Simple Go/No go focus test

[siwatkins]

Non-measurebators look away now :)

But if you still have concerns about focusing accuracy, are not sure if it's user error etc, are dubious about some of the ruler tests worth (as I am), then this method is very simple to use, requires no special equipment, and takes only a minute to do.

Very quick and easy way to check a number of aspects of your focus system repeatably and accurately using no special equipment:

Mount camera on tripod. Use remote release, or self-timer for taking the test shots. If you don't use a tripod, I think you should consider your results invalid.

Step 1:


Display something predominantly white, with text all over it on your monitor screen. Maximising your Outlook inbox screen works well.

Ensure your camera back is parallel and centred with respect to your screen. Focus on your computer monitor screen manually. Notice that at the point of exact focus, you see a Moire pattern in the viewfinder. (Squiggly interference lines).

Take a picture.

Step 2:


Move the lens out of focus, and this time use AF to onto the screen again, and you should see the moire pattern in your viewfinder.

Take the picture.

Step 3:

Review the images.

That's it.

Possible Problem Indicators:


If you can see the moire pattern in the viewfinder when manual focusing, but not when Autofocusing, or not in the resultant image on the CF card, that would indicate that there is some kind of focus alignment issue that might warrant further investigation. The moire is always strongest at the point of exact focus, so the AF and MF tests should appear very similar.


Additional test for advanced/obsessive measurebators :)

Instead of using Outlook for the above test, open up FVU and display one of your previous images with the FVU focus point overlay.

Align your camera at the correct distance and position, so that your in-finder focus boxes, overlay those displayed

Now when you review your images, you can also see if the focus boxes in your camera, correlate with those shown in FVU.

The reason I added this step, was that someone on this forum was concerned that their in-camera focus boxes appeared to disagree with those shown in FVU.

Quick, easy and repeatable check of your finder, sensor, and AF alignment.

Simon

[CyberDyneSystems]

For these steps to be repeatable,. we may need to have the sa,e monitor that you have, but at the very least we will need the same resolution and refresh rate.

[siwatkins]

CyberDyneSystems wrote:
For these steps to be repeatable,. we may need to have the sa,e monitor that you have, but at the very least we will need the same resolution and refresh rate.

Shouldn't make a lot/any difference - it's the dot pitch of the phospors that causes the moire. With certain shutter speeds it could be feasible that you catch the screen being redrawn, but the persistence of the phosphor is such that even that won't make any difference - certainly not in terms of what the test is trying to acheive. Most computer CRTs have a fairly fine dot pitch these days.

I take it from your comments you've not actually tried it yourself?

Simon

[justme_dc]

I have found that ruler tests, alignment charts and similar set ups show Consistant back focus issues with my 10D. I don't believe that these tests are all flawed as canon uses a version of the test for their in house evaluation. That leads me to believe the tests must have some merit. Perhaps I am doing something wrong but I doubt it. The reason I did these tests in the first place is because the back focus was visable in my everyday shots. I only ever use the center focus point so it isn't a focus point selection issue.

I have been putting off sending it into canon. I am not really sure why. I just have.

I'll try your test tonight and see what results I get.


A few questions if I may....

Have you ever seen a camera fail this test? In doing so proving the test valid for detecting a flaw?

Is it posible that the camera can't fail such a test?

What was your basis for principles for this test? i.e. is this a proven optical phenomenon?

I don't question you out of disbelief but a desire to understand the principles involved.

thanks
DC

[CyberDyneSystems]

I did try it,. I don't see any moire,. but I do get large horizontal bands of differing levels of light across the screen,. which would be caused by the shutter speed freezing the horizontal refresh rate....

Shouldn't make a lot/any difference - it's the dot pitch of the phospors that causes the moire.

.. but it could be the difference in Dot pitch from one monitor to another,... In fact I don't have a dot pitch.

