I suppose it's only fitting...#@$%%# BACKFOCUS!!!

[pixydust]

Apparently there WAS enough contrast, that's why it locked. How the
hell was the camera to know the photographer was aiming for the
blade not for the objects behind it ???

Isn't that why they put the AUTO in autofocus? If I'm given a
reference point with which the aim the camera, say that little
center autofocus square in the viewfinder, and I point that square
at something that I want in focus, isn't it it's job to know that's
what I"m pointing at? Otherwise, why have autofocus, or an
autofocus selection point?

does that mean the car drive itself? Sorry, I just can't resist

 

[Vern Stevens]

Apparently there WAS enough contrast, that's why it locked. How the
hell was the camera to know the photographer was aiming for the
blade not for the objects behind it ???

Isn't that why they put the AUTO in autofocus? If I'm given a
reference point with which the aim the camera, say that little
center autofocus square in the viewfinder, and I point that square
at something that I want in focus, isn't it it's job to know that's
what I"m pointing at? Otherwise, why have autofocus, or an
autofocus selection point?

does that mean the car drive itself? Sorry, I just can't resist

--
'Deceive, Inveigle, Obfuscate.' - The X-Files (Teliko)

http://www.pbase.com/vsteven

 

[Vern Stevens]

The lens is on it's way back to 17th St. Photo. The gave me an RMA# to exchange the lens. I think that probably means a quicker turnaround than repair, and hopefully, if it's the lens and not the camera, my problem will be solved when it returns. Thanks Olga for the very good (and obvious) suggestion! (which I uncharacteristically overlooked)

It's funny shipping the box at the Fedex place when you tell them you want to insure the box for $1650 and they are like "What's in that box?!?!"

Thanks all for the suggestions and comments. Even those I may have disagreed with, I do appreciate that you took time to give input.

I'll post some picks (hopefully by the end of next week) with the new lens. I will probably try some of the same shots, to see IF there is a difference in performance between the two lenses, and some easier more basic shots as well. It may or may not demonstrate where the problem actually was.

VES

 

[Ed Rotberg]

Good on ya mate!

= Ed =

The lens is on it's way back to 17th St. Photo. The gave me an
RMA# to exchange the lens. I think that probably means a quicker
turnaround than repair, and hopefully, if it's the lens and not the
camera, my problem will be solved when it returns. Thanks Olga for
the very good (and obvious) suggestion! (which I
uncharacteristically overlooked)

It's funny shipping the box at the Fedex place when you tell them
you want to insure the box for $1650 and they are like "What's in
that box?!?!"

Thanks all for the suggestions and comments. Even those I may have
disagreed with, I do appreciate that you took time to give input.

I'll post some picks (hopefully by the end of next week) with the
new lens. I will probably try some of the same shots, to see IF
there is a difference in performance between the two lenses, and
some easier more basic shots as well. It may or may not
demonstrate where the problem actually was.

VES

--
= Ed Rotberg =

'A waist is a terrible thing to mind'
http://www.edrotberg.org/gallery

 

[Victor Engel]

does that mean the car drive itself? Sorry, I just can't resist

Actually, yes. It's not dependent on a horse (which drives the carriage). What it doesn't do is steer itself.

 

[keef]

i never mentioned rulers

i sggested a test that takes 60 seconds

but of course, its much better to send of a camera to have fixed and wait several weeks than to spend 60 seconds finding out if it needs fixing.

and of course, taking back the lens is the best. I think we all assumed they would not change it because this is so obvious. I really have to wonder what the original poster is thinking about. As i mentioned, last wekend I bought a lens, discovered it front focused, returned it, refund, new lens, end of problem. I did all this and a lens test in the time the poster has spent on this one thread.

http://www.waddo.net

keith

 

[processor]

10D's AF is a transplant from Elan obviously without any adjustment of CoC (acceptable amount of blur) to match tiny 7.4 micron pixels.

Canon's standard CoC is 0.035mm. Draw a circle of 4-5 pixels in diameter on your picture and you will see that the area under the bottom AF point is blurred right by this amount. It's Elan's AF system and it thinks that this area is reasonably in focus.

