How to check infinity focus?

StewartR

StewartR

Messages
11,513
Name
Stewart
Edit My Images
Yes
We encountered a technical conundrum in the office yesterday that's got us stumped. I wonder if anybody out there can suggest a solution?

The short version is that we want to find an easy, reliable, repeatable way of checking that a lens can focus to infinity.

Here's why. Recently a customer hired a lens (Canon EF 24mm f/1.4 L II USM) for a trip to Norway to photograph the aurora borealis. But when he got back, he told us that he hadn't been able to get any good pictures because the lens wouldn't focus to infinity. The two professionals leading the trip he was on agreed with the diagnosis.

We were surprised because we test every lens every time it comes back to us, and we had never noticed anything wrong with this lens. So we started to investigate, but after half an hour of head scratching we realised that we were struggling.

Out of our office window the most distant object we can see is a factory chimney which is about 400m away. We photographed it with the suspect lens and with another identical lens which we knew to be good. (Ironically, because another customer had taken it on the very same Norway trip and had had no problems.) We took photos with the lens focussed manually at infinity, and then focussed automatically on the chimney (starting from the infinity stop), and pixel-peeped the results.

Here's what we got:
- suspect lens, manual focus at infinity - sharp image
- suspect lens, auto focus - sharp image (and no discernable AF movement)
- good lens, manual focus at infinity - soft image
- good lens, auto focus - sharp image (and a slight AF movement)

What it looks like is that when the suspect lens is manually focussed at what we think ought to be infinity, its actually focussed at some medium distance of several hundred metres. That would explain why the AF doesn't move when the lens focusses on the chimney, and it would explain why the manually focussed image is sharp.

But - we haven't actually proven whether or not the lens can actually focus at infinity. And we don't have an infinitely distant subject in or near the office to test it with. We tried using clouds but there's not enough contrast there to do a reliable test. If the moon had been out we could have tried that, but that wouldn't be a procedure which we could employ day-in day-out.

Any suggestions?
 
Last edited:
SussexBlue said:
This thread may help Stewart http://forum.mflenses.com/viewtopic.php?t=23534&view=next
Click to expand...
Thanks for the suggestion. But I'm afraid I can't see anything in that thread that's particularly helpful to my situation. Was there something specific in it that you thought was relevant?

One of the contributors to that thread suggests that focussing on an object 400m away is effectively the same as focussing at infinity (He says: "If there's still focusing range left, your set-up goes beyond infinity.") But that's actually the opposite of the problem I have - on my "good" lens there is a difference between focussing at 400m and focussing at infinity, but on my "suspect" lens there isn't. So if I took his advice I'd conclude that my good lens was bad and my bad lens was good.....
 
Have you's tested the lens at night? It's a daft question, but that's when the customer was primarily using it.I've had a problem in past when shooting at night when I think the lens is focused, but when I upload to my pc it's out of focus. I really have no-one else to blame bar myself and a general user error for using Af at night,when I should be using manual.

Another theory, and a valid one(i hope) is could it be an issue with the customers Camera? By which I mean could he had used the inbuilt micro adjustment software that came with the camera(if the camera has this ability). Or alternatively, could the customer have a problem with his camera?

I would have thought that if a lens isnt sharp and focusing on one camera, it wont do it on another. But you've seemed to have proven that the lens is actually working.

Just a few thought's, Some of which may be wrong.
 
John Mc said:
Have you's tested the lens at night?
Click to expand...

I refer the honourable gentleman to my previous statement:
StewartR said:
The short version is that we want to find an easy, reliable, repeatable way of checking that a lens can focus to infinity.
Click to expand...

John Mc said:
I would have thought that if a lens isnt sharp and focusing on one camera, it wont do it on another. But you've seemed to have proven that the lens is actually working.
Click to expand...
No, I think you've got that the wrong way round. We have proven that something is wrong with the lens - because it seems to be focussed at ~400m when it ought to be focussed at infinity, and its behaviour is quite different from the other lens. So I don't think we need to worry about user error, camera error etc. But our basic problem remains:
StewartR said:
But - we haven't actually proven whether or not the lens can actually focus at infinity. And we don't have an infinitely distant subject in or near the office to test it with. We tried using clouds but there's not enough contrast there to do a reliable test. If the moon had been out we could have tried that, but that wouldn't be a procedure which we could employ day-in day-out.
Click to expand...
 
Last edited:
I'd shoot the moon, but as you say you can't rely on it. Where's the office?
 
Last edited:
LOL. If the office is at your posted location, Maidenhead, I can't think of any national TV masts that are near enough to use. Maybe a mobile phone mast?
 
