the compensation required is much the same at both ends of the zoom range?
The other problem is how the auto focus varies according to whether it is approached from the near end or the far end. If approaching from the far end the focus obtained will be further back than if approaching from the near end. It's very easy to spot if you pixel peep. The difference is equaivalent to about 2/3rds of depth of field wide open. I was disappointed that this didn't seem to improve much with the K5-II. I assume this has been improved with the K3?
Posted 02/12/2013 - 12:55
Link
Sounds like a focussing issue with that particular lens, so I doubt it. A K-3 may even make the problem seem worse if you pixel peep.
John K
John K
John K
Add Comment
To leave a comment - Log in to Pentax User or create a new account.


7671 posts
17 years
Smegland
Well we have an intelligent audience here, so they can make up their own minds I'm sure.
Lenses: Pentax DA 10-17mm ED(IF) Fish Eye, Pentax DA 14mm f/2.8, Sigma 17-70mm f/2.8, Pentax-A 28mm f/2.8, Sigma 30mm F1.4 EX DC, Pentax-A 50mm f/1.2, Pentax-A 50mm f/1.4, Pentax-FA 50mm f/1.4, Pentax-A 50mm f/1.7, Pentax DA* 50-135mm f/2.8, Sigma 135-400mm APO DG, and more ..
Flash: AF-540FGZ, Vivitar 283