Why don't all lenses have infinity exactly on the infinity stop?

peterh337 · 01-12-2011, 02:37 AM

This seems to be common; one has to back off a little to get the true infinity focus.

I've been a keen photographer since the 1970s, with countless Olympus (Zuiko) lenses, and more recently Pentax lenses.

Most of them seem to have this problem.

It is a bit of a hassle, because a lot of the time you are trying to shoot something distant, the camera won't auto focus (for whatever reason) and you just want to turn the focus ring to the stop.

I take a lot of pics from a light aircraft and the windows sometimes prevent AF, and obviously infinity is exactly what is required. So I tend to shoot at F8 in the hope that that will be OK if one is trying to work fast. F4 (wide open 16-45 lens) definitely blurs the image if the lens is on the stop.

Tommot1965 · 01-12-2011, 02:40 AM

good question Peter..not that I have the answer...but its puzzled me too

simico · 01-12-2011, 02:58 AM

Most obvious reason: expansion under different temperatures. Then mechanical tolerances and the change of these over time and use.

Use the viewfinder and try to focus on the distant object using your eye.

stanic · 01-12-2011, 03:24 AM

I have the DA 35 macro ltd, and yes, I have to back off a little but the infinity mark is then aligned with the focus mark on the barrel
so there is no stop as used to be for infinity, but the infinity focus IS exactly on the infinity mark

roliver · 01-12-2011, 03:34 AM

I heard it was because they're easier to make in the factory that way, less preconfiguration I guess. And I suppose it's always better to be able to focus slightly past infinity than slightly before it!

newarts · 01-12-2011, 05:39 AM

A lens is made by mechanical processes, each of which have tolerances. Infinity focus on the other hand is a mathematically precise concept. Near infinity focus, very small changes in lens position cause very large changes in focus distance.

As an example, lets see how much of a lens position change is needed to focus at 1000 meteres compared to 2000 meters for a 50mm lens. Using thin lens formulas,

1/1000 + 1/S = 1/2000 +1/S'; hence (2000-1000)/(2000*1000) = (S'-S)/(S*S')
Since S ~ F = .05M

S'-S ~ 0.00125mm or about 1.25 micrometers!

That's a small distance indeed and can easily result from machining errors or even temperature changes.

It is so small that it brings up a related question: just what do we mean when we say that a lens is "focused at infinity"?

Dave

PS I believe that almost any lens' "infinity focus" limit can be adjusted internally somehow.

peterh337 · 01-12-2011, 06:18 AM

A small distance indeed, but equally this implies that an AF camera is going to have a very hard time achieving optimal focus, and is relying on the depth of field to get away with it.

Optics design was never a strong point for me

A while ago I tried to work out some optics to put around an image intensifier tube to make it into a 'telescope' and never finished the project.

Can anyone indicate what F number is likely to make the infinity focus point moot i.e. below the resolution of the lens or the camera? It would not suprise me if F8 did it.

newarts · 01-12-2011, 06:43 AM

Originally posted by peterh337

A small distance indeed, but equally this implies that an AF camera is going to have a very hard time achieving optimal focus, and is relying on the depth of field to get away with it.......

I think you are correct and that the tolerance on perfect focus likely relates closely to "depth of focus" (the equivalent of "depth of field" at the image plane.) I think I recall the total depth of focus is something like 4 times the Airy disk diameter*.

This implies that the total tolerance on focus would be similar; for a 50mm lens, this is about

tolerance ~ (+/-) 2*2.44*(F-Number)*0.55 micrometers for average light

At F:8 this is:

Tolerance ~ (+/-) 21 micrometers..... still quite small*.

Dave

EDIT *I've also seen references say the total depth of focus is 2 times the Airy disk diameter so the tolerance would be half as large.. Gaussian optics theory seems to give yet different results (depth of focus changes with f-number squared I think.)

RE-EDIT The above is related to how precise the focus can possibly be and not to what is "Good Enough" for practical photographic purposes, in which case "Circle of Confusion", should be substituted for Airy disk diameter. With this change,

focus position tolerance ~ (+/-) C*(F-Number)

For an APS-C sensor, C ~ .020mm, so for F:8,

focus position tolerance ~ (+/-) 0.02*(8) = (+/-) 0.16mm ..... quite sloppy in comparison, but still only .006 inch.

I apologize for dragging this out with informal, sloppy math & no references but it was interesting to me and I thought I'd use this opportunity to record & report the results. Reporting will help me remember & maybe elicit responses which can help us better understand the situation.

peterh337 · 01-12-2011, 08:04 AM

This is very interesting.