The aurora are about 100km up - there isn't anywhere on Earth you can use to test focus at this distance. Infinity is a long way away.

The only objects at distances which are for all intents and purposes are infinity are the stars. The only way I can think of doing the test to see if the lens will focus that far is to manually focus on stars (it's a pain to do that as it's a question of tethering the camera and looking at the capture or trying to judge focus on a tiny screen...
 
Stewart,

Here's an indoor usable test setup for you although you'll need to do get a little data first.
Adding a short extension tube behind the lens will bring "infinity" inside your room with all but the longest focal length lenses. With something like a 24mm lens and a 12mm tube then "infinity" will be on your desk.
The setup will work best for lenses between 50mm and 200mm as the distance involved will allow for some small percentage errors when you measure. Short lengths will require a much better resolution on the scale whilst longer FL's will be down the hallway.

Get the gist of the idea?

Bob
 
eccles said:
LOL. If the office is at your posted location, Maidenhead, I can't think of any national TV masts that are near enough to use. Maybe a mobile phone mast?
Click to expand...

Oh, I see what you were after. Yes, we're in Maidenhead. We tell people it's in a secret underground bunker for security reasons, but the office is actually above ground and has windows.

But I've already answered your question:

StewartR said:
Out of our office window the most distant object we can see is a factory chimney which is about 400m away.
Click to expand...
 
Canon Bob said:
Stewart,

Here's an indoor usable test setup for you although you'll need to do get a little data first.
Adding a short extension tube behind the lens will bring "infinity" inside your room with all but the longest focal length lenses. With something like a 24mm lens and a 12mm tube then "infinity" will be on your desk.
The setup will work best for lenses between 50mm and 200mm as the distance involved will allow for some small percentage errors when you measure. Short lengths will require a much better resolution on the scale whilst longer FL's will be down the hallway.

Get the gist of the idea?

Bob
Click to expand...
Get the gist? I think so. Its certainly ingenious. I'll have to wrap a cold towel around my head and think through the practicalities....

Can you point me to the formulae involved, so I can work through the maths?
 
StewartR said:
....
Can you point me to the formulae involved, so I can work through the maths?
Click to expand...

I believe so Stewart, can you give me until later today....I've got a few bits to do but I'll sort something this evening (and do a few tests myself).
The bonus we have is that the FL of the lens is specified at infinity (maybe it's rounded up or down a little) but it eliminates the reductions in FL seen with modern retro-focussing designs.

Bob
 
Well, I've been trying a few things and come up with a small problem and not a little confusion.

The problem first....
The focal length of a lens is specified at infinity focus...a good thing for Stewart's tests. However, the stated FL is actually rounded up or down to the nearest recognised value. A 24mm lens might be 23mm or 25mm on anywhere in between. To calculate the new infinity distance with some rear extension means that we need to know an accurate figure....more so with shorter FL lenses as the extension has a far greater effect on the magnification. Extension of 12mm (the shortest I've got) has too large an effect on a 24mm lens. Working with a 300mm lens is much easier and meaningful (if not totally accurate) figures can be obtained. I worked with 24mm and 300mm as they're lengths that I have two or more primes that comply.

The confusion....
A 24mm lens focussed at any distance greater than 13.6m on a full frame body (21.4m on a 1.6 crop) should have an object at infinity within the region that is, by convention, sharply focussed.....ie, a huge available range of focus (the hyperfocal distance if you wish).
So why couldn't the hirer achiveve sharp images? I believe that it must be a registration distance problem rather than a focus issue per se.
In short, it is a depth of focus problem as depth of field should never be an issue at 24mm focussed towards the long end.

One further point concerns the internal calibration of the lens. This is performed using small links in a binary code which are termed BFCV links (BFCV=Best Focus Correction Value). I have no reason to believe that the 24/1.4II is any different in this respect and it could be that it has been calibrated close up and this is having an adverse effect at long distances...I would expect that the calibration procedure does specify the maximum value to be coded before issues are seen at the opposite end of the range...but who knows?

I'll play a little more over the weekend but I'm not so sure it's going to help with this lens.

Bob
 
Last edited:
I'm not sure I understood much of that, Bob, but I would have thought (as you seem to imply) that if something @ 400m was in focus, then so would something @ 100km on a wide angle lens.

Does this imply the 24mm was significantly front-focussing?

Phil
 
There's only one way to be certain, and that's to focus on an infinity subject. Which isn't that far away with a 24mm lens, even at f/1.4. I would have thought your 400m mast would be fine up to say, I dunno you'd have to check, but maybe 100mm? For anything longer or with a very low f/number, you'll have to jump in the car or wait for the moon.

