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

Reply
 
Thread Tools Search this Thread
Old Oct 4, 2006, 9:19 AM   #11
Senior Member
 
Widowmaker's Avatar
 
Join Date: Jul 2003
Location: Bakersfield, CA
Posts: 851
Default

Ok heres more input

Looking at the two shots using the memory card, the card is definitely soft but look at the text "polygram video" and compare on both. If the canon is backfocusing would this not be sharper? It appears to be the same to me on both, maybe slight color differences.

Heres what I would do.

Take a ruler or yard stick and lay on the table or floor. Set your tripod/camera up so that it is shooting down the ruler at a 45 degree angle. Choose the same mark on the ruler as the focus point for each camera. I'm not sure what the minimum focus distance is on that lens but if its to long to let you set the ruler on the floor or table then fasten the ruler to a wall and shoot at a 45 degree angle down the wall. When you are ready to shoot make sure you are set up about 1 foot or more longer than the minimum focus distance. For example if the min on that lens is 8 feet, set up at least 9 feet from the focus point.

This will be much more accurate. Also do not use the flash. Use your timer to activate shutter and set mirror lockup on. The reason I say do not use the flash is because you are just adding in another variable. You need to eliminate anything that might change the outcome.

Something like this however I was not at 45 degrees.. I just wanted you to see what I am talking about.

Attached Images
 
Widowmaker is offline   Reply With Quote
Old Oct 4, 2006, 10:25 AM   #12
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

Mark1616 wrote:
Quote:
This I think is the most alarming test as I believe the camera should not be losing focus on a stationary object while taking multiple shots (with AI Servo). The first 2 are OK but then things start to go wrong.
Mark

This may be your whole problem, I've stressed this may time "AI servo" is (1) predictive and (2) it is shutter-priority - This is why I shoot in "one-shot" AF even on the MrkII, this mode is focus-priority instead.

-> Try this to put the lens cap on your lens and press the shutter - The camera will release anyway to keep up the frame rates regardless if the subject is in focus or not!
Now try it with "one-shot" AF - You can't release the shutter can you? the AF need to be lock-on before the camera release - It also maintains focus on subsequent shot, but won't track the subject (all you have to do then is to let go of the shutter all the way and the camera will re-lock again)

IMO the camera is mistaken your camera shake as the stationary object is actually moving :idea:
(i.e. It's predicting a movement which is not there)
NHL is offline   Reply With Quote
Old Oct 4, 2006, 10:31 AM   #13
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

Mark1616 wrote:
Quote:
That's taken care of the moving subject where there seems to be a lock of clarity in all bar one which is not too bad. The rest seem to be clearer on the left and centre of the frame (feet, legsand middle) rather than the head so is there an issue with a portion of the lens or a focus plane issue? I don't think it is depth of field as the legs etc cover more than the range of the head but they are in focus.
Try the newspaper test I suggested in my previous post.

It's normal to see some softness in corners, and a focus plane is not going to be flat with most lenses. See the diagrams in this article:

http://www.mhohner.de/recompose.php?lang=e

But, you should have equal softness in all 4 corners if the camera is square to a target like a newspaper. Otherwise, you could have a misaligned element in the lens, or even a misaligned CCD sensor (or rather CMOS in this model's case) in the camera body.

I'd also suggest looking at the images using Canon's software to see what focus points were being used. I think it can show that to you. If not, the camera itself can show the focus point being used when you review images on it's LCD.

You could have misaligned AF sensors, mirrors, etc, impacting AF accuracy. The AF sensors can also be misaligned so that one focus point is OK and another is not. Or, it could be you (focus point being selected, locking focus while subject still moving, depending on focus mode, etc.).

Also, even the temperature of the light can sometimes impact AF accuracy (incandescent versus daylight), as can the focal length of a lens being used (it may be OK at one end, and not at the other).

I've gpt a feeling that NHL will dispute that a lens can be the cause of AF accuracy. :-)

But, I've seen it before. From what I understand, lenses have a table in EEPROM that is used to determine the amount of movement needed based on instructions from the camera (and also feeds this info back to the camera). I don't understand all of the technical details, but I've seen the problem demonstrated before (one lens may backfocus and another may be fine on the same camera body).


