[ebooks4pentax]

Dear Pentax users and friends:

Before I start, I would like to make sure that we don't all get into a dispute about the Charts presented here versus other charts available on the internet. Actually, my goal to to set this "auto-focus testing" controversy once and for all. With your constructive comments, I am confident that we'll get there. I will be trying to get Pentax's approval on the Charts presented here and will modify them to suit their recommendations if I receive any reply.

Autofocus Adjustment for the Pentax K20D, Custom Setting No 35, and how to check the Auto-focus accuracy

AUTOFOCUS ACCURACY, BACK & FRONT FOCUSING PROBLEMS

I personally never had any auto-focusing problem with any lens or camera I ever owned. It seems that these days, members of every blog and forum sites are talking about checking and adjusting the auto-focus on their lenses or DSLR. I think there is a perception that many lenses or DSLR cameras are having front or rear focus problems. In reality, I am inclined to think that novice photographers, “Pixels Peepers”, and inquisitive hobbyists, influenced by the available internet information, are finding problems where there is none. If a particular lens consistently gives out-of-focus results, when all other lenses used with the same DSLR are okay, you might have a focusing problem with that particular lens. If most of the pictures taken with the same DSLR, but with various lenses, are out-of-focus, you might have a DSLR camera-focusing problem.

If you think there is a focusing problem, several auto-focus testing charts are available from forums, blogsites and websites. I probably downloaded most of them and found inaccuracies with all of them. They are either too small, or too cluttered, or have measurements that are not to scale, etc. My own chart, published on my blog site earlier on, was too small and too cluttered for lenses with a minimum focusing distance of more than six to eight inches. One thing charts seem to have in common is that no one is ever sure of which exact point of the chart the camera is actually focusing on.

Back or front focusing problems are more notorious with subjects that are within a short depth of field, such as macro pictures or selective focusing pictures, and with the lens used at its widest aperture. The three images below illustrate this. The middle picture is the way it should be as the cat’s eye was the focusing target. The picture on the left shows a front- focusing problem, and the picture to the right shows a back-focusing problem.

(Photo1)

That alone would not mean that the lens or camera has a focusing problem, it could be the photographer’s error. However, similar results time after time could be the first hint that you might have a focusing problem. Perhaps a logical next step would be to test the lens in question under a controlled environment. A controlled environment could be the inside of a building where there is no wind, with the camera set on a tripod, and with good lighting. An auto-focus testing chart could be used, preferably one approved by the manufacturer if they have one available.

Of all of the DSLR cameras manufactured by Pentax, only the K20D has the option of adjusting the front / back focusing from within the camera (Custom Menu No.35). All auto-focus adjustments on other Pentax models should be made by a Pentax trained technician. Several forums members have published articles, explaining procedures to modify DSLR Firmware. Doing so could void the warranty and damage the camera. It is better to leave specialized work to specialists.


PROPOSED SOLUTION

I do not believe that Pentax has an AF checking chart available to the public. That is the primary reason I took it upon myself to design one. After reading every blog and information about the subject on the internet, after reading comments from the readers of my blog site and threads on various forums, I deducted that one chart could not do it all. Three charts were designed because the minimum focusing distance varies from lens to lens. The smaller chart (Chart–1) works well for close-up lenses and lenses that have macro capabilities. The medium sized chart, (Chart-2) works well for normal lenses, say 30mm to 100mm, which have a minimum focus distance needing a target a little bigger than Chart-1. The third chart (Chart-3) is for lenses that cannot focus very close. One could keep enlarging the last chart, but I believe that the three charts attached herein will be functional for the majority of lenses.


Chart No. 1


Chart No. 2


Chart No. 3


Download the Charts: http://www.k10dbook.com/newchart.pdf

There are various opinions whether a chart should be viewed at 45° from the lens centerline axis, or at 30°, or any angle or even flat. The reality is that it should give good results at any angle between 30° and 60°. Chart-1 and Chart-2 were design for viewing at 45° from the lens’ centerline axis. The measurements on the viewing surface were corrected for accurate reading when viewed at a 45° angle. The Charts could have been designed for viewing at 30°, giving more viewable depth of field for both front and back. However, I opted for the 45° because the charts could fit on a 8 ½” x 11” sheet of paper (Letter size).

I chose a focusing area shaped as a circle. The bottom half is black and the top half is white. When viewed perpendicularly or flat, it has an oval shape, because the chart is meant to be viewed at 45°. When viewed at 45°, the focusing area appears as a perfect circle. That also helps verifying that the lens is at a 45° angle..



