Go Back   Steve's Digicams Forums > Digital SLR and Interchangeable Lens Cameras > Canon EOS dSLR

View Poll Results: Have you experienced focus problems with Canon EF lenses?
No - My EF lenses are near perfection 9 64.29%
No, but I wonder sometimes 2 14.29%
Yes, but Canon fixed the problem 2 14.29%
Yes, and Canon failed to fix the problem 1 7.14%
Voters: 14. You may not vote on this poll

Reply
 
Thread Tools Search this Thread
Old Mar 14, 2005, 12:30 PM   #21
Member
 
TDM_Canon_User's Avatar
 
Join Date: Jan 2005
Posts: 90
Default



I never missed your point, just was in front of your logic... I guess you aren't getting what I am saying - I'm saying a working D60 focuses better than a broken 1D. Normally, I prefer the 1D operation.


A basic concept: what does SIZE have to do with anything here? Only the image area subtended by the focus rectangle. If that's the same, then of course it has everything to do with resolution. Algorithmically the number of pixels used allows better/higher Nyquist on the resulting focus results. Plus, two dimensions instead of one allows much better spacial information. I think we agree on the resolution part - all else being equal.

I also believe I understandhow youthink AF works - but you're wrong. AF doesn't work the way you think it does. It doesn't "look" at anything - it simply looks at an area and maximizes the contrast between pixels. It uses two focal points within that area to decide which focuses better, then moves the lens accordingly. More pixels means better capability. Higher efficiency pixels and better algorithms means better low contrast capability.

And: the coop-de-graw: one important detail you're obviously NOT aware of: a single focus rectangle in my D60 is IDENTICAL in area to one in my 1D. So your argument fails on that point alone.

Here's from the two manuals.In reality they enclose identical areas as far as I can tell - point at something that fits inside a rectangle. Change bodies, look at same object, does it fit?



Attached Images
 
TDM_Canon_User is offline   Reply With Quote
Old Mar 14, 2005, 2:45 PM   #22
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

My fault then:

From Steve's Review the D60's AF sensor 'look' larger (I don't have a D60):



IMO the AF algorithm does not so much look for contrast between pixels since they can achieve contrast detection on the pixels with one lens - The reason they have two secondary lenses (actually two sets) in this AF system is such that they can detect phase difference i.e. the near/far defoccus info and move the lens accordingly - without hunting back and forth before achieving focus

-> One can decrease the contrast with front filter (or lower the light) and the camera can still focus, actually a wider aperture is more important here than contrast: This is a quote from the previous link "There are two sets of such lenses, one acting on each eye, as well as upper and lower "eyes" not found in humans. The CMOS area AF sensor binds these four parallax images together" which is why I reasoned a larger area of the image would help...

So this broken 1D doesn't focus at all or just not as good as the D60 in some 'real-life' situation??? :?:
NHL is offline   Reply With Quote
Old Mar 14, 2005, 3:22 PM   #23
Super Moderator
 
peripatetic's Avatar
 
Join Date: Nov 2004
Posts: 3,599
Default

TDM_Canon_User wrote:
Quote:
First - I take photos of real life objects, not test targets. I need my camera to focus on real life targets. My D60 does fine - my 1D once did, now is broken, just like my first 1Dm2 body was (from teh factory no less).
You obviously know a great deal about photography and cameras, far more than I. But I do know about troubleshooting complex systems with lots of variables. The fastest way, sometimes the only way, to an accurate diagnosis is to isolate the variables under controlled and artificial conditions until you can vary the parameters one at a time.

I don't imagine anyone would suggest that you should henceforth only take pictures of test charts. After all you could print out a couple and dispense with the camera altogether. The aim is to have a camera that works on real life subjects - but what you're doing at this point is fault diagnosis; you have already established that the camera isn't working, now you need to find out why.

To that end insisting on only photographing real world objects in real world situations is rather odd, because it results in you making claims that you cannot verify. The methodology you're using is such that if you do hit upon the right answer it will be by accident and you will find it very difficult to replicate your results.