It's an aperture grill which has a "Stripe Pitch" of .22... aperture grills use a totally different form of "dots" from a Shadow Mask CRTs "dots". Aperture Grills have no "Dots" at all,.


http://www.sonystyle.com/is-bin/INTERSHOP.enfinity/eCS/Store/en/-/USD/SY_DisplayProductInformation-Start;sid=JKQHpahqLiEHl5aLFboNrudkmN2nRfOPcbE=?Cat alogCategoryID=snAKC0%2eNk8IAAAD190RbXJwZ&ProductID=cpAKC0%2eNPgYAAAD10RBbXJwe&Dept=cpu

[spark]

Great method introduced ......

[siwatkins]

CyberDyneSystems wrote:

I did try it,. I don't see any moire,.

It's an aperture grill which has a "Stripe Pitch" of .22... aperture grills use a totally different form of "dots" from a Shadow Mask CRTs "dots". Aperture Grills have no "Dots" at all,.


Surprised you don't see any moire - it might be that your lens isn't resolving sufficient lpmm for the effect to occur, or that you are too close - try moving further back, or using a TV screen instead.

BTW, I'm using an Aperture Grill - there is still a very fine line pattern that will cause moire. As long as you are at the right distance, you should see it.

Simon

[siwatkins]

justme_dc wrote:
A few questions if I may....

Have you ever seen a camera fail this test? In doing so proving the test valid for detecting a flaw?

Is it posible that the camera can't fail such a test?

What was your basis for principles for this test? i.e. is this a proven optical phenomenon?

I don't question you out of disbelief but a desire to understand the principles involved.

thanks
DC



Well, should the focus screen be at a different path length from the rear element than the sensor, then one would expect MF and AF results to differ.

This effectively is just an advanced version of the "Newspaper" test that Canon France and Canon Germany suggested for determining if there is a problem. The issue with their tests, is that MF on a ground glass screen is very difficult to precisely acheive without some kind of focusing aid. So a comparison between AF and MF results is difficult, because the human has been responsible for determining the manual focus result.

By focusing on the CRT display in MF, the very fine make up of the screen causes a moire pattern at the point of exact focus that is very easy to see, and use as a focus aid. Needless to say, the AF system should be able to focus similarly well and if performing correctly, will show a similar level of moire.

It might need some experimentation because monitors do vary, some lenses will have lower resolving power in the first place (so won't cause the moire) and distance to screen may be important.

As to whether a camera with a focus issue will fail - theoretically yes - maximum moire only occurs at a very precise point in focusing terms, so a camera with a misaligned focus screen or sensor/calibration issue should show issues. Mine doesn't fail, but mine doesn't have a focus issue. Thus it would be nice to see someone who believes they do have a genuine issue try this, and see if it does fail and what the failure mode is.

Of course, there are some people who would be happier believing that their camera is faulty, and who would be reluctant to do a test that might prove their assertion was wrong - thus would prefer to use a flawed test in the first place. I don't doubt there are faulty cameras out there - too much smoke for there not to be any fire, and my first 10D exhibited problems that I could clearly show with a modified version of the Canon test.

On the Canon test, the many "ruler" type grids around do make a camera appear faulty - that's why Canon doesn't use such a method. Pekka's suggested method, which is basically a recreation of the Canon test is fine - it uses a plane parallel target, and moves the measuring grid off to one side. Those that are just using the sloping target will potentially see problems.

For example - I can make my current camera, using a sloping target vary from FF to BF just by altering by a few degrees the angle of the target, or by changing the lighting conditions, or by changing my distance to the target, or even by switching lenses! How many variables!?

By using a properly modified version of the Canon test (a la Pekka's example), it's 100% fine, with all lenses.

More importantly, it's fine in the real world too! :)

There seems to be a general reluctance, and I don't know if it's human psychology or not, to use any test method that might show that there isn't a problem.