The unfortunate thing is that when using manual focus, the best thing one has is the tiny viewfinder image, and it is difficult to get that spot-on.

I recall, from my OM1/OM2/OM4 days, that a lot of people were changing the prisms. Is this still the case, and can it be done? It would be great to have the Olympus type prism where you align the two lines.

newarts · 01-12-2011, 08:30 AM

Originally posted by peterh337

This is very interesting.

The unfortunate thing is that when using manual focus, the best thing one has is the tiny viewfinder image, and it is difficult to get that spot-on.

I recall, from my OM1/OM2/OM4 days, that a lot of people were changing the prisms. Is this still the case, and can it be done? It would be great to have the Olympus type prism where you align the two lines.

It is easy to change the viewfinder screen to one that has a split prism focus indicator.

Many people, including me have successfully used the "Jinfinance" type screen available on ebay. Here's a thread discussing it:
https://www.pentaxforums.com/forums/pentax-beginners-corner-q/72081-jinfinanc...reen-good.html

Dave

sewebster · 01-12-2011, 10:37 AM

Originally posted by peterh337

Can anyone indicate what F number is likely to make the infinity focus point moot i.e. below the resolution of the lens or the camera? It would not suprise me if F8 did it.

A hard question to answer I think. Let's start with accidentally focussing closer than infinity (f=50mm for examples). The problem here is that, as has been pointed out, even a tiny turn of the focus ring causes a huge change in focus distance. If you accidentally focus at 1000 m instead of infinity, then even at f/1.4 the DOF goes from 80 m to infinity, so no problem. BUT there is basically no difference on the focus ring for 1000 m to infinity. A more realistic example perhaps is focussing at 10 m instead of infinity. Now, you might say that, whoa, 10 m is nowhere near infinity! But they aren't very far apart on the focus ring. In this case you'd need f/16 to get infinity in focus. f/8 wouldn't cut it.

As for focussing PAST infinity, which, I think, is what we are discussing more here, I don't know. I could probably figure it out eventually, but maybe someone with more optics expertise can let us know. I'm not sure it works the same way... but if it does, then unfortunately even turning a relatively small distance past the infinity point could make a big difference, and require stopping down a LOT to compensate.

normhead · 01-12-2011, 10:46 AM

Quote:

Can anyone indicate what F number is likely to make the infinity focus point moot i.e. below the resolution of the lens or the camera? It would not suprise me if F8 did it.

What you need is an old manual lens with a DOF guide on it. It used to be standard knowledge on every lens. In these days of auto_focus, is that even something anyone thinks about. IN the old days I would focus to find out how far away something was, and then use the DOF scale to position my lens accordingly. What the viewfinder told you about the focus was irrelevent unless it was really bright and you were looking with the manual DOF stop down preview thingy depressed. Auto focus isn't like that unless there's a setting that will put the back end of your DOF at infinity for the set F-stop in the AV setting. Someone write Pentax and suggest it. Tell them I'm going to switch to a Canon or a Nikon if they don't.

That's actually a good reason to go with older lenses and manual focusing.

Quote:

I recall, from my OM1/OM2/OM4 days, that a lot of people were changing the prisms. Is this still the case, and can it be done? It would be great to have the Olympus type prism where you align the two lines.

Ah the old split screen focusing circle in the middle of the view finder... the good old days. Came standard on every Pentax I had, until the *ist.

peterh337 · 01-12-2011, 10:47 AM

How to DSLRs autofocus?

I assumed they drive the focus motor while operating the sensor in video mode, and looking at the spectrum of the video coming out of the selected spot(s) on the sensor, and they drive the motor so that the high frequency components of the spectrum are maximised. Then you have the sharpest focus.

It will always involve a bit of hunting around the optimum point, but this will never be 'perfect'. At some stage the camera has to say 'this is good enough'.

sewebster · 01-12-2011, 10:57 AM

Originally posted by normhead

What you need is an old manual lens with a DOF guide on it. It used to be standard knowledge on every lens. In these days of auto_focus, is that even something anyone thinks about. IN the old days I would focus to find out how far away something was, and then use the DOF scale to position my lens accordingly. What the viewfinder told you about the focus was irrelevent unless it was really bright and you were looking with the manual DOF stop down preview thingy depressed. Auto focus isn't like that unless there's a setting that will put the back end of your DOF at infinity for the set F-stop in the AV setting. Someone write Pentax and suggest it. Tell them I'm going to switch to a Canon or a Nikon if they don't.