Simply putting a different model of camera next to your malfunctioning 1D and taking photographs of the same subject proves almost nothing; you can end up arguing about the area of autofocus sites and the algorithms used on the different models till you're blue in the face and it won't get you any closer to finding out what's wrong with your camera. The user manual and reviews won't give either of you the answer to which system uses larger AF zones, etc. for that you will need access to the technical manuals that Canon engineers have access to. There is nothing to indicate that the AF sites have to match the etching on the viewfinder, they could be much larger on the 10D or smaller or exactly the same size. And here's the coup de grace: none of that gets you any closer to finding out what's wrong with your 1D. If you put another 1D next to it you'd be removing a variable from the equation which would get you closer to proving something. But not very much closer because of the type of photos your taking to do these "tests".

Quote:
Second - did you actually look at the PDF file? If not, look at the images then make the statement its a sharpness problem (then I'll know to ignore such comments).
I did look at the pdf file and most of the shots were suggestive of a focus problem, however some of them might not have been. Too many variables to tell.

Quote:
By the way - what's not shown are images that ARE in focus
Ah right so we're supposed to look at the pdf and gather evidence of shots NOT shown in it somehow?

NHL strongly suspects that there is nothing wrong, my guess is that there almost certainly is something wrong, (after all you have 30 years' experience of taking photos of "real life objects" and that must count for something) but the way you're going about proving it you will have a very difficult time convincing anyone (including Canon) of that fact.


peripatetic is offline   Reply With Quote
Old Mar 14, 2005, 10:47 PM   #24
Member
 
TDM_Canon_User's Avatar
 
Join Date: Jan 2005
Posts: 90
Default



I think teh reviewer is incorrect. I have measured my camera - for angle, not for active area -and the 1D and D60 are equivalent. The 1D system indeed uses two "regions" to determine focus. With two you can determine both slope and offset (ideally) and predict where to focus more quickly.

--

Peri,

Regarding minimizing variables... it took a while plus the purchase of the extra focus screen but I am down to one item that I think is the problem. The mirror. Sometimes it returns to "home", sometimes it doesn't quite get there.

Even manual focus with the split image screen was giving blurry results. And those results were identical to the image given with teh auto-focus. Basically, I'd manually focus on the distant object (near infinity), and check the lens indicator. Often it was not at "prime" infinity (not even close). Pressing the shutter to acheive focus resulted in the exact same focus position. Releasing the shutter resulted in a blurry image.

After running the camera at full-tilt for a hundred or so frames, focusing on an object on the horizon positioned the lens indicator at infinity. And, wah-laa, the image was in focus (well, I can nit-pick and say I've seen better). Anyway - right now the problem has dropped to an annoyance level, not an unusable level. Thus the plan to do a few more tests to include with my notes to Canon factory service.

Tim




TDM_Canon_User is offline   Reply With Quote
Old Mar 15, 2005, 2:06 AM   #25
Super Moderator
 
peripatetic's Avatar
 
Join Date: Nov 2004
Posts: 3,599
Default

A question - is there any benefit do you think to locking up the mirror and then using a timer or cable release and testing that?

My ignorance shows here, will the AF still work? Certainly manual focus would have to work.

If the shutter's not returning properely could it be dirt in the mechanism?

Sounds like you're pretty much there though. :-)
peripatetic is offline   Reply With Quote
Old Mar 15, 2005, 6:09 AM   #26
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

peripatetic

If the main mirror is up - both the photographer's viewfinder at the top and AF sensors in the bottom of the body will become blind (ie black-out) - the projected image go straight to the film plane:





peripatetic wrote:
Quote:
NHL strongly suspects that there is nothing wrong
Theses are what lead me to that opinion:
1. "I just had the body and all lenses at Canon for "adjustment" (in January)".
-> The entire system was just calibrated...

2. "Since this is my second 1D mk II to go nuts like this"
-> Canon QC can not be this bad, two in a row is rather unlikely!

3. "I just installed and tried a split image screen to ensure focus was where I wanted it. Autofocus ALWAYS matched manual focus - amazingly well - so the algorithm and lens control is working as expected."
-> The focusing screen is independent from the AF system and the camera focuses correctly!