The way they do it in labs is with a collimator (lens) which simulates a target at infinity, and sounds expensive.
 
philthejuggler said:
I'm not sure I understood much of that, Bob, but I would have thought (as you seem to imply) that if something @ 400m was in focus, then so would something @ 100km on a wide angle lens.
Click to expand...
Not strictly Phil but maybe I'm being picky on the wording.
Something at 400m can be in focus if the subject plane is 13.15m (full frame at f/1.4). If the subject plane is 13.6m then an object at infininty should comply with the normal rules of being acceptably sharp

philthejuggler said:
Does this imply the 24mm was significantly front-focussing?

Phil
Click to expand...
Either that or the error existed behind the lens.

Bob
 
Canon Bob said:
Something at 400m can be in focus if the subject plane is 13.15m (full frame at f/1.4).
Click to expand...
No. Something at 400m is in focus only if the subject plane is at 400m.

Canon Bob said:
If the subject plane is 13.6m then an object at infininty should comply with the normal rules of being acceptably sharp
Click to expand...
Yes. But we were pixel-peeping to check whether there was a difference between the two lenses. When you do that, conventional notions of "acceptable sharpness" don't apply.
 
John, did you read post 1 in this thread?
StewartR said:
... we want to find an easy, reliable, repeatable way of checking that a lens can focus to infinity.
Click to expand...

Maybe we have different ideas regarding what constitutes "easy", but there's no way I'm opening up a DSLR and taping a target over the sensor in order to follow this procedure.
 
StewartR said:
John, did you read post 1 in this thread?


Maybe we have different ideas regarding what constitutes "easy", but there's no way I'm opening up a DSLR and taping a target over the sensor in order to follow this procedure.
Click to expand...

Ok Stewart but perhaps you should re-read it, you use an old film camera.
Yes we do appear to have different ideas, I always believed you got thanked for trying to help.
 
a collimator is the same thing as a tube - cardboard (black) would do - all a collimator does is focus everything in to a small thin area - disclaimer - I'm saying this as a physicist not a photographer - I've never even seem a collimator lens - I just know what a collimator does :)

in other words same as what Bob suggested
 
hee hee @woof woof
 
SussexBlue said:
Yes we do appear to have different ideas, I always believed you got thanked for trying to help.
Click to expand...
You did.
StewartR said:
Thanks for the suggestion.
Click to expand...
And I do appreciate help, seriously. If I seemed less than gracious it's because I just can't / couldn't work out HOW this was supposed to help.

SussexBlue said:
Ok Stewart but perhaps you should re-read it, you use an old film camera.
Click to expand...
OK, I've re-read it. (It's not the most clearly written explanation - or, to be more charitable, it may be written for the benefit of people who are more familiar with these things than I am.) I think I understand the procedure now - I use an old film camera and tape the target behind the shutter mechanism of it, where the film would be.

But I'm really struggling to understand how to apply this to my problem. The procedure talks about disassembling the lens and putting it back together so that the target is sharp. Obviously I'm not going to do that. I guess I'm going to focus my lens at infinity and see what happens. If the target is sharp, then my lens focusses to infinity, and if it isn't then it doesn't ... is that it?

Is that it? If so, how sharp does it need to be? How accurate is the test - is it likely to be able to distinguish between a lens that can focus at infinity and a lens that can focus at 400m but not at infinity?

And here's another thought. Presumably the collimator lens needs to be tested first so that I know it's good, because if the collimator isn't spot on then the whole test is undermined. But how do I do test the collimator?
 
Nothing complicated. When you're on the roof you can generally see much further so there will be more likely to be something that requires infinity focus :) Sorry it's nothing actually that helpful!
 
Hi stewart!

I hired that very same lens from you for the very same purpose last January but the difference is that I was with a professional aurora photographer. This is what he had to say when I had the same issue.

My first 50 shots I discovered came out very soft and I had to keep adjusting off the infinity mark in order to gain focus his explanation was that due to the cold temperatures the focus range indicator on the lens body isn't truly accurate. So essentially what the range on the body says isn't necessarily corresponding to the internal of the lens.

He proved this to me by showing me his lens and it was off infinity by a few mm but he was producing sharp shots.

Later in the week we were in the warmth of the day taking photos of Tromso from the top of the cable car ride it was a lot warmer and we were both producing sharp shots set on infinity.

Is there a possibility the hirer was caught unaware with the effects of cold on the lens and maybe your strive to provide the best service you can is finding faults in perfectly functioning equipment?
 
Last edited:
You must log in or register to reply here.
Back
Top