Front Focusing

gman

gman

Messages
11,500
Name
Graham
Edit My Images
Yes
Could use a little advice as this is outside my comfort/knowledge zone.

My Tokina appears to be front focusing but I can get it pretty much bang on when manually focusing (see photos below).

Is this just the lens that needs calibrating or can the camera also be at fault? The latter wouldn't make sense as my Canon 50mm f1.8 works perfectly with both auto focusing and manual focusing on the same camera - but then this is where I could be missing some knowledge?

Any advice would be appreciated.


Cheers :)

Tokina%2080mm%20Auto.jpg


Tokina%2080mm%20Manual.jpg
 
From what you say and the images posted I agree if seems to be Front Focusing. Do I assume that your 5D doesn't have Micro adjust, in which case getting the lens calibrated is the only solution.

Before doing so though I would check with a decent vertical target at a reasonable distance first.
 
Cheers.

No micro-adjustment on the Canon 5D unfortunately. I still can't see how a camera can have any cause for this as I would have thought that all cameras of a particular brand should be setup to work properly with all lenses designed for that make of camera?

I see micro-adjustment as being a luxury feature for cater for those lenses which aren't quite flawless and saves them going off for calibration? But I don't know enough about this to know for sure.

It's still under warranty so will send it off for calibration. What is the vertical target test and how does this work?


:)
 
Focussing on a high contrast target that is flat in front of the lens with the ruler alongside, rather than shooting at 45 degrees to a ruler on the ground. Like this

images
 
Last edited:
Cool, will give it a shot. :thumbs:
 
I might be a thicko but if you carry out this test and find it is out then what do you do to correct it.?
 
I'm not sure if you can calibrate a lens yourself and it may have to be sent away. Unless you've got micro-adjustment in your camera which allows you to make the adjustments in the camera instead to compensate for the shift.

Credit to MPB Photographic, the lens is still under warranty and they've been very quick, polite and helpful with an immediate offer to have it calibrated. However, due to the mail problems I may just take it to AJ Johnstone in the city and front the cost myself to avoid it potentially going missing in transit.

That would be my choice though so no ill wish to MPB.
 
Hmmm, they won't give me a ball park figure over the phone but want to inspect it first. Fair enough but surely a rough idea of how much it costs to calibrate a lens can be provided?

Does anyone know?
 
my guess would be between £30-£60, but this is purely based on the Sigma tariff. of course, you don't know whether it's just a firmware/software fix or if a strip down is needed.... but based on my readings of the forum that sort of amount is common for a standard calibration.
 
gman said:
I still can't see how a camera can have any cause for this as I would have thought that all cameras of a particular brand should be setup to work properly with all lenses designed for that make of camera?

I see micro-adjustment as being a luxury feature for cater for those lenses which aren't quite flawless and saves them going off for calibration? But I don't know enough about this to know for sure.

It's still under warranty so will send it off for calibration. What is the vertical target test and how does this work?


:)
Click to expand...
You may find this article enlightening - http://www.lensrentals.com/news/2008.12.22/this-lens-is-soft-and-other-myths

May as well add this too - http://www.lensrentals.com/news/2010.03.06/this-lens-is-soft-and-other-facts

I haven't read them, but they have more articles here - http://www.lensrentals.com/news. I suspect many will be informative.
 
Last edited:
The problem with running a test like the one in the OP, is that you're actual focus point may not be exactly where the camera says it is. I would bet that the focus point you see in the viewfinder covers more than the area which is to be considered "in focus" and therefore, the point actually focused on is within the field of the visible focus point in the viewfinder.

I've had some issues with front/back focus - and found the best way to check it was as follows. Note - doing any of these types of tests at the closest focus distances is not a good idea since the lens probably isn't going to perform at it's best at either end of the focus spectrum.

Arrange 3 books (if you have 3 very similar books it's better in my opinion) staggered on a flat surface, one slightly behind and to the left or right of the previous one.

