20D Frontfocus, need advice

[Monique]

Hello,

I have a Canon 20D, very soon i noticed that the picture where very soft.
I did some testing with my 50mm 1.8 and 70-200 2.8L IS, seems the problem is frontfocus. I contacted the dealer after one day and he adviced me to do the update from the firmware. As i did and still no sucses, totaly out of focus. When i contacted the dealer again and asked him for replacement, he refused because i did the update I contacted Canon, and they adviced me to bring the camera in to the Canon Service Centre, as i did a week ago. Some people i know also have the same problem, same dealer. They also bring the camera to the Service Centre and the say that the camera is ok. But they still have the frontfocus problems.
I'm so affraid that they tell me the same. The camera is frontfocusing and i know how to handle dsrl, before i had the 10D, never problems with the focus.
Are there people over here who have also this problem?

Monique

[JZaun]

Monique
Go to the link below. Read and then do the test. It is one of the best ways to know if the cam is in specs or not. I also thought I had a focus issue with my 20 but the Bob Adkins test proved to me otherwise. My problems were just me :)

http://www.photo.net/learn/focustest/

JZ

[Monique]

Jerry i have done that, and this
http://www.pbase.com/zippo/image/34935219
http://www.pbase.com/zippo/image/34913950
All on tripod.

[KennyG]

I look at a lot of these focus tests and most of them are flawed. The main flaw is that the object that is the point of focus should be in exactly the same plane as the camera sensor, not at an angle. It should also be larger than the AF sensor point area which does extend quite a way outside the box in the viewfinder.

Canon's own test has a flat target and a 'ruler' at one side, angled upwards. To re-create this is quite simple. Use a cereal box and tape a target (black X printed on a white sheet of paper) on the large face and, using whatever method you want, attach a ruler to the side of the box angled with the lower edge towards you. The 'hinge' point of the ruler should be in line with the centre of the X.

Do your tests with bright daylight, not house lighting. This method is close enough to the one Canon use to get a more accurate test. In the early days of the 10D AF 'problem' (I had two 10D and never had the problem) Canon managed to leave images of their tests on some CF cards when the cameras were returned. I'll try to find a link to one of the images of the set up, but the above does work and is far more accurate than an angled target which has very little free space around the focus point.

[Monique]

Ken,

I do not have the camera here, he is in the Service Centre since one week.
I'm just sick of all these tests, all give the same, flawed en frontfocused.
Tomorrow we go the the Centre with all our lenses to do the final test.
I am not the only person with this problem, i hope Canon will admit that there is a serious problem with some 20D's :x

[amorrison]

Sorry you are having the problem Monique. I would love to see a micro focus adjustment parameter in the firmware so you could adjust the focus point slightly. That might save a bunch of returns to the service department. I wonder what they do to actually move the focus point?

[JZaun]

Focus is not an across the board problem with the 20. I think yours is an exception.

Here is a link to test of my 20.. Its spot on.:?

100 mm http://www.bytegallery.com/data/3191/51100mm-macro.jpg

50mm http://www.bytegallery.com/data/3191/5150mm.jpg

24-85mm http://www.bytegallery.com/data/3191/5124-85mm-_-85mm.jpg

JZ

[phili1]

The chart you were given in the link is not the same chart that you used. Your chart looks like its layed on a 45 % angle from the floor up , because of paralex would throw off the test, like a falling building and give you that results. His test was on a vertical plane on a 45% angle from lens which would show true front to back focus. Download his chart and do it his way. I have a 20 D and thought I had a problem with it and it was the lens I was using. My pictures are tack sharp

[FlipsidE]

Pardon my beginner question, but what is "frontfocus?"

Thanks

FlipsidE

[BoySpot]

Am I being an idiot or is Moniques test chart front focussed while the statue looks like the face is in focus, i.e. behind. This may (I only say may because I wasn't there) suggest the control of the focus point may be an issue.