Yep, DOF scales are good. One thing to remember though is that the DOF scales for the old lenses are for 24x36mm frames on film, and with the APS-C sensors you get less DOF (all else being equal...). So you can't rely on the old DOF scales with an APS-C camera.

grainbelt · 01-12-2011, 11:13 AM

Originally posted by sewebster

Yep, DOF scales are good. One thing to remember though is that the DOF scales for the old lenses are for 24x36mm frames on film, and with the APS-C sensors you get less DOF (all else being equal...). So you can't rely on the old DOF scales with an APS-C camera.

This is correct and very important. I did some lazy street shooting using the distance scale on a manual focus 28mm lens, set to F*8 and the indicated hyperfocal distance, only to find that it was very significantly different for pentax digital SLRs.

For the OP, suggest a visit to DOFmaster website to make a quick table of hyperfocal distances for your most used focal lengths and apertures.

Hyperfocal Distance and Depth of Field Calculator - DOFMaster

It really doesn't have to be as complicated as all the math posted above makes it out to be.

01-12-2011, 02:37 AM	#1
peterh337 Veteran Member Join Date: Jan 2009 Location: London Posts: 573	Why don't all lenses have infinity exactly on the infinity stop? This seems to be common; one has to back off a little to get the true infinity focus. I've been a keen photographer since the 1970s, with countless Olympus (Zuiko) lenses, and more recently Pentax lenses. Most of them seem to have this problem. It is a bit of a hassle, because a lot of the time you are trying to shoot something distant, the camera won't auto focus (for whatever reason) and you just want to turn the focus ring to the stop. I take a lot of pics from a light aircraft and the windows sometimes prevent AF, and obviously infinity is exactly what is required. So I tend to shoot at F8 in the hope that that will be OK if one is trying to work fast. F4 (wide open 16-45 lens) definitely blurs the image if the lens is on the stop.

01-12-2011, 06:43 AM	#8
newarts Inactive Account Join Date: Dec 2008 Location: Ames, Iowa, USA Photos: Albums Posts: 2,965	Originally posted by peterh337 A small distance indeed, but equally this implies that an AF camera is going to have a very hard time achieving optimal focus, and is relying on the depth of field to get away with it....... I think you are correct and that the tolerance on perfect focus likely relates closely to "depth of focus" (the equivalent of "depth of field" at the image plane.) I think I recall the total depth of focus is something like 4 times the Airy disk diameter. This implies that the total tolerance on focus would be similar; for a 50mm lens, this is about tolerance ~ (+/-) 22.44(F-Number)0.55 micrometers for average light At F:8 this is: Tolerance ~ (+/-) 21 micrometers..... still quite small. Dave EDIT I've also seen references say the total depth of focus is 2 times the Airy disk diameter so the tolerance would be half as large.. Gaussian optics theory seems to give yet different results (depth of focus changes with f-number squared I think.) RE-EDIT The above is related to how precise the focus can possibly be and not to what is "Good Enough" for practical photographic purposes, in which case "Circle of Confusion", should be substituted for Airy disk diameter. With this change, focus position tolerance ~ (+/-) C(F-Number) For an APS-C sensor, C ~ .020mm, so for F:8, focus position tolerance ~ (+/-) 0.02(8) = (+/-) 0.16mm ..... quite sloppy in comparison, but still only .006 inch. I apologize for dragging this out with informal, sloppy math & no references but it was interesting to me and I thought I'd use this opportunity to record & report the results. Reporting will help me remember & maybe elicit responses which can help us better understand the situation. Last edited by newarts; 01-12-2011 at 08:57 AM.

01-12-2011, 10:46 AM	#12
normhead Pentaxian Join Date: Jun 2007 Location: Near Algonquin Park Photos: Gallery \| Albums Posts: 40,451	Quote: Can anyone indicate what F number is likely to make the infinity focus point moot i.e. below the resolution of the lens or the camera? It would not suprise me if F8 did it. What you need is an old manual lens with a DOF guide on it. It used to be standard knowledge on every lens. In these days of auto_focus, is that even something anyone thinks about. IN the old days I would focus to find out how far away something was, and then use the DOF scale to position my lens accordingly. What the viewfinder told you about the focus was irrelevent unless it was really bright and you were looking with the manual DOF stop down preview thingy depressed. Auto focus isn't like that unless there's a setting that will put the back end of your DOF at infinity for the set F-stop in the AV setting. Someone write Pentax and suggest it. Tell them I'm going to switch to a Canon or a Nikon if they don't. That's actually a good reason to go with older lenses and manual focusing. Quote: I recall, from my OM1/OM2/OM4 days, that a lot of people were changing the prisms. Is this still the case, and can it be done? It would be great to have the Olympus type prism where you align the two lines. Ah the old split screen focusing circle in the middle of the view finder... the good old days. Came standard on every Pentax I had, until the ist. Last edited by normhead; 01-12-2011 at 10:57 AM.*