Focus on the text on the back (or front) of the middle book, and examine the results. You can move the front and rear books closer or further away if required.

I can't speak for Canon, but Nikon certainly do not like the "focus chart" or ruler type tests because of the different positions of the focus points above and because there is little contrast for the AF to pick up on.
 
OutLore said:
I can't speak for Canon, but Nikon certainly do not like the "focus chart" or ruler type tests because of the different positions of the focus points above and because there is little contrast for the AF to pick up on.
Click to expand...

Canon feels the same way. That's why I use a flat focus target, oriented parallel with the sensor plane, and an angled ruler from which to assess the exact positioning of the sharp plane of focus....

20100220_161721_0120_LR.jpg


You should not focus on a target which itself is at an angle to the camera.
 
They got back with an £80 + VAT unless additional work is required. The confirmed after looking at the example photos that there is definitely something wrong.

Thanks for all the tips, make sense to me to use a vertical target. I'll see if I can do it tonight :)

The thing about the focus point of the camera, were this not accurate then surely the 50mm lens would be out as well?

Cheers for the links, I'll have a read in a moment :D
 
gman said:
The thing about the focus point of the camera, were this not accurate then surely the 50mm lens would be out as well?
Click to expand...

It's not that the focus point is inaccurate, but when presented with a target at an angle the focus point has a choice of distances in the target from which it has to pick one distance at which to focus. Which should it choose? In practice the design of the focus system is such that it will favour whatever is closest to the camera. Furthermore, the actual focus points are larger than the focus point squares indicate. Let's say you aim exactly at the 30cm mark. The focus point square might line up with 29cm at the front edge and 31cm at the back edge. The actual focus point sensor might be able to see all the way from 28.5cm to 31.5cm. It will probably end up focusing at 28.5cm rather than the 30cm that you would have liked. You will think the camera/lens is front focusing. The truth is that it probably isn't but your test method is flawed if you use an angled target. To properly test AF accuracy you should present the camera with a well lit, clearly defined, unambiguous target that is at only one distance from the camera.

As for the 50/1.8, in my experience the AF performance is sufficiently vague and variable that the focus will rarely settle in the exact same place twice in a row. If you focus, then refocus, then refocus, then refocus you will probably find that the focus alternates back and forth between a couple of posiitons. Which is right? It beats me. In practice it means at best you have a 50:50 chance of the focus being where you want it to be, which is a real problem when shooting with wafer thin DOF.
 
Last edited:
Thanks for the response. I understand what you are saying but it seems too much of a coincidence to me that every shot I did with the Tokina is off and every shot I did with the Canon was fine? (there was a lot more photos, just didn't post them) I did some studio test shots also with the Tokina focusing clearly on the eyes and not one had sharp focus on the actual eyes.

It's probably best to do the vertical test and go from there I reckon! Cheers again :thumbs:


EDIT: I skimmed the first link but as said I've been taking studio test shots at F8, F11 etc and they are all soft. I can only obtain sharp focus by doing it manually. 50mm is fine though - this still leaves me with the assumption that it is the lens?
 
Last edited:
A consistent error is a good thing. It means everything is working but the calibration between body and lens is off. It's as simple as that and the first link I provided earlier should explain.

Unfortunately, with only one body and two lenses to judge with it's pretty hard to conclusively figure which has the problem. By the sound of it the likely villain is the Tokina, but it's not impossible that the Tokina is spot on and the the body is off, and the 50/1.8 is also off but in a way that negates any error in the body.

e.g. your body might have a -2 error and the 50/1.8 might have a +2 error, leading to perfect focus when the two are combined. However, the Tokina might have a -3 error, which combined with your body ends up with a -5 error, which is too much to ignore. Yet all the kit might be operating individually within acceptable manufacturing tolerance. If this were the case you could have the body adjusted and the Tokina might reach an acceptable level of performance, but the 50/1.8 would go just a little off, but still be usable.