[drisley]

Am I being an idiot or is Moniques test chart front focussed while the statue looks like the face is in focus, i.e. behind. This may (I only say may because I wasn't there) suggest the control of the focus point may be an issue.
BoySpot, I noticed the same thing. The test looks frontfocused, but the statue looks backfocused.
I feel your pain Monique, but I can happily report that my 20D has been spot-on with perfect Autofocus.
I dont think there is a focus problem with the 20D (actually I'm sure of it).
However, there are always going to be some cameras with problems.
By the looks of things you may have a problem, so hopefully Canon will fix it. Good luck

[phili1]

Guys look at the angle of his chart. His chart is horr. but his lens is not level to the chart, just like when you take a picture of a building and tilt the lens up the building is falling over. Well the same thing will happen with focus.

Jerry gave him the chart and the directions have it vertical on plane with the lens 45% to the lens front to back.

Look at it this way if you take a macro shot of a fly and the head is at the top of the picture and you focus on the head the tail will be out of focus.

[Red Squirrel]

If the test posted by JZaun ( http://www.photo.net/learn/focustest/ ) is flawed why is he getting consistent results showing his camera to be correctly focussed?

I have tried the exact same test and it confirmed my suspicions that my 10d was front focussing. I tried it with the following lenses

17-40L ( 7-8mm front focussing )
24-85 ( 15mm front focussing )
tamron 24-75 f2.8 ( 10mm front focussing )

I will be sending the camera to a canon service centre for calibration.
The question is though, __does anyone know exactly what Canon do__??
I wonder if it's possible to adjust this myself??

[phili1]

He is not getting the same results.

Bring his chart up and look at the coverging lines, it shows you that his camera is tilted. it fills the paper at the bottom but bows in at the top, that means his camera is not level to the chart and will give you that focus problem.

His chart shows front focusing.

Now look at the statue,he focused on the face and the wings are sharp, in that one it shows back focusing.

When yo do tests like this you have to do it properly or you will get tainted results. I am not saying he does not have a problem but if he uses the system Jerry posted he will get the best test results.

His test show improper set up as far as I am concerened.

Now I will hit you with someting else. the 20 D has nine point auto focus. If you use this the Camera selects the focus point and in my test it focus on the near subject first, not nec. the one you want. To do this test you have to use the center focus point. I never use the nine point, I select my own. When I use the nine point I get out of focus shots.

[evilenglishman]

I think Monique is a lady :D

[Red Squirrel]

What makes you say that? :wink:

[evilenglishman]

two reasons;
1. monique is a female name
2. from her site:



...When i got married with my sweet husband Luc...

[jboyd]

I will be sending the camera to a canon service centre for calibration.
The question is though, __does anyone know exactly what Canon do__??
I wonder if it's possible to adjust this myself??

Honestly, I don't think they do anything. Took almost 2 weeks to get my 10D back from them, and I noticed no difference.

[amorrison]

As info, here is Canon's official verbage on digital SLR focusing issues - from a FAQ on their website:

begin quote...

"AF (Auto Focus), focusing subjects automatically, is a convenient function with certain cameras. However, as a limitation of AF accuracy, the best focus point sometimes may shift forward and back a small amount. The AF accuracy of the Canon digital SLR camera is judged with the same viewing prerequisite as regulated in film cameras: "AF accuracy is judged when print out the whole image and view it from a normal viewing distance."

When using digital cameras, it is easy to magnify or print out just a part of the image once downloaded to your computer. This can magnify small errors in AF that would otherwise not be apparent when being viewed at a more traditional print size.

For example, imagine observing an image taken with the EOS10D (6,000,000 pixels) on the 17 inch SXGA CRT monitor (1280x1024) from a distance of 16-20 inches (40-50 cm) away. Under these conditions, the size of the image displayed on the monitor is equivalent to approximately a 32 in. x 21 in. (82cm x 54cm) print size, viewed at a distance of 16-20 in. from you. This is quite a different circumstance from film. In this case you are seeing slight errors in AF magnified many times over when they would not normally be visible when the image is printed. Even if small errors in AF are visible on the computer monitor, these errors may not be visible when the image is viewed as a full size print."

... end of quote.

[Cadwell]

that sounds like complete rubbish to me

:lol:

No, no... it's quite clear. The camera isn't getting it wrong, it's just our expectations that are too high. Silly us ;)

:lol:

[defordphoto]

Actually, it is true.

Until digital came upon us, we (well, 99% of us anyway) never magnified our photos 500% or more to check focus.

And AF only can do so much, however, on the other side of the coin, if the problem is consistent, or can be repeated consistently, then there is either a weakness in the AF system or it is truly defective.