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
k-7 blurry @ infinity w/ all lenses	wicked1	Pentax DSLR Discussion	31	09-25-2010 04:27 PM
adjustment of infinity focus on Kiron lenses	Lowell Goudge	Pentax SLR Lens Discussion	24	04-12-2010 05:19 PM
Kit lenses and Infinity	justDIY	Pentax SLR Lens Discussion	6	01-21-2010 03:42 PM
infinity focus adjustment on M42 lenses	Lowell Goudge	Pentax SLR Lens Discussion	11	08-20-2009 01:49 PM
Voigtlander SL lenses - Infinity focus question	thePiRaTE!!	Pentax SLR Lens Discussion	29	10-30-2007 10:22 AM

01-12-2011, 02:40 AM	#2
Tommot1965 Veteran Member Join Date: Sep 2010 Location: Perth Western Australia Photos: Gallery \| Albums Posts: 1,026	good question Peter..not that I have the answer...but its puzzled me too

01-12-2011, 02:58 AM	#3
simico Veteran Member Join Date: Jan 2009 Location: Budapest Posts: 821	Most obvious reason: expansion under different temperatures. Then mechanical tolerances and the change of these over time and use. Use the viewfinder and try to focus on the distant object using your eye.

01-12-2011, 03:24 AM	#4
stanic Veteran Member Join Date: Aug 2010 Location: Zakopane Photos: Gallery \| Albums Posts: 389	I have the DA 35 macro ltd, and yes, I have to back off a little but the infinity mark is then aligned with the focus mark on the barrel so there is no stop as used to be for infinity, but the infinity focus IS exactly on the infinity mark

01-12-2011, 03:34 AM	#5
roliver Junior Member Join Date: Mar 2010 Location: Southampton Posts: 42	I heard it was because they're easier to make in the factory that way, less preconfiguration I guess. And I suppose it's always better to be able to focus slightly past infinity than slightly before it!

01-12-2011, 05:39 AM	#6
newarts Inactive Account Join Date: Dec 2008 Location: Ames, Iowa, USA Photos: Albums Posts: 2,965	A lens is made by mechanical processes, each of which have tolerances. Infinity focus on the other hand is a mathematically precise concept. Near infinity focus, very small changes in lens position cause very large changes in focus distance. As an example, lets see how much of a lens position change is needed to focus at 1000 meteres compared to 2000 meters for a 50mm lens. Using thin lens formulas, 1/1000 + 1/S = 1/2000 +1/S'; hence (2000-1000)/(20001000) = (S'-S)/(SS') Since S ~ F = .05M S'-S ~ 0.00125mm or about 1.25 micrometers! That's a small distance indeed and can easily result from machining errors or even temperature changes. It is so small that it brings up a related question: just what do we mean when we say that a lens is "focused at infinity"? Dave PS I believe that almost any lens' "infinity focus" limit can be adjusted internally somehow.

01-12-2011, 06:18 AM	#7
peterh337 Veteran Member Join Date: Jan 2009 Location: London Posts: 573 Original Poster	A small distance indeed, but equally this implies that an AF camera is going to have a very hard time achieving optimal focus, and is relying on the depth of field to get away with it. Optics design was never a strong point for me A while ago I tried to work out some optics to put around an image intensifier tube to make it into a 'telescope' and never finished the project. Can anyone indicate what F number is likely to make the infinity focus point moot i.e. below the resolution of the lens or the camera? It would not suprise me if F8 did it.

01-12-2011, 08:04 AM	#9
peterh337 Veteran Member Join Date: Jan 2009 Location: London Posts: 573 Original Poster	This is very interesting. The unfortunate thing is that when using manual focus, the best thing one has is the tiny viewfinder image, and it is difficult to get that spot-on. I recall, from my OM1/OM2/OM4 days, that a lot of people were changing the prisms. Is this still the case, and can it be done? It would be great to have the Olympus type prism where you align the two lines.