On the other hand you might have a +5 body, a -5 Canon lens and the Tokina might be spot on, but sadly not with your +5 body. It would be daft to send the Tokina away for adjustment if the problem was with the body.

The problem is in determining exactly which equipment has what sort of error. Without proper measurement and/or a larger sample size you will struggle to know. This is why when I got fed up with microadjusting my bodies and lenses I just dumped the whole lot with Canon and asked them to sort it. And sort it they did. Everything within warranty was adjusted for free, but I had to pay for the lenses that were not in warranty. I felt it was worth it to have properly functioning equipment.
 
Last edited:
Ok now that makes more sense to me now! Sorry, I only skimmed the links as there was a lot there.

The Tokina is a £300 lens and whilst the images are great when I manually focus, £80+vat is a fair whack to pay compared to the price of the lens to get it calibrated. I think I may just take my chances with Royal Mail and return it to MPB for calibration under warranty.

In the meantime I'm on the lookout for an 85mm prime :)

Thanks for taking the time to explain it all, I enjoy learning something new so much appreciated :D
 
UPDATE:

Just thought I'd mention that MPB Photographic have just got in touch saying they'll front half of the cost of me getting the lens calibrated in Glasgow rather than sending it away to them.

This is after it was discussed that it may be even more beneficial to have the lens calibrated to my specific camera and also my concerns at posting a lens in the current postal climate.

Fair enough they probably save a little money and hassle here but it will also save me worry and time and they've responded to my concerns.

In my book this is top stuff and what customer service is all about. :thumbs:
 
I have one of these lenses and thankfully don't have any focussing issues. I did have a Tokina 24-200mm a while back which needed loking at so emailed the Tokina importers.
When you say it is under warranty I assume you mean under MPB warranty?

Below is a copy of the response I received:

Dear Michael,

We can have a look at the lens for you, £20 + VAT inspection charge and our standard repair charge is £84 + VAT plus the cost of any parts required.

As it is an older lens, there may be some issues with availability of spares.

If you would like us to have a look, please send your lens along with a note detailing the problem to the address below.

FAO Steve Trute
Kenro Ltd
Greenbridge Road
Swindon
Wiltshire
SN3 3LH
 
Yep it is MPB's own 6 month warranty which is not bad considering their prices are sometimes better than 2nd hand on here and ebay! lol
 
I purchased Lencal to check front/back focus. Doing it with home made devices is Ok providing its done correctly. First the target should be level with the camera for the first photo, then the target slanted at the correct angle for focus check. Also the distance the target is from the camera can make all the difference,this should be done at different distances for an accurate accessment. You shot at a focal length of 80mm in the pictures but not shown the same at 28mm which makes a difference.

A common error is to set the target closer than the minimum focus distance of the lens. I would advice different distance and different lens mm settings for a more accurate reading


I made this mistake as well at first which really threw the readings right out the window.

Realspeed
 
Last edited:
I've still not got round to doing the vertical focus. I did check at various focal lengths though. As said, it's out in real life studio shots also and unless I go manual I can't get the eyes in focus - it's only with this lens.

Will see how the calibration comes back and fingers crossed it sorts it as it really is a nice lens :)
 
Just thought I would update on this.

Took it into AJ Johnstone a few days ago and they've confirmed that the electronics were not working correctly, which is what was causing the focus problems. Phew! Wasn't just me then! lol

Straightforward repair at around £80 and will have it back in a week or so :D

Just to reiterate, MPB Photographic were good to their word and transferred half of this cost to me a few weeks ago as mutually agreed and promised. Top supplier in my books from the way they handled this from start to finish. :thumbs:
 
gman said:
Just to reiterate, MPB Photographic were good to their word and transferred half of this cost to me a few weeks ago as mutually agreed and promised. Top supplier in my books from the way they handled this from start to finish. :thumbs:
Click to expand...
:thumbs:

Mine too. They do a cracking job and I would have no problems buying anything of value form them.
 
You must log in or register to reply here.
Back
Top