Most of these focus tests are also bogus. You cannot test your camera by shooting a bunch of lines on a piece of paper. The lines should be used as a guide only, and a high contrast object placed on or near the paper for the camera to actually focus on. The camera will be easily confused trying to AF on a bunch of lines.

[JZaun]

Actually, it is true.


Most of these focus tests are also bogus. You cannot test your camera by shooting a bunch of lines on a piece of paper. The lines should be used as a guide only, and a high contrast object placed on or near the paper for the camera to actually focus on. The camera will be easily confused trying to AF on a bunch of lines.

If the bob Atkins test is done as per his instructions say. You are only focusing on (1) high contrast line at an exact 45 degree angle. The lines above and below are for reference where the lens is in focus. I think his test is very accurate if used correctly.

Just my opiniion

JZ

[CyberDyneSystems]

Monique,

I agree your Camera is front focusing :(

If the service center can not fix it.. demand another Camera from your retailer.

If they will not replace it.. then I am not at all sure what recourse you have,.. but it seems to me they must have some replacement policy.. how long has it been since you got the camera? Why on earth wouldn't a retailer offer a replacement to a customer who paid a few thousand for a product that may be defective?

[Monique]

Monique,

I agree your Camera is front focusing :(

If the service center can not fix it.. demand another Camera from your retailer.

If they will not replace it.. then I am not at all sure what recourse you have,.. but it seems to me they must have some replacement policy.. how long has it been since you got the camera? Why on earth wouldn't a retailer offer a replacement to a customer who paid a few thousand for a product that may be defective?

I have the camera 2 weeks and is now since one week in repair.
The retailer refused, and i'm not alone who bought one in that store, having the same problems , no replacement. They have bring the camera in for repair too, and after the repair they told that the camera whas ok.
Its seems so that they will not will admit there is something wrong with some 20D's, Canon has to do something about that , find a solution very fast, this is not good for the name of Canon, and lots of people will be glad to here that the camera they own is not good, so it can be replaced by Canon. It is so frustrating, please Canon, do something!

[CyberDyneSystems]

Could you tell us hwere you are located?

It is my feeling that the responsibility lies with the Retailer to replace the camera during there return period...

Canon may have rules about when they will replace a camera when the time period falls within a retailers return policy?

Two weeks.. really sounds like a Retailer issue.

If the retailer refuses to replace the camera, they should then at least refund your money.

Again,. I am curious where you are located and who the retailer is,. as in the US I do not think it is legal to refuse to refund or replace a malfuntioning product.

Also.. I agree with others that the Statue photo does not show front focus.. so your issue does not look consistant :?

I,.. think most of us here will agree that any focus issues that any single 20D may exhibit is not indicative of a "trend" like the one that manifested with the 10D early in it's release.

If you have a 20D that is defective,. (this is certainly possible) I do not think this neccesarily means that there is an inherent problem. To the contrary it seems that with the 20D launch we have seen very few of these issues beyond the "lock up"

[evilenglishman]

a post of mine seems to have disappeared from this thread?????????

[JZaun]

a post of mine seems to have disappeared from this thread?????????

Yep it did...?

JZ

[phili1]

Jerry I agree with you, you have to do the test properly. I can not figure out is why some who know photography can not figure it out.

If you hold the chart in the photo horr and it was in plane with the lens you could not see the chart. When you tilt it to see it you create front or rear focus.

It has to be in the vertical plane with the right side pushed back 45% to get a true result.

In the sample photos you can see they contridict themselves.

I had a problem with my Sigma lenses, sent them back to Sigma, they were rechipped and now they focus properly. When I di a firmware update it furthe coreected the problem.

There is a communication problem with certain lenses and firmware 11.0.5 fixed it for me.

Do Jerry's test and then if there is a problem you can scream and yell.

[Andy_T]

a post of mine seems to have disappeared from this thread?????????

Was there any particularly offensive content ?

Best regards,
Andy

[Steven M. Anthony]

The test outlines by Bob Atkins (included in a post way above) is great--BUT IT MUST BE DONE ABSOLUTELY CORRECTLY TO WORK--I.E., GIVE YOU ACCURATE RESULTS THAT LEAD TO A CORRECT CONCLUSION ABOUT YOUR LENS.