In other words, a 4-5 pixel blur on 10D is within the spec by design. It's just the design is flawed:
http://www.canondslr.com/articles/af-flaw/

Mishkin argues that 10D's AF should have been revamped to provide at least 1.6x more accuracy than Elan. To shoot consistently pixel-sharp pictures, it should be 4x more accurate.

in so shallow a DOF; yet there has to be a sane and reasonable
explanation as to why it hit the cheeks to the side of , under and
in front of the eyes while missed the eyes altogether.

Where the heck is Mishkin when you need him!! LOL

 

[processor]

has been documented by Mishkin
http://www.canondslr.com/articles/ai-servo/
http://www.canondslr.com/articles/af-accuracy/

By the way, yesterday I was shooting some test pictures of a paper
target of a color grid in order to make profiles for neat image. I
noticed that with the camera tripod mounted and pointed at the
target, if I repeatedly half-pressed the shutter release (asking
the camera to refocus) and observed the focus ring, I could see it
barely adjusting itself sometimes. I know this has been discussed
on this forum before. I think it clearly demonstrates a random
component in the focusing system. I'm curious how many others here
get similar results. Note that this test is only valid if the
subject does not extend beyond the focal plane -- there is no
ambiguity in what is really in focus.

 

[Mark Rodrigues]

You are the voice of reason.

Canon's standard CoC is 0.035mm. Draw a circle of 4-5 pixels in
diameter on your picture and you will see that the area under the
bottom AF point is blurred right by this amount. It's Elan's AF
system and it thinks that this area is reasonably in focus.

In other words, a 4-5 pixel blur on 10D is within the spec by
design. It's just the design is flawed:
http://www.canondslr.com/articles/af-flaw/
Mishkin argues that 10D's AF should have been revamped to provide
at least 1.6x more accuracy than Elan. To shoot consistently
pixel-sharp pictures, it should be 4x more accurate.

in so shallow a DOF; yet there has to be a sane and reasonable
explanation as to why it hit the cheeks to the side of , under and
in front of the eyes while missed the eyes altogether.

Where the heck is Mishkin when you need him!! LOL

 

[processor]

I noticed that most of your examples are shot in the shade (cool light).

My experience particularly with 70-200IS is that it backfocuses in the shade (esp. in the low light). And it frontfocuses in warm light.

Could the light temperature affect your shots?

I suggest using Mishkin's technique to quantify the problem before you send the lens for adjustment:
http://www.canondslr.com/articles/m-dream

Example: before - 60 microns backfocus, after the 1st service - 80 microns frontfocus, after the 2nd service - 0 microns.

 

[Victor Engel]

That should be precise, not accurate. It's the precision that is the problem, not accuracy.

Canon's standard CoC is 0.035mm. Draw a circle of 4-5 pixels in
diameter on your picture and you will see that the area under the
bottom AF point is blurred right by this amount. It's Elan's AF
system and it thinks that this area is reasonably in focus.

In other words, a 4-5 pixel blur on 10D is within the spec by
design. It's just the design is flawed:
http://www.canondslr.com/articles/af-flaw/
Mishkin argues that 10D's AF should have been revamped to provide
at least 1.6x more accuracy than Elan. To shoot consistently
pixel-sharp pictures, it should be 4x more accurate.

in so shallow a DOF; yet there has to be a sane and reasonable
explanation as to why it hit the cheeks to the side of , under and
in front of the eyes while missed the eyes altogether.

Where the heck is Mishkin when you need him!! LOL

 

[Victor Engel]

Yeah, that's what I was referring to, but I didn't have the link handy. BTW, I think he should be using a front surface mirror to avoid any refraction effects from the mirror itself.

By the way, yesterday I was shooting some test pictures of a paper
target of a color grid in order to make profiles for neat image. I
noticed that with the camera tripod mounted and pointed at the
target, if I repeatedly half-pressed the shutter release (asking
the camera to refocus) and observed the focus ring, I could see it
barely adjusting itself sometimes. I know this has been discussed
on this forum before. I think it clearly demonstrates a random
component in the focusing system. I'm curious how many others here
get similar results. Note that this test is only valid if the
subject does not extend beyond the focal plane -- there is no
ambiguity in what is really in focus.