As an engineer, I found *way* too much variation in a sloping target to have any confidence in it at all. It could make any camera fail I feel, yet people are still flocking to it in their droves and concluding their camera must be bad if it fails that test, when it's probably the test that is failing their camera.

EDIT: What this test *doesn't* tell you, is the degree of error. That's why it is a go/no go type check. But it's very simple to carry out in the first instance and easier to set up than Pekka's test. But I would recommend, if you still feel that there is a problem after doing this test, to check with Pekka's test set up.

This test is perhaps more appropriate for those who feel that they have an issue with the MF versus AF results from their camera. A misaligned focusing screen or mirror being the common cause.

Simon

[PaulB]

Simon writes,
"There seems to be a general reluctance, and I don't know if it's human psychology or not, to use any test method that might show that there isn't a problem"

On this, as on other, forums there seems exactly the opposite. Most people are hoping that there WILL be a problem with their camera.
Is there a conspiracy by Canon to sell us all dud 10Ds and then deny all knowledge of the faults?
I somehow doubt it myself, it would be commercial suicide - why is there a general movement from Nikon to Canon digital systems if Canons are so bad?
Over the years I have seen so, so many people blame their photographic equipment for bad technique when the main problem is their own lack of understanding of how to use what they have.
There is an old saying, "Only a bad workman blames his tools".
Tell me, how DID Ansell Adams do such magnificent work without a digital camera anyway?

Come on get real and start learning to take pictures.(if that really is the reason for owning all this expensive gear.)

[siwatkins]

PaulB wrote:
On this, as on other, forums there seems exactly the opposite. Most people are hoping that there WILL be a problem with their camera.


That's exactly what I said in different words, and is my point. There seems to be some psychology at work, whereby people are happy to undertake a flawed test, in spite of of problems being found with the test itself.

I trust therefore that the "get real" comment wasn't directed at me. I sit somewhere in the middle. I'm not so naive as those who proclaim that there isn't a problem - I've seen it for myself. There are two extremist views going on - those who are so doggedly brand loyal that they blind themselves to any possible flaws or faults with the mighty Canon - see that everywhere, but long term Canon users seem to be the worst of breed for some reason!

Then there are those who having spent so much money are determined to find fault, perhaps as some kind of purchase regret. Don't know.

But I'm also sure there were more problems with the earlier batches and that the later batches seem to be much better. There are still people using tests that have been shown to be unreliable and reflect the camera in a bad light, but won't do the revised tests - such as advocated by Pekka. Either through laziness or something else. Perhaps they don't want to acknowledge that the problem they though existed and kicked up hell about, doesn't exist. It would be natural and human for that to be the case, but it doesn't remove the fact that there is probably nothing wrong.

I trust the "get real" comment was *not* directed at me.

Simon

[defordphoto]

PaulB wrote:
On this, as on other, forums there seems exactly the opposite. Most people are hoping that there WILL be a problem with their camera.


So, so true. And that seems to be the trend. We've even seen a few D60 and 1Ds threads regarding focus problems. I even saw a post (somewhere) regarding focusing problems on a film-based camera!! Amazing. You say 'rain' on a sunny day and invariably, someone will definitely feel rain even when there is not a cloud in the sky.

Now we see all these threads regarding lens focus problems and a myriad of other focus-based issues. There have been probably a million photos shot of rulers since the 10D was released.

What a waste.

Seems like some folks have purchased cameras to just test them the rest of their lives and then troll the forums.

Get out and shoot!!

[siwatkins]

What amazes me on some of these forums, is how some people can't comprehend what is being said.

Perhaps it is just my UK English versus American English or differing standards of education - don't know.

To coin a phrase I read yesterday here, perhaps some ought to read what I said again, "this time taking in what it says, not what you want it to say!"

I have no problems with my camera, and am very happy with it. It troubles me as an engineer (I don't know why) when I see some users going off half-cocked because they have brainwashed themselves into seeing a problem, using fallible tests that identify a problem (incorrectly), and get themselves into a logjam and cycle of events of sending their cameras back unnecessarily, instead of being out there enjoying their cameras in real life.

