Help! Lens with back-focus?!
Posted 15/11/2005 - 21:42
Link
Yes, it can happen. I've read reports of some Sigma lenses showing this (on both Canon and Nikon). Most of the problems can be fixed with a chip-change in the Sigma...
...but now you come to mention it I'm slightly puzzled as to how you can get focus errors when the body is doing the control?! I'll look forward to any comments on this
Matt
...but now you come to mention it I'm slightly puzzled as to how you can get focus errors when the body is doing the control?! I'll look forward to any comments on this
Matt
Posted 16/11/2005 - 08:19
Link
I am so baffled by this that I didn't even describe the culprit! I'm talking about my new Tokina AF205 28-105mm F3.5-4.5. The body is my *ist DS.
In fact, the first time I used the lens at an event (my mother-in-law's birthday) all the photos taken with the Tokina had bizarre focus points... I thought I just had too much french champagne that day to be able to compose right.
Then I made a small test (wich I think I already mentioned here) were the Tokina surprised me by being very sharp, even wide open. It's very susceptible to flare and veiling, but I could live with that. Thing is, I made that test with all lenses manually focused.
But this time (in the international chocolate festival), I realized the lens really has a problem, although I cannot understand how on earth can this be; save for a few taken to maximize DOF, all pics were back-focused. Badly.
Yesterday, I made another small test: I went outside and auto-focused on some leaves in my backyard. The effect can be seen in naked-eye on the viewfinder: in 28mm, it never focuses the leaves I wanted! It seems to do a better job at bigger focal lengths; for example, at 105mm it seems to focus right.
Then I took the lens in and did a small head-to-head with two of my zooms: a Pentax SMC 18-55 and a Pentax 35-80. The test involved three objects at about 80cm from the camera, separated by 15cm each in a line going away from the camera. I took three shots, focusing on each object, and I wanted to see only one object in focus. The lenses were all wide open to minimize DOF.
The Pentaxes did marvelously well (I think the 18-55 is actually quite a good lens except at 18mm wide open where the corners are bof-bof, and the 35-80 actually surprises me with it's image quality), but the Tokina focused the wrong object all the time. I mean, when I focused the first it (more or less) focused the one behind it.
I can see it with my own eyes, but I still can't understand it! Any ideas?
I'll post some pictures later in the evening.
In fact, the first time I used the lens at an event (my mother-in-law's birthday) all the photos taken with the Tokina had bizarre focus points... I thought I just had too much french champagne that day to be able to compose right.
Then I made a small test (wich I think I already mentioned here) were the Tokina surprised me by being very sharp, even wide open. It's very susceptible to flare and veiling, but I could live with that. Thing is, I made that test with all lenses manually focused.
But this time (in the international chocolate festival), I realized the lens really has a problem, although I cannot understand how on earth can this be; save for a few taken to maximize DOF, all pics were back-focused. Badly.
Yesterday, I made another small test: I went outside and auto-focused on some leaves in my backyard. The effect can be seen in naked-eye on the viewfinder: in 28mm, it never focuses the leaves I wanted! It seems to do a better job at bigger focal lengths; for example, at 105mm it seems to focus right.
Then I took the lens in and did a small head-to-head with two of my zooms: a Pentax SMC 18-55 and a Pentax 35-80. The test involved three objects at about 80cm from the camera, separated by 15cm each in a line going away from the camera. I took three shots, focusing on each object, and I wanted to see only one object in focus. The lenses were all wide open to minimize DOF.
The Pentaxes did marvelously well (I think the 18-55 is actually quite a good lens except at 18mm wide open where the corners are bof-bof, and the 35-80 actually surprises me with it's image quality), but the Tokina focused the wrong object all the time. I mean, when I focused the first it (more or less) focused the one behind it.
I can see it with my own eyes, but I still can't understand it! Any ideas?
I'll post some pictures later in the evening.
Posted 16/11/2005 - 14:01
Link
This is very interesting and certainly something that I've never heard of before. Basically I think you're saying that the autofocus system is choosing a point slightly behind the subject, do you get focus confirmation (correctly on the subject) when you focus manually?
If you don't, then I suspect that the electronic contacts between camera and lens (or more specifically the chip within the lens) is somehow confusing the camera's AF system. I don't see how this can be, if you are using single-point AF with the subject perfectly positioned within the AF frame but I wonder if the lens is initially focusing correctly and then a pulse is being sent to the motor which causes the lens to shift slightly as you press the shutter release.
If this is the case, then it looks as though you've identified another example of an independant make of lens that is not quite 100% compatible.
As you don't have this problem with a Pentax lens, I suppose the only way to be sure would be to try another Tokina of the same generation but perhaps a different model.
I suggest a different model from the same range because a) they should communicate with the camera in the same way and b) the problem could be unique to your model because of it's construction (zoom/focus mechanism).
On the other hand, trying another identical lens could determine whether or not your particular lens has a fault and you've just been unlucky!
Do you have another AF body that you can try the lens on - perhaps a 35mm Z or MZ series, for example?
