PDA

View Full Version : Multiple cameras, multiple lenses


Tony Field
June 1st, 2006, 12:03 PM
One of the interesting things I recently found out is that my 1D-IIn and 5D cannot be perfectly calibrated to all of my 10 lenses. The service tech here in Calgary indicated that it is possible to almost perfectly calibrate (within human ability) all lenses to a single camera however it is "impossible" (except by manufacturing accident) to perfectly do that with multiple bodies.

Here is the resulting situation. My 1D-IIn virtually perfectly focuses my 24 1.4, 35 1.4,50 1.4, 85 1.2, 100 F2, 135 F2, 300 F2.8, and zoom lenses - the point if maximum sharpness is where it should be - 1/3 to 1/2 into the depth of field region.

The 5D, on the other hand consistently focuses in the back 2/3's of the DOF for all lenses.

The Canon service tech (who is VERY competent, IMHO) says he cannot do better than that- it is part of the design constraints of the cameras. Since the consistency of focus is very fine indeed for each camera, he obviously has the skill to do the very best possible.

The minor down-side of this is that the 1D-IIn becomes the camera of choice for large-aperture shots with limited DOF. On the other hand, since the focus consistency of the 5D is so well done for all the lenses, I guess it is reasonable to "learn" to focus on alternate spots of the subject to allow sharp wide-open shots.

It would be interesting to know if this is really the case :-) Is it possible for perfect calibration for many lenses, many bodies?
tony

dhphoto
June 1st, 2006, 12:58 PM
I take it by calibrate you mean autofocus?

I assume that all the lenses would be in focus on all the bodies if carefully focused manually, that is, it is the autofocus tolerances you are talking about?

David

Tony Field
June 2nd, 2006, 09:55 AM
Yes, it is autofocus that I am talking about. IMHO, the autofocus should always be virtually perfect and yield images which are sharply focused in the approximate centre of the depth of field for that lens. For example, if you have the 85mm F1.2, wide open it should have virtually perfect auto-focus ability on any body.

Daniel Harrison
June 3rd, 2006, 05:40 AM
This is very interesting, I hear alot about this. In my understanding a camera focusses optically and therefore it should not be the cause of the lens. But I know the lens does affect the focussing becuase I am having 2 of my sigma lenses calibrated and one basically rebuilt.

The sigma service people seem convinced that different Canon bodies focus differently and that they do not usually adhere to a perfectly calibrated AF standard. (although they believe it can be properly calibrated -maybe :-) ) I am not sure what to make of it but I will say that a Canon 70-200mm f/4 seems to focus perfectly on my 1D and a 10D.

I would think that if you had your cameras AF adjusted they should work the same with your lenses. I can't see a reason why they can't adjust them to be the same. Keep at them, and let us know what they say.

Daniel

Michael Tapes
June 3rd, 2006, 06:06 AM
I will ask Chuck Westfall to comment on this thread. Thanks for the heads up. Great to know, but depressing if true...

Chuck Westfall(Canon USA)
June 5th, 2006, 10:20 AM
Hi, Tony:

With all due respect to the service technician, the information that you are reporting here is not 100% accurate. What it really comes down to is that no individual camera can produce so-called 'perfect' autofocus for all photos, no matter which lens is in use. The fact of the matter is that there is always a tolerance. As long as the focusing data coming into the camera indicates that the focus is within the tolerance, the camera will report that the subject is in focus.

The specifications of the tolerance may vary. As you probably know, Canon is fairly unique in the world of autofocus SLRs by providing so-called 'normal precision' AF that sets the tolerance to a predetermined depth-of-focus, as well as 'high precision' AF that sets the tolerance to 1/3 the level of 'standard precision' AF.

Regardless of the degree of precision involved, there is always going to be a range of focusing distances that the camera deems acceptable for any given lens. Even under ideal testing conditions, such as mounting the camera on a steady tripod, setting the focus mode to One-Shot AF, and using a stationary, readable target, there is going to be a slight variation in the range of focusing distances that the camera considers acceptable. If one were to perform a true test, where the lens was intentionally preset to an out-of-focus condition, then autofocused, and this procedure was repeated 100 times in a row, the resulting data would prove beyond a shadow of a doubt that there would be some variation. Every shot would most likely be within tolerance, but not every shot would be focused identically, and therefore, not every shot would be 'perfect.' It's unavoidable.

This is what tolerances are all about. The best that any technician can do is to adjust your equipment so that it performs within the tolerances it was designed for. In this respect, it most certainly is possible to calibrate any number of camera bodies and lenses so that they all perform according to their design specifications.

Now, this is not to say that the aim point for exact focus can't be tweaked on individual cameras and lenses. However, no matter where the aim point ends up, there will be some variation in AF performance on a per-shot basis because of the tolerances built into the system. Bottom line: Aim points can be tweaked, but tolerances cannot.

