Why do you calibrate lens?

norahbattie

norahbattie

Suspended / Banned
Messages
348
Name
Nads
Edit My Images
Yes
A basic question.....
If a camera is brand new and straight out of the box why do people talk of calibrating lenses and bodies etc? And if you calibrate don't you run the risk of your other lenses being off?

Also how often are you supposed to service your camera?
Thanks
Nads
 
Sometimes a lens maybe a fraction out of focus, barely noticable sometimes and can be micro adjusted with some canon bodies.

If its really bad send it to Canon for calibration, 99% of the time its fine to use out of the box
 
norahbattie said:
A basic question.....
If a camera is brand new and straight out of the box why do people talk of calibrating lenses and bodies etc? And if you calibrate don't you run the risk of your other lenses being off?

Also how often are you supposed to service your camera?
Thanks
Nads
Click to expand...


Calibration is done per lens... so diff calibration for each lens... only usually noticable on lens with f2.8, f2, f1.8 etc with narrow DOF

how often to service depends on how much you use it.. I take thousands of pics a week.. if you take dozens of pics a week then our service times would be miles different :)
 
The only one I've had to do it on was my 85mm f/1.4 which was about 3-5mm out at the closest-focus distance.
Same amount of adjustment required on all three of my camera bodies.
It's the Lens' CPU telling the camera that it's in-focus when in fact it isn't.
Not a hardware problem, just a software glitch. All you're doing when performing an AF-fine-tune is telling the camera to override the AF information for that lens.

If it's a very long way out or different amounts for different camera bodies then it needs to be re-calibrated by the manufacturer or authorised service centre.
 
Most people do it so when they are pixel peeping at 300% in PS4 they can see a slight improvment in sharpness.

Some people do it because the manufacturing tolerances of the lens and camera can result in a visible degree of miss focus.
 
a1ex2001 said:
Most people do it so when they are pixel peeping at 300% in PS4 they can see a slight improvment in sharpness.

Some people do it because the manufacturing tolerances of the lens and camera can result in a visible degree of miss focus.
Click to expand...

It's not manufacturing tolerances - it's software.
 
Because Sigma don't calibrate them at the factory.


(I joke - I'm a sigma fanboy if I'm honest).
 
Arkady said:
It's not manufacturing tolerances - it's software.
Click to expand...

But surely the software is the same on every lens as it's just blatted onto a chip it is one of the few things in the manufacturing process that should be 100% the same on every single lens (of a given type) made?
 
So basically people spend hundreds of pounds on tiny adjustments which can hardly be seen?

I took my camera for a clean today and was wondering what the calibration was all about. I just want the dust off my sensor and tried to do a dry clean but failed :( so thought I would leave it to the pros
 
norahbattie said:
So basically people spend hundreds of pounds on tiny adjustments which can hardly be seen?
Click to expand...

Not Quite some lenses are really far of the mark, to the point where a shallow DOF you cant get a focus where you want it.

I have had to send a lens back because of this. it only cost £30 for the job to be done.
 
norahbattie said:
So basically people spend hundreds of pounds on tiny adjustments which can hardly be seen?
Click to expand...

Hundreds of pounds ? It took me 15 minutes and cost me nothing. Simply downloaded and ran a little utility that puts a calibration image on the screen of the laptop. Manual focussed on that then let the camera auto focus and that it done. As it was there was no change between manual and auto focus, if there had been I could have corrected it in the 7D settings.

Look Here : http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html
 
Somewhere on the Canon site Canon say that if you take three pictures you shouldn't be surprised to see differences between them and that this is normal. I don't have MA on my camera and I've not had any significant focus issues and in my innocents I therefore think that some people get a little too obsessed with MA and perhaps expect more than can actually be achieved. I also wonder how relevant it is to test and adjust for a particular distance and / or focal length but there must be something to it, I suppose, as it now seems to be a must have feature for any new DSLR to be taken seriously.

I've no doubt that I'll be obsessed with MA when I have a camera that has it.
 