Chart No. 2 viewed perpendicularly or flat. (Photo5)


Chart No. 2 viewed at 45 degree angle (Photo6)


CAMERA SETTINGS

Set the camera to:

* Autofocus single.
* Leave the Shake Reduction off if you use a tripod (recommended).
* Use the Aperture Priority mode (Av).
* Set the lens aperture to its maximum.
* Set the AF point to center.
* Use a remote for the shutter release or use the 2-second timer to avoid any movement.



PART-2 of this subject will be posted in the next few days. It will consist of "How to perform the test", "Understanding the results", "A word about auto-focus sensors", "The K20D custom setting No. 35 explanation".

There may be a PART-3 consisting of inputs received by Pentax and you, the readers.

Again, thank you very much for reading my blog.

Regards,


Yvon Bourque


P.S. I couldn't put all of the photos in this thread as there are a maximum of five allowed. The rest are on my blogsite.

Attached Images

         

---

Last edited by ebooks4pentax; 06-27-2008 at 12:28 AM.

[hessenwulf]

A good start watering my mouth - can't wait to see the real beef coming.
Thanks for your efforts!

[christinelandon]

Thank you, Yvon. I've not noticed any FF/BF problems with my equipment, but appreciate any concrete efforts in the community to help others who worry over the results they obtain with theirs.

[coloseu]

Hello,

Thank you very much for your thread, the time and the concerning in helping us all.
Nevertheless, from diferent forums consultation, I got the impression that due to charts inacuracy to give consisten results, the easier way to check for front/back issues is by using 3 AAA batteries, separated by 1cm from each other in a diagonal line and one should focus in the middle batterie.
Maybe I´m just wrong. I have always been afraid of charts because I never knew how to use them right for testing the lens.

Thank you

[ckanthon]

Dear Pentax users and friends:

Before I start, I would like to make sure that we don't all get into a dispute about the Charts presented here versus other charts available on the internet. Actually, my goal to to set this "auto-focus testing" controversy once and for all. With your constructive comments, I am confident that we'll get there. I will be trying to get Pentax's approval on the Charts presented here and will modify them to suit their recommendations if I receive any reply.

Autofocus Adjustment for the Pentax K20D, Custom Setting No 35, and how to check the Auto-focus accuracy

AUTOFOCUS ACCURACY, BACK & FRONT FOCUSING PROBLEMS

I personally never had any auto-focusing problem with any lens or camera I ever owned. It seems that these days, members of every blog and forum sites are talking about checking and adjusting the auto-focus on their lenses or DSLR. I think there is a perception that many lenses or DSLR cameras are having front or rear focus problems. In reality, I am inclined to think that novice photographers, “Pixels Peepers”, and inquisitive hobbyists, influenced by the available internet information, are finding problems where there is none. If a particular lens consistently gives out-of-focus results, when all other lenses used with the same DSLR are okay, you might have a focusing problem with that particular lens. If most of the pictures taken with the same DSLR, but with various lenses, are out-of-focus, you might have a DSLR camera-focusing problem.

If you think there is a focusing problem, several auto-focus testing charts are available from forums, blogsites and websites. I probably downloaded most of them and found inaccuracies with all of them. They are either too small, or too cluttered, or have measurements that are not to scale, etc. My own chart, published on my blog site earlier on, was too small and too cluttered for lenses with a minimum focusing distance of more than six to eight inches. One thing charts seem to have in common is that no one is ever sure of which exact point of the chart the camera is actually focusing on.

Back or front focusing problems are more notorious with subjects that are within a short depth of field, such as macro pictures or selective focusing pictures, and with the lens used at its widest aperture. The three images below illustrate this. The middle picture is the way it should be as the cat’s eye was the focusing target. The picture on the left shows a front- focusing problem, and the picture to the right shows a back-focusing problem.

(Photo1)

That alone would not mean that the lens or camera has a focusing problem, it could be the photographer’s error. However, similar results time after time could be the first hint that you might have a focusing problem. Perhaps a logical next step would be to test the lens in question under a controlled environment. A controlled environment could be the inside of a building where there is no wind, with the camera set on a tripod, and with good lighting. An auto-focus testing chart could be used, preferably one approved by the manufacturer if they have one available.

Of all of the DSLR cameras manufactured by Pentax, only the K20D has the option of adjusting the front / back focusing from within the camera (Custom Menu No.35). All auto-focus adjustments on other Pentax models should be made by a Pentax trained technician. Several forums members have published articles, explaining procedures to modify DSLR Firmware. Doing so could void the warranty and damage the camera. It is better to leave specialized work to specialists.