Hope this helps!

Best Regards,

Chuck Westfall
Director/Media & Customer Relationship
Camera Marketing Group/Canon U.S.A., Inc.

Tim Dolan (Longwatcher)
June 5th, 2006, 02:37 PM
Followup question,
I am somewhat confused if the original question by the original poster has been answered so I am asking again in a different way.

I understand lenses can be adjusted to set the optimum focal point to match the camera's sensors (within tolerances of course)

My question is:

Can the camera's sensor also be adjusted?

To say a common focal point so that a lens that has been properly adjusted will match two or more cameras. If only the lenses can be adjusted then they can only truly match one camera. If only select cameras can be adjusted a list would be handy.

I have a mental picture of one camera that comes out of the factory at one end of the tolerance spectrum and then a second camera coming out at the other end and then technitions matching the first camera's tolerances some of which will be even farther out from the second camera and some that will be closer to the second.

I would expect that the cameras can be adjusted, but have also learned to ask just to make sure. I follow an advance UAV sensor program where they forgot a couple of things I thought were common sense to add but because no one actually asked them to add the features they were left off the prototype.

Thanks,

Chuck Westfall(Canon USA)
June 5th, 2006, 03:53 PM
>>Can the camera's sensor also be adjusted? <<

Yes. Each item (each camera and each lens) can be calibrated according to its own standards by a Factory Service Center technician.

In practice, each camera is calibrated with a 'tool lens' with known, consistent performance characteristics.

Here's my point: For the sake of discussion, let's assume that a specific camera/lens combination is calibrated 'perfectly.' Even when that is true, the actual focusing distance set by the camera can vary from shot to shot within the tolerances established by the design specifications of the equipment. This is normal, and also unavoidable.

Let's further assume that Tony's description is accurate when he says:

>>My 1D-IIn virtually perfectly focuses my 24 1.4, 35 1.4,50 1.4, 85 1.2, 100 F2, 135 F2, 300 F2.8, and zoom lenses - the point [o]f maximum sharpness is where it should be - 1/3 to 1/2 into the depth of field region.

The 5D, on the other hand consistently focuses in the back 2/3's of the DOF for all lenses.<<

If that's the case, then his 5D is not calibrated to the same aim point as his 1D Mark II N, even though both may be operating within the tolerances they were designed to meet. It is possible to calibrate both cameras to the same aim point, but it is not possible to change the tolerance range of the AF system.

Best Regards,

Chuck Westfall
Director/Media & Customer Relationship
Camera Marketing Group/Canon U.S.A., Inc.

Mike Spinak
June 5th, 2006, 04:09 PM
Chuck,

Thank you for your illuminating answers, so far.

As you probably know, Canon is fairly unique in the world of autofocus SLRs by providing so-called 'normal precision' AF that sets the tolerance to a predetermined depth-of-focus, as well as 'high precision' AF that sets the tolerance to 1/3 the level of 'standard precision' AF.

I didn't already know about this. Could you please clarify this point and elaborate upon it? Canon provides "high precision" AF to whom, and provides it how?

Thank you.

Mike

Tony Field
June 6th, 2006, 12:20 AM
Hi Chuck,

Thanks for the careful discription of the Canon's basic focus process. I will soon be revisiting the local Canon office and get this one minor point resolved.

tony

Frank Werner
June 6th, 2006, 01:52 AM
Chuck,

I didn't already know about this. Could you please clarify this point and elaborate upon it? Canon provides "high precision" AF to whom, and provides it how?

Thank you.

Mike

The 1 series cameras(and the 5 D/20D/30D as I know of) have two kinds of AF Sensors. The 1 Ds II by example has 45 AF sensors, 38 of whom are only sensitive to vertical lines. 7 are cross sensors and sensible in two ways horizontal and vertical and therefore more sensitive then the others. The sensibility of the cross sensors is about three times higher then the sensibitly of the vertical ones. The 5D/20D/30D have only one high precision AF Point which is the central AF Point. If a AF Point can act as High Precision AF Sensor depends also on maximum aperture of the used Lens. See an article of Chuck here:http://www.bobatkins.com/photography/eosfaq/eos3af.html

Or have a look in the manual of your Camera if you have a 1 series Camera its explained in the AF Section of it.

Frank

Chuck Westfall(Canon USA)
June 6th, 2006, 08:29 AM
Hi, Mike:

Frank's information is correct. Here is a diagram of our 45-point AF system, as used in the EOS-1 series of digital SLRs.

Best Regards,

Chuck Westfall
Director/Media & Customer Relationship
Camera Marketing Group/Canon U.S.A., Inc.

Mike Spinak
June 6th, 2006, 11:00 AM
Ah... this is what you meant.

Yes, I was aware of that. I had misunderstood.

Thank you.

Mike