The last two contributors of this thread seem to have completely misread my intention - to simply provide a more reliable method for those users unhappy, or uncertain to do a simple no-go test, then get on with their photographic lives. I see it as a rather more productive effort in helping those reach that goal of taking real photographs, when understandably, with all the noise about focus problems, some users have concerns. Just telling them to get on with it, is less than productive, as their confidence has been eroded by the noise. Providing a simple check that those with underlying concerns can undertake and satisfy themselves that their equipment is working properly, might just enable that same goal that those without problems preach - taking photos.

I don't think that's too difficult a concept too understand, but if necessary I'll clarify further. There seems to be far too much nastiness and unproductive clamouring to blame the user - especially for those who haven't experienced a problem camera. I'll leave it at that, but please read properly and digest the message before shooting it down.

Simon

[photography By Evangelos]

All these tests are driving me nuts. I have 2 canon EOS 10D cameras and am suffering alot of out of focusing issues. I am a canon EOS system user for over 10 Years now and it seems to me all of the Digital cameras suffer some kind of A/F problem. The D30,D60 and now the 10D. The 10D is the best so far. I use only the center A/F point as I have been for the last 10 years on all my EOS film cameras. I never had this problem with the film cameras. I guess it is digital growing pains. There is a problem some ware in the EOS system which causes the out of focus problems of the 10D. When I use the camera and I am not working at a wedding it is fine. But when I soot alot at my weddings I have this problem I avg. 500 to 600 per wedding and am often suprized when I know they sould me spot on at F 8 and F 11. I often shoot more than one of the same shot so as trying to work around the problem. But some times things happen so fast and you only get one chance. I often miss my EOS 3 A/F system and wish I had it on my 10D. I guess I will send bowth of them in when I get a chance as I am a CPS member so I get fast turn around times. Good luck to all of you as this camera seem to have a problem and canon will not admit it just like the did not with the D60's A/F in low light. The D60 should have had a waring on the canon box.

[defordphoto]

There's nothing wrong with having another test, especially if it's an easier one than shooting a graph or ruler. When I got my 10D I did my tests real-world and was very happy. Yup, I shot a couple of rulers and my camera is spot-on. I am extremely happy with my 10D and would buy another one in a heartbeat. It's very reliable and the build quality is definitely several steps up from the D30/D60's. And I also still have my D60 and still love that camera too. Yes, it sucks in low light, but that's when the MF switch gets turned on. And, if everyone would read their manuals, all Canon AF cameras whether film or digital explain that AF is NOT PERFECT. Never has been and I doubt it ever will be. People just need to get out and shoot and quit nit-picking.

Have fun!

[justme_dc]

On the contrary, I am willing and eager to try a test that will prove that there is or isn't a problem with my camera.

I tried your test this weekend and my 10D failed miserably.

I used a canon 50mm f1.8 and a canon 70-200 f2.8L to recreate your test. Everything was lined up and level. I did in fact see the moire pattern in the finder in both AF and MF shots but the resultant images were all out of focus. This can only mean three things, a problem with the prizm/focusing screen group, bad focus calibration or inproper sensor alignment. I have ruled out the Lenses. The lenses I used in the test have given me countless sharp and in focus images over the last 10 or so years. Images that have appeared in magazines, newspapers and on product packaging.

Although I am glad that this problem doesn't effect you all, I feel like most of the time people think my methods or photographic skills are the problem. I am not a hack, I am a skilled and competent photographer. There is a problem with some of these bodies.


It goes off to canon first thing tomorrow morning. I'll keep you posted.

[CyberDyneSystems]

Well,. I figured out I am too close,. and i can't change it,. (I know that's weird but its the room,.. there is literally no room behind my chair),.. and well I will have to try it somewhere else.

Moving the FW900 is a major military procedure! :D :D