01-12-2011, 08:30 AM	#10
newarts Inactive Account Join Date: Dec 2008 Location: Ames, Iowa, USA Photos: Albums Posts: 2,965	Originally posted by peterh337 This is very interesting. The unfortunate thing is that when using manual focus, the best thing one has is the tiny viewfinder image, and it is difficult to get that spot-on. I recall, from my OM1/OM2/OM4 days, that a lot of people were changing the prisms. Is this still the case, and can it be done? It would be great to have the Olympus type prism where you align the two lines. It is easy to change the viewfinder screen to one that has a split prism focus indicator. Many people, including me have successfully used the "Jinfinance" type screen available on ebay. Here's a thread discussing it: https://www.pentaxforums.com/forums/pentax-beginners-corner-q/72081-jinfinanc...reen-good.html Dave

01-12-2011, 10:37 AM	#11
sewebster Veteran Member Join Date: Jun 2008 Location: Vancouver, BC Photos: Gallery Posts: 544	Originally posted by peterh337 Can anyone indicate what F number is likely to make the infinity focus point moot i.e. below the resolution of the lens or the camera? It would not suprise me if F8 did it. A hard question to answer I think. Let's start with accidentally focussing closer than infinity (f=50mm for examples). The problem here is that, as has been pointed out, even a tiny turn of the focus ring causes a huge change in focus distance. If you accidentally focus at 1000 m instead of infinity, then even at f/1.4 the DOF goes from 80 m to infinity, so no problem. BUT there is basically no difference on the focus ring for 1000 m to infinity. A more realistic example perhaps is focussing at 10 m instead of infinity. Now, you might say that, whoa, 10 m is nowhere near infinity! But they aren't very far apart on the focus ring. In this case you'd need f/16 to get infinity in focus. f/8 wouldn't cut it. As for focussing PAST infinity, which, I think, is what we are discussing more here, I don't know. I could probably figure it out eventually, but maybe someone with more optics expertise can let us know. I'm not sure it works the same way... but if it does, then unfortunately even turning a relatively small distance past the infinity point could make a big difference, and require stopping down a LOT to compensate.

01-12-2011, 10:47 AM	#13
peterh337 Veteran Member Join Date: Jan 2009 Location: London Posts: 573 Original Poster	How to DSLRs autofocus? I assumed they drive the focus motor while operating the sensor in video mode, and looking at the spectrum of the video coming out of the selected spot(s) on the sensor, and they drive the motor so that the high frequency components of the spectrum are maximised. Then you have the sharpest focus. It will always involve a bit of hunting around the optimum point, but this will never be 'perfect'. At some stage the camera has to say 'this is good enough'.

Thread Tools	Search this Thread
	Search this Thread: Advanced Search

01-12-2011, 10:57 AM	#14
sewebster Veteran Member Join Date: Jun 2008 Location: Vancouver, BC Photos: Gallery Posts: 544	Originally posted by normhead What you need is an old manual lens with a DOF guide on it. It used to be standard knowledge on every lens. In these days of auto_focus, is that even something anyone thinks about. IN the old days I would focus to find out how far away something was, and then use the DOF scale to position my lens accordingly. What the viewfinder told you about the focus was irrelevent unless it was really bright and you were looking with the manual DOF stop down preview thingy depressed. Auto focus isn't like that unless there's a setting that will put the back end of your DOF at infinity for the set F-stop in the AV setting. Someone write Pentax and suggest it. Tell them I'm going to switch to a Canon or a Nikon if they don't. Yep, DOF scales are good. One thing to remember though is that the DOF scales for the old lenses are for 24x36mm frames on film, and with the APS-C sensors you get less DOF (all else being equal...). So you can't rely on the old DOF scales with an APS-C camera.

01-12-2011, 11:13 AM	#15
grainbelt Veteran Member Join Date: Nov 2008 Location: Minneapolis Photos: Albums Posts: 587	Originally posted by sewebster Yep, DOF scales are good. One thing to remember though is that the DOF scales for the old lenses are for 24x36mm frames on film, and with the APS-C sensors you get less DOF (all else being equal...). So you can't rely on the old DOF scales with an APS-C camera. This is correct and very important. I did some lazy street shooting using the distance scale on a manual focus 28mm lens, set to F*8 and the indicated hyperfocal distance, only to find that it was very significantly different for pentax digital SLRs. For the OP, suggest a visit to DOFmaster website to make a quick table of hyperfocal distances for your most used focal lengths and apertures. Hyperfocal Distance and Depth of Field Calculator - DOFMaster It really doesn't have to be as complicated as all the math posted above makes it out to be.