PROPOSED SOLUTION

I do not believe that Pentax has an AF checking chart available to the public. That is the primary reason I took it upon myself to design one. After reading every blog and information about the subject on the internet, after reading comments from the readers of my blog site and threads on various forums, I deducted that one chart could not do it all. Three charts were designed because the minimum focusing distance varies from lens to lens. The smaller chart (Chart–1) works well for close-up lenses and lenses that have macro capabilities. The medium sized chart, (Chart-2) works well for normal lenses, say 30mm to 100mm, which have a minimum focus distance needing a target a little bigger than Chart-1. The third chart (Chart-3) is for lenses that cannot focus very close. One could keep enlarging the last chart, but I believe that the three charts attached herein will be functional for the majority of lenses.


Chart No. 1


Chart No. 2


Chart No. 3


Download the Charts: http://www.k10dbook.com/newchart.pdf
......

PART-2 of this subject will be posted in the next few days. It will consist of "How to perform the test", "Understanding the results", "A word about auto-focus sensors", "The K20D custom setting No. 35 explanation".

There may be a PART-3 consisting of inputs received by Pentax and you, the readers.

Again, thank you very much for reading my blog.

Regards,


Yvon Bourque


P.S. I couldn't put all of the photos in this thread as there are a maximum of five allowed. The rest are on my blogsite.

That provide people three scales to check the three ranges (macro, near fields and application in longer distances). Sound very good. No need to worry. In my opinon, it is rather scientific and could provide accurate, repeatable, reproducible and reliable results if the sensor area is well covered. I saw a vertical object image plane on the N forum. But it is not always necessary if a single-critical contrast area is covered and hit by the focus sensor.

= Good job
= Minimized Uncertainty
= "PUSH" (Support)

[falconeye]

I would like to make sure that we don't all get into a dispute about the Charts presented here versus other charts available on the internet.

Hi Yvon,

as one of the persons who criticized your previous work let me say now that I see no (serious) flaw with your current work. Well done!

I have a question: In the PDF version, there is a dotted zick-zag line within the black half of the oval. Is this on purpose?

If not, it could lead to wrong focussing, although I think it shouldn't with a proper functioning ÁF system. Did you make tests?

If this is on purpose and you tested it (in order to have something near the center to validate focussing) than this is a great invention from your part.

May I make a suggestion pushing this idea?

How about a dense point pattern of 10% white in the black and 95% white in the white half, rather than the zick-zag line? If you tested that the pattern causes no errors, this could be even more useful leading to increased accuracy.

For most printers, just using non 100% colors for black and white will create a dense point pattern anyway. At the point of accurate focus one should be able to see the printer dot pattern then.

In my own tests, I was able to see the paper's grain structure to verify focus. However, this won't work anymore in the black half.

[pentkon52]

My BF problems were noted in real life, field shooting. The tests just confirmed it.

Sending the camera in alone, proved fruitless, however sending in the camera and one of the lenses that was doing the BF proved correct.

Camera and lens is back with me know, and all looks good. Focus point is were i want it, not 4-6 mm behind it.

It is a problem and seems fixable.

Nikon D1 had the same issues(I know as i have one) and my D200 also has these issues and is back getting looked at.

Dave

[brothereye]

This is an excellent write-up, although like you, I've never noticed any problem with my equipment either. I also, like you, believe that much of the problem has been made larger than it truly is by the "web effect". I'm not saying that there isn't any problem at all ever, but I think the incidents of problems are actually fewer than you might be led to believe by some writers on forums. Unforunately if there is a problem that affects 1% of users, my being in the 99% that have no problems doesn't help them any. Thoughtful write-ups like this are helpful though.

[ckanthon]

Focus senor cover a small circle area (!?)
Hei Guys.

In fact several days ago I had read a Forum concerning how those A350's user check the possiblility of FF/BF presented on their system. Noticed their findings during their investigation:
The Focus sensor could cover a small circle area (attached picture from them.)

Original link (simplified Chinese):
h**p://forum.xitek.com/sorthread.php?threadid=523169

[ebooks4pentax]

Hi Yvon,

as one of the persons who criticized your previous work let me say now that I see no (serious) flaw with your current work. Well done!

I have a question: In the PDF version, there is a dotted zick-zag line within the black half of the oval. Is this on purpose?

If not, it could lead to wrong focussing, although I think it shouldn't with a proper functioning ÁF system. Did you make tests?

If this is on purpose and you tested it (in order to have something near the center to validate focussing) than this is a great invention from your part.

May I make a suggestion pushing this idea?