I did a quick set-up in my office friday afternoon--probably came close to the "perfect" setup-and the tests showed my 10D was front-focusing.

Needless to say, I was upset.

So earlier today, I set up the test with more accuracy--to ENSURE I had the axis of my lens at 45 degrees to the test page.

THIS TEST IS EXTREMELY SENSITIVE. Prior to my exact measurment of the 45 degree angle I eye-balled it several times and the lens that front-focused on friday was now back-focusing! So I finally got out a protractor (the 1 benefit of having a teenager in the same house!) and a piece of string. Then I lined up the bottom and center AF points on the string to ensure straight-line orientation to the 45 degree line and eased the camera so the center AF point was on the target line on the test page.

Repeated exposure of several lenses using this method showed near central focusing.

This is not to say that ALL 10D/20D cameras have accurate focusing. But even if you set up the test and you are only off a couple of degrees in the set-up, you are not going to get a true picture of front/back focusing.

[phili1]

Ok Steven just what myself and Jerry have been saying. I do not know why people do these test without the proper set up it confuses everything.

[evilenglishman]

a post of mine seems to have disappeared from this thread?????????

Was there any particularly offensive content ?

Best regards,
Andy

no. it was relating to the fact that magazines have been blowing up slides over 500% for years.

[Steven M. Anthony]

There are a couple of threads with this topic--maybe you posted it on one of those...?

[Steven M. Anthony]

phili1--I think people just don't realize how precise it needs to be to work.

[phili1]

I agree. It upsets me when conclusions are jump to. I guess that what makes life go around.

[Monique]

I agree. It upsets me when conclusions are jump to. I guess that what makes life go around.

I asume you have a good 20D, lets say you have on that not is in focus, and you did all the tests precise, you send it back to the Service Centre, with your lensens, and after the test are completed there still is the problem. Keep in mind that Canon has told you all is ok, no problem whas found, what would you say by each photo you take in open field, all is out of focus, what would you say....

Well, thats what makes the life go arround...they say....

[Steven M. Anthony]

did canon show you their test shots? If not, I would ask to see them.

I will go back to see if you linked any shots you took so I can look at them.

One other thought is that it's possible that something to the edge of the photo (not in the center) is getting "picked" by the camera as the AF focus spot. You can check this with some of the software that comes with the camera (i think). (I haven't read through the entire thread here--so if someone already suggested that and you checked it already, sorry for repeating!)

[phili1]

If you read what I said about your situation earlier is that you might just have a bad one. Now I know a liitle about photography and when I look at your test shots the way you did it is not the right way, so I question the results. What I was saying is that allot of focus problems that some people have compalined about is not the proper use of the Camera, mistakes like holding at a slow shutter speed.

One factor in focusing is the nine point system is not perfect for photographic shots all the time, it is a specialized focus system for only certain situations. If you use it you might get your subject matter out of focus because it has a tendency to focus on near subject.

If you want to test your lens then do it the way Jerry has outlined it is the proper way. Your test print was on a horr line to the lens and tilted and you will get either front focus or rear focus that way . Your shot 1 showed front focus but the other showed back focus.

In all this you did not say what lens you used. I had a problem with my Sigma, every picture was blurred and when I checked it out I found that the 20 D has a lens incompatabilty problem with Sigmas and some Canon lenses. Sigma rechipped my lens and it works fine now. When I updated the firmware it coreected the rest.

I am not doubting your problem but your test methods, if you do the test properly and show it to Canon then they have to say yes you are right.

The conclusion jumping was not meant for you, you are trying to resolve the problem.

I do not know your photographic experience but when I have a problem which I did I sit an evaluate what I did to see if it was me, because I make mistakes and then I look at the picture to see if it is motion blurred, if I see it blurred all over then I know I moved. An example I took a shot of a flower with my Tamron 9 inces away, when I got home it was soft at 1/125 of a sec. When I evaluated what I did I realised that I had move a fraction changing the focus, on close up work it is fractions of an inch.

All I am saying is go thru an evaluation of everything you are doing and eliminate all human possabilities, do Jerry's test and if you still have a problem I will contact Canon for you for you to get satisfaction. If a bunch of us complain they have to do something.