What the problem is (I think), in some situation (real life objects, not test targets) it looks as if the D60 (with an inferior AF system) focuses better than the 1D :?

My reasoning came about this fact (righly or wrongly) - the active AF areas on the 10D (which is improved upon the D60) are much larger than the etched red indicator in the viewfinder. Anyone can try this: point the inside indicator of the AF marker to a featureless area of an object, the details outside the AF 'red square' if any will be locked on instead -> This 'design flaw' has led to many front/back focus issues on the 10D (just look @ the AF active sensor area vs what's etched in the viewfinder in the previous link).


On the other hand TDM_Canon_User is pretty much set on the 'mirror', let see if he can convince Canon
Remember - the sub mirror is oval (ever wonder why a satellite dish is shaped that way?) :idea:
NHL is offline   Reply With Quote
Old Mar 15, 2005, 10:04 AM   #27
Member
 
TDM_Canon_User's Avatar
 
Join Date: Jan 2005
Posts: 90
Default

Quote:
3. "I just installed and tried a split image screen to ensure focus was where I wanted it. Autofocus ALWAYS matched manual focus - amazingly well - so the algorithm and lens control is working as expected."
-> The focusing screen is independent from the AF system and the camera focuses correctly!
NHL, no the above isn't the case. The autofocus and manual agree, but the actual image taken isstill blurry. Hence a difference in the optical path between the sensor and the focusing mechanisms (manual and auto). The only thing common to both focus screen and AF sensors and not the CCD image sensor is the primary reflex mirror.

The first 1D NEVER gave a focused image, not even once, no matter what I did.

The second (current) 1D gave focused images for a while. Now doesn't.

Repeat: the problem is that the 1D began NOT focusing. Its been working just fine for about six months. It changedhow it works (after a trip to Canon, mind you, but that I think is just a coincidence).

The D60 has been fabulously consistent in provideing great focus.








TDM_Canon_User is offline   Reply With Quote
Old Mar 15, 2005, 3:24 PM   #28
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

TDM_Canon_User wrote:
Quote:
The only thing common to both focus screen and AF sensors and not the CCD image sensor is the primary reflex mirror.
True but:

1. Isn't this primary reflex mirror semi-transparent for the AF (ie not used in the reflective mode)?
2. Why is the sub-mirror elliptic (whereas the D60/10D sub-mirror is flat)?
NHL is offline   Reply With Quote
Old Mar 16, 2005, 12:54 AM   #29
Member
 
TDM_Canon_User's Avatar
 
Join Date: Jan 2005
Posts: 90
Default



1. If one mirror is misaligned, so is the other mirror; unless it too is sticking.

Assume the primary does a 90 degree fold to the focus screen, and to other mirror is mounted to the primary such that it points 90 degrees the other direction - or 180 degrees from the primary path.If the primary is off by 1 degree, the other mirror will be off by 1 degree. Simple geometry. The result is a change in back-focus distance (the distance between the lens and the focal plane). This is a common malady of reflex mirror systems.

2. The D60 sensor can "see" enough area without expensive concave mirrors, so a flat one works fine. The 1D focuses a greater visual area onto a small sensor area.


Anyway - when the dual mirror system is out of alignment, the autofocus system may not even be seeing the same object to focus on.
Attached Images
 
TDM_Canon_User is offline   Reply With Quote
Old Mar 16, 2005, 2:25 AM   #30
Super Moderator
 
peripatetic's Avatar
 
Join Date: Nov 2004
Posts: 3,599
Default

It seems to me that if the main mirror is not returning to the correct position that would affect both manual and auto focus.

However my question is would it affect them the same? I would suspect that if the mechanism doesn't return fully then the angle of the mirror would be different to it's correct value. Would that not make the AF and what is visible in the finder for manual focus disagree?

The other possibily though very unlikely is that the sensor has shifted away from being flat to the correct plane, if that were the case then we would expect to see one half of the picture more in focus than the other.
peripatetic is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 2:43 AM.