How about a dense point pattern of 10% white in the black and 95% white in the white half, rather than the zick-zag line? If you tested that the pattern causes no errors, this could be even more useful leading to increased accuracy.

For most printers, just using non 100% colors for black and white will create a dense point pattern anyway. At the point of accurate focus one should be able to see the printer dot pattern then.

In my own tests, I was able to see the paper's grain structure to verify focus. However, this won't work anymore in the black half.

The zig-zag pattern was the result of converting the AutoCAD Drawing file to PDF. However, on my tests, when printed, the zig-zad disapeared. Were you still having the zig-zag lines when printed?

I appreaciate your comments and will wait for additional comments to react. I wrote to Pentax and received an answer to my email. They are testing and reviewing the chart and will let me know their opinion, comments, approval...actually I'll have to see what their answer is. I will post it here when received. They were actually very appreciative for the charts and very nice.

[ebooks4pentax]

Focus senor cover a small circle area (!?)
Hei Guys.

In fact several days ago I had read a Forum concerning how those A350's user check the possiblility of FF/BF presented on their system. Noticed their findings during their investigation:
The Focus sensor could cover a small circle area (attached picture from them.)

Original link (simplified Chinese):
h**p://forum.xitek.com/sorthread.php?threadid=523169

If this is the case with the Pentax sensors, since we perform the test with the center sensor which is a cross type sensor, the area of focus is smaller than the surrounding sensors and should make our test chart work even better.

I'm surprise to see only one cross sensor on the A350, or at least that's what the image you provided indicates.

[ckanthon]

This is an excellent write-up, although like you, I've never noticed any problem with my equipment either. I also, like you, believe that much of the problem has been made larger than it truly is by the "web effect". I'm not saying that there isn't any problem at all ever, but I think the incidents of problems are actually fewer than you might be led to believe by some writers on forums. Unforunately if there is a problem that affects 1% of users, my being in the 99% that have no problems doesn't help them any. Thoughtful write-ups like this are helpful though.

I don't know how your "1%" and "99%" came from. But I can give you one example:
My last job was working in R&D department of an electronic factory in China. we did produce customer goods for the shipment to Japan as well. One day the Japanese Manager flighted to our factory and had a hot talk in the resturant with our department managers. He required us to achieve "Zero-defect". So "1%" could be a large number in this regard

[konraDarnok]

In reality, I am inclined to think that novice photographers, “Pixels Peepers”, and inquisitive hobbyists, influenced by the available internet information, are finding problems where there is none.

I'm inclined to trust my eyes. Full pixel crops:

00um factory k10d setting


+60um Debug setting




If people *see* focus problems, they probably have one. Sorry, I kinda hate the ubiquitous talk of how 'hobbyists' don't know what they're doing. Like being a 'hobbyist' is the same thing as being 'blind'.

I, and others, have backfocus issues with their K10D. The Debug setting fixed my issue just fine -- with all lenses (except a teleconverter problem I'm having)


Honestly I think these issues have probably existed for years -- but the vast majority of photos were (and are) printed to 4x6.

In the digital world it's nice to crop a photo or otherwise look at it in a way that wasn't possible before (like fake macros) -- thus focus issues are much more pronounced.

[ckanthon]

Hei,

Good Job on your own fine-tune !! Enjoy!!.
I am regret of buying K200d (although it is my baby).
If I knew the softness of K10d only due to focus/third party lenses quality/large aperatures etc, I would have considered to buy K10d because K10d can do this....

I sent my baby out to hospital (repair centre) and waiting at home anxiously!!!

[brothereye]

I don't know how your "1%" and "99%" came from. But I can give you one example:
My last job was working in R&D department of an electronic factory in China. we did produce customer goods for the shipment to Japan as well. One day the Japanese Manager flighted to our factory and had a hot talk in the resturant with our department managers. He required us to achieve "Zero-defect". So "1%" could be a large number in this regard

2+2=5 for very large values of 2

Really, I don't have any hard numbers for 1% and 99%. It was mostly my recognition that my being in the group of people that don't have defective equipment doesn't really help the small percentage of users that are affected by a defect. And it's hard to convince many people that most manufacturers do occasionally have an item that is defective get shipped, even if most other specimens of the item are NOT defective.

The problem I see with zero-defect is that I would think it tends to make whatever the product is more expensive to produce. I think manufacturing any sufficiently complex product is more difficult than it probably seems from the outside, and unfortunately for most goods there is a certain level of defects, small though it may be. And most manufacturers do have small failure rates, if they had a failure rate that was large, I would think it would be difficult to stay in the manufacturing business for long.