I take pictures wiht my kit and calibration turned off.. the pics are poor.. I have calibration turned on and the pics are good .. i may use the wrong words but the important thing is.. it makes a real difference...

its not pixel peeping and its not useless... canon didnt add it to top end cameras as a toy..
 
woof woof said:
Every picture?
Click to expand...

every picture i take using a 400mm lens at f2.8 yes..

the only way it wouldnt be every pic is if i go to say f4 and the adjustment isnt noticable because the DOF os bigger
 
Thanks iPhoto, I was on my phone, and missed the link, sorry mate
Thanks
 
a1ex2001 said:
But surely the software is the same on every lens as it's just blatted onto a chip it is one of the few things in the manufacturing process that should be 100% the same on every single lens (of a given type) made?
Click to expand...

Sorry - I suppose in that sense you're absolutley right. I had a think about it last night while surrounding some more whisky and given that the software is the same then the only thing that could account for this is the manufacturing/assembly of the components as that's the only variable.
At least one other person here has the exact same issue with their 85mm f/1.4 as I do, so it may be a batch-to-batch issue.

As to norahbattie's remark that it can hardly be seen: 3mm out on a full-face portrait shot at minimum focus is pretty dramatic.
No point spending £1000 on a lens that can't be used wide-open, is there?
 
norahbattie said:
So basically people spend hundreds of pounds on tiny adjustments which can hardly be seen?

I took my camera for a clean today and was wondering what the calibration was all about. I just want the dust off my sensor and tried to do a dry clean but failed :( so thought I would leave it to the pros
Click to expand...

Not at all...there are two types of calibration that Canon offer...

One is the micro-adjustment which is user controlled. As Arkady states, it's the lens CPU telling the camera it's in focus, when it's actually out. Micro-adjustment alters the software side of things to get it back in focus.

If you're just not satisfied with that, or want a "sure-fire" method, then Canon are able to calibrate lenses and bodies to one-another. It costs £90 per body/lens combo, but you can only have one lens calibrated to a body. It alters the actual hardware configuration...literally moving the bits of glass around to ensure absolute calibration. It is MUCH better than micro-adjustment.

I went for the latter simply because I use one lens with one body much more regularly, and I can retain the micro-adjustment settings if I still need them for any other (less commonly used) lenses.
 
Micro Focus Adjustment is very good when you've got a prime lens, it's not so hot with zooms as you'll probably find that you require different MFA values at different focal lengths. When Canon do a lens/body calibration they can adjust it so that they work optimally throughout the FL range.
 
"every picture i take using a 400mm lens at f2.8 yes.."

Sorry, I was just being a bit flippant but I'm still not convinced that every picture can be out before and in after MA adjustment for two reasons. Firstly because autofocus systems simply aren't infallible or that consistent, they just aren't, hence Canon's comment about taking three pictures and seeing differences between them and this being perfectly normal. Secondly because of reason one, because it's such a biggie, and also because we focus on things that are different distances from us.

I'm not completely dismissing MA but I do think that it possibly isn't the must have or fix all that some believe.
 
I think that somebody who knows the exact process of AF needs to come here to explain the mechanics of it all. As a mechanical engineer myself (who knows very little of electroncs/computers) I would think that the camera should decide that focus had been achieved, and not the lens telling the camera that focus had been achieved?
 
I take this only applies to Canon cameras.
 
diggerz said:
I think that somebody who knows the exact process of AF needs to come here to explain the mechanics of it all. As a mechanical engineer myself (who knows very little of electroncs/computers) I would think that the camera should decide that focus had been achieved, and not the lens telling the camera that focus had been achieved?
Click to expand...

Neither the lens or the camera 'decide'. The AF system of the camera figures out how much the lens needs to move to get the focus and tells the lens to move that much. And that's it. There is no feedback to determine if focus has actually been achieved.

Neither of the processes (deciding how much movement is required and moving the required amount) can possibly be exact. Both will be subject to random and systematic errors. Canon's manufacturing tolerances are to get both processes correct to within one-half or one-third (at least I think those are the right numbers) of the depth of field, depending on the camera model and the specific AF point used.