JimC is offline   Reply With Quote
Old Oct 4, 2006, 10:36 AM   #14
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

JimC wrote:
Quote:
I've gpt a feeling that NHL will dispute that a lens can be the cause of AF accuracy. :-)
... or may be the KM focus is better than the Canon's

Seriously though I think it's the 'predictive' AI-servo - All he has to do is to try again in "one-shot" AF mode instead and see if the accuracy improve...

-> Clearly the camera/lens achieved focus, but it's the susequent shots (or the inconcistency):
"This I think is the most alarming test as I believe the camera should not be losing focus on a stationary object while taking multiple shots (with AI Servo). The first 2 are OK but then things start to go wrong."
NHL is offline   Reply With Quote
Old Oct 4, 2006, 12:24 PM   #15
Senior Member
 
Widowmaker's Avatar
 
Join Date: Jul 2003
Location: Bakersfield, CA
Posts: 851
Default

Just for reference, I shot a stationary subject as you did. Using AI Servo I took a burst of 20 shots and the focus remained accurate all the way through.
Widowmaker is offline   Reply With Quote
Old Oct 4, 2006, 12:27 PM   #16
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

NHL wrote:
Quote:
... or may be the KM focus is better than the Canon's
It might be with that Sigma on it. I understand that it's geared to be a fast focusing lens on a KM DSLR. :-)

I don't have one for my KM 5D. If I were looking for a new f/2.8 zoom in that focal range, I might be tempted to get the Sigma.

But, I would seldom use anything longer than my 100mm f/2. My longer 135mm f/2.8 is not used very much in comparison to my Minolta 24-85mm f/3.5-4.5 or my Tamron 35-105mm f2.8 or my Minolta 100mm f/2 . I use the Minolta 28mm f/2 more than any other prime, and the Minolta 24-85mm f/3.5-4.5 is my most used walk around.

These are the lenses I've got for use on my KM Maxxum (a.k.a., Dynax) 5D, all are Autofocus and use the the AF motor built into the camera body (as all lenses for this camera do, with the exception of Minolta's SSM lenses):

Minolta
----------
28mm f/2
50mm f/1.7
100mm f/2
135mm f/2.8
18-70mm f/3.5-5.6
24-85mm f/3.5-4.5
35-70mm f/4 Macro

Tamron
-----------
20-40mm f/2.7-3.5
35-105mm f/2.8

Vivitar
--------
70-210mm f/2.8-4

JimC is offline   Reply With Quote
Old Oct 4, 2006, 12:28 PM   #17
Super Moderator
 
Mark1616's Avatar
 
Join Date: Oct 2005
Posts: 7,397
Default

Thanks Guys,

Jim have done the newspaper test and that seems OK, but light was not good so had to use 3200 ISO so will try that again when it is better.

Widowmaker what kit are you using again?
Mark1616 is offline   Reply With Quote
Old Oct 4, 2006, 12:29 PM   #18
Senior Member
 
Widowmaker's Avatar
 
Join Date: Jul 2003
Location: Bakersfield, CA
Posts: 851
Default

1D MKII and 300 2.8 IS
Widowmaker is offline   Reply With Quote
Old Oct 4, 2006, 12:34 PM   #19
Super Moderator
 
Mark1616's Avatar
 
Join Date: Oct 2005
Posts: 7,397
Default

Cool, so it does not jump around (focus that is while looking at the same subject)? While on full zoom looking at my fathers face it was constantly adjusting itself, could this suggest that the AF is too sensitive?
Mark1616 is offline   Reply With Quote
Old Oct 4, 2006, 12:42 PM   #20
Senior Member
 
Widowmaker's Avatar
 
Join Date: Jul 2003
Location: Bakersfield, CA
Posts: 851
Default

If I just halfpress it will adjust but then I am hand holding so it should since my distance to/from subject is changing slightly... it does not jump though like its hunting. There are various adjustments in the MKII for AI like sensitivity and expanding the focus area around the chosen focus point but I am set at default so it shouldn't matter.
Widowmaker 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 2:47 PM.