 

[processor]

The thickness of the mirror (2mm) corresponds only to 1 micron error at the sensor in Mishkin's setup. There's no need in surface mirror (although it would be nice if not expensive).

By the way, yesterday I was shooting some test pictures of a paper
target of a color grid in order to make profiles for neat image. I
noticed that with the camera tripod mounted and pointed at the
target, if I repeatedly half-pressed the shutter release (asking
the camera to refocus) and observed the focus ring, I could see it
barely adjusting itself sometimes. I know this has been discussed
on this forum before. I think it clearly demonstrates a random
component in the focusing system. I'm curious how many others here
get similar results. Note that this test is only valid if the
subject does not extend beyond the focal plane -- there is no
ambiguity in what is really in focus.

 

[Vern Stevens]

Isn't Mishkin's technique aka processor's technique? :?

I had already satisfied my reasoning with real world shots such that I didn't believe any further testing would be necessary. I know this isn't enough for some folks, but it was enough for me. However, in the example you cited, having tested the camera with Mishkin's system didn't prevent you from having to send it to Canon for repair, and it didn't ensure that they would get it repaired right because you had to send it in for the second repair. I'm not saying that to be rude, just saying it as an observation.

In my case, I (easily) saw it had problems with regular pictures, and more than just what I had posted. IF I thought it had a problem, but it wasn't as obvious to me in regular pictures, THEN I might resort to more precise testing of line charts, etc. I think the brick shot I posted, although not as precise, is very close to shooting lines.

Thanks for your comment.

VES

I noticed that most of your examples are shot in the shade (cool
light).

My experience particularly with 70-200IS is that it backfocuses in
the shade (esp. in the low light). And it frontfocuses in warm
light.

Could the light temperature affect your shots?

I suggest using Mishkin's technique to quantify the problem before
you send the lens for adjustment:
http://www.canondslr.com/articles/m-dream

Example: before - 60 microns backfocus, after the 1st service - 80
microns frontfocus, after the 2nd service - 0 microns.

--
'Deceive, Inveigle, Obfuscate.' - The X-Files (Teliko)

http://www.pbase.com/vsteven

 

[Vern Stevens]

It's my time to waste. You shouldn't take this so personally. The only thing your test would have done was convince YOU. I was already convinced there was a problem therefore, 60 seconds or not, I didn't NEED to do your test. Since I'm the one with the investment, NOT YOU, I have to make decision based on my criteria, NOT YOURS.

If you are so personally offended, please don't waste any more of your time worrying about whether I'm wasting my time.

Thanks,

VES

i never mentioned rulers

i sggested a test that takes 60 seconds

but of course, its much better to send of a camera to have fixed
and wait several weeks than to spend 60 seconds finding out if it
needs fixing.

and of course, taking back the lens is the best. I think we all
assumed they would not change it because this is so obvious. I
really have to wonder what the original poster is thinking about.
As i mentioned, last wekend I bought a lens, discovered it front
focused, returned it, refund, new lens, end of problem. I did all
this and a lens test in the time the poster has spent on this one
thread.

http://www.waddo.net

keith

--
'Deceive, Inveigle, Obfuscate.' - The X-Files (Teliko)

http://www.pbase.com/vsteven

 

[Victor Engel]

Note that I was not referring to the placement of the reflecting surface but to the refraction of the glass in front of it. Using a rear-surface mirror, you're essentially adding another lens element to the mix. A rear surface mirror also has internal reflections and reflections from the glass and the metal surface, causing multiple images.

Are these effects insignificant? Perhaps.

By the way, yesterday I was shooting some test pictures of a paper
target of a color grid in order to make profiles for neat image. I
noticed that with the camera tripod mounted and pointed at the
target, if I repeatedly half-pressed the shutter release (asking
the camera to refocus) and observed the focus ring, I could see it
barely adjusting itself sometimes. I know this has been discussed
on this forum before. I think it clearly demonstrates a random
component in the focusing system. I'm curious how many others here
get similar results. Note that this test is only valid if the
subject does not extend beyond the focal plane -- there is no
ambiguity in what is really in focus.