The random errors (about which one can do nothing) are why Canon say that repeat shots may show some differences in focus. The systematic errors are what can be fixed by MFA.
 
But surely the sensor of the camera can "sense" that the selected area is sharp edged ie. in focus at any given point?
 
Surely the camera has most input into this? I thought it looked for the hardest most contrasty edges it could get across its focus points? That doesn't explain why different lenses would be differently calibrated on the same body though....
 
The problems of AF micro-adjustment are all mechanical/optical in practise, not software related.

Phase-detect AF works by comparing two images from either side of the lens, which is why it is much more dependent on a physically wide aperture than it is on bright light. It also operates within the zone of depth of focus rather than depth of field (as it must do) and while that is affected by f/number in the same way as depth of field, in terms of focal length they work completely opposite. Long lenses have lots of depth of focus, and wide angles have very little.

Assuming that the AF gets a decent lock on accurate focus, what it then does is instruct the lens to move by a certain amount. The accuracy of that movement is dependent on mechanical tolerances. At a given focal length and focusing distance, it is easily possible to get pretty much absolute accuracy every time. The problems arise when the lens has to move a different amount according to focal length and focusing distance, when a whole new set of mechanical tolerances apply each time - the lens has to move much further at closer range, and further at longer focal lengths too. So there are lots of permutations to cover, and lots of potential for error.

Which is why getting a zoom to be spot on at every possible setting is difficult, and why the lens is as critical to accurate AF as the camera. TBH I'm amazed that it is both as accurate and consistent as it is, in my experience.

On the other hand, contrast-detect AF (as in live view, and as in all these new mirrorless jobbies) doesn't suffer from any of these mechanical issues. In that sense, contrast-detect can never be anything but absolutely accurate every time, but the problem with that system is that it is both less reliable and slower at getting an accurate fix in the first place.

When those issues are sorted, and it's getting better all the time, contrast-detect will be brilliant. Imagine being able to lock on to a subject and track it accurately anywhere in the frame in real time :thumbs:
 
diggerz said:
I think that somebody who knows the exact process of AF needs to come here to explain the mechanics of it all. As a mechanical engineer myself (who knows very little of electroncs/computers) I would think that the camera should decide that focus had been achieved, and not the lens telling the camera that focus had been achieved?
Click to expand...

I cant help but Canon Can linky to full article.

Autofocus (AF)
How AF works
Autofocus is so easy to use that the focus switch on some of your lenses may never have been moved to the manual position.

Most of the time autofocusing gives excellent results. It is faster than manual focusing, and often more accurate. However, autofocusing operates by a series of rules (algorithms). There may be times when these do not give the results you want.

EOS cameras use a highly sensitive line sensor for autofocus ranging. Called BASIS (BAse Stored Image Sensor), it consists of two 48-bit line sensors and associated amplifier circuitry.

The sensor is in the base of the camera. A sub-mirror behind the camera’s reflex mirror reflects light down to the sensor. This light is split by a small lens assembly to form two separate images. One image is formed on the first line sensor, the other on the second line sensor. If there is no deviation between the two images seen by the sensors, the lens is focused. However, if the spacing of the two images is not correct, a signal is sent to the lens motor to bring the subject into sharp focus.

There's some info around which of the focus points can actually be used with which type of lens also.
 
Last edited:
Having read all this, it sounds like a very good reason for buying Nikon.
 
Thanks to everyone who commented, I think I would have to leave it to the experts as I don't think I could fiddle with MFA, I would probably muck it up myself considering I managed to scratch the focussing screen with my rocketblower! DOH!!
 
Coaster said:
Having read all this, it sounds like a very good reason for buying Nikon.
Click to expand...

I thought the newest Nikons had their equivalent of MFA? And, if they don't it means either -

1. Nikon can defy the laws of the universe and manufacture their gear to infinite accuracy. Or,

2. Nikonians have to put up with mis-matches in tolerance between lens and camera.

In any of these cases, I can't see what the advantage of Nikon is?
 
You must log in or register to reply here.
Back
Top