If the lens is a current model, perhaps a local dealer will allow you to try a few others from the range to see if it's a generic fault
If you don't, then I suspect that the electronic contacts between camera and lens (or more specifically the chip within the lens) is somehow confusing the camera's AF system. I don't see how this can be, if you are using single-point AF with the subject perfectly positioned within the AF frame but I wonder if the lens is initially focusing correctly and then a pulse is being sent to the motor which causes the lens to shift slightly as you press the shutter release.
If this is the case, then it looks as though you've identified another example of an independant make of lens that is not quite 100% compatible.
As you don't have this problem with a Pentax lens, I suppose the only way to be sure would be to try another Tokina of the same generation but perhaps a different model.
I suggest a different model from the same range because a) they should communicate with the camera in the same way and b) the problem could be unique to your model because of it's construction (zoom/focus mechanism).
On the other hand, trying another identical lens could determine whether or not your particular lens has a fault and you've just been unlucky!
Do you have another AF body that you can try the lens on - perhaps a 35mm Z or MZ series, for example?
If the lens is a current model, perhaps a local dealer will allow you to try a few others from the range to see if it's a generic fault
Die my dear doctor, that's the last thing I shall do!
Posted 16/11/2005 - 16:07
Link
Just remembered something (Lightbulb! Ding!)
Apparently the *ist-D firmware was designed to focus to the hyperfocal point. It uses information from the lens (aperture and length) to determine where that point is.
Now, if the lens inaccurately reports either the aperture or focal length then the body would mis-focus on what it believed was the hyperfocal point, and the AF would be wrong.
I suspect that if you manually focus the indicator should be correct (I would guess the body reverts to "regular" focusing).
Other things to check - are you using single, centre, focus?
If that fails, I would suggest trying another of the Tamron lenses (if that's possible).
Further to this point, I notice that the Pentax browser software (at least the one I have with the *ist-D) reports the 4-digit hex ID of the lens, and will map this to a lens name. I wonder whether your Tamron is being wrongly identified by the body as a Pentax lens, and the body is (wrongly) focussing it as if it was another lens. Perhaps you could check this in your EXIF information with the browser?
Matt
Apparently the *ist-D firmware was designed to focus to the hyperfocal point. It uses information from the lens (aperture and length) to determine where that point is.
Now, if the lens inaccurately reports either the aperture or focal length then the body would mis-focus on what it believed was the hyperfocal point, and the AF would be wrong.
I suspect that if you manually focus the indicator should be correct (I would guess the body reverts to "regular" focusing).
Other things to check - are you using single, centre, focus?
If that fails, I would suggest trying another of the Tamron lenses (if that's possible).
Further to this point, I notice that the Pentax browser software (at least the one I have with the *ist-D) reports the 4-digit hex ID of the lens, and will map this to a lens name. I wonder whether your Tamron is being wrongly identified by the body as a Pentax lens, and the body is (wrongly) focussing it as if it was another lens. Perhaps you could check this in your EXIF information with the browser?
Matt
Posted 16/11/2005 - 16:18
Link
Looks like we're thinking along the same lines Matt but you're more technically knowledgable than myself.
BTW, it's a Tokina lens not a Tamron!
BTW, it's a Tokina lens not a Tamron!
Die my dear doctor, that's the last thing I shall do!
Posted 16/11/2005 - 17:12
Link
Thanks for the correction Kimbo
Posted 16/11/2005 - 22:35
Link
Kimbo and Matt,
Yes, I get focus confirmation on manual focusing (on the wrong place), either by slowly focusing from back to front or from front to back. I'm using Select Focus Point (on the center, and I've tried Single Center, it's the same). And I kept my fingers on the (generous) focus ring; no movement is taking place after the autofocus beeps and during the shutter work.
The Pentax PhotoBrowser reports the lens as... well... as unkown. It reports "-- -- --" where it normally reports the full names of my Pentax lenses. The aperture and focal length are well reported on the EXIF info as F/3.5 and 28mm respectively.
It does a much better job at 105mm, autofocusing practically perfectly... this is really weird.
As for another AF body, I have a Pentax SFXn, but 28mm on the SFXn is too wide to be able to discern focus on the viewfinder... I can't really see if focus is right. I'll have to shoot some film and develop it to check.
As for local dealers... here in Portugal that's a bit difficult. I'll try to see if I can contact Tokina to see if there is something that can be done. Either that or www.foto-schneider.de where I bought the lens (via eBay).
Do you think I should upgrade my *ist DS' firmware? How can you see wich firmware mine has (didn't find anything on the net after a quick search).
Thanks!
Yes, I get focus confirmation on manual focusing (on the wrong place), either by slowly focusing from back to front or from front to back. I'm using Select Focus Point (on the center, and I've tried Single Center, it's the same). And I kept my fingers on the (generous) focus ring; no movement is taking place after the autofocus beeps and during the shutter work.
The Pentax PhotoBrowser reports the lens as... well... as unkown. It reports "-- -- --" where it normally reports the full names of my Pentax lenses. The aperture and focal length are well reported on the EXIF info as F/3.5 and 28mm respectively.
It does a much better job at 105mm, autofocusing practically perfectly... this is really weird.
As for another AF body, I have a Pentax SFXn, but 28mm on the SFXn is too wide to be able to discern focus on the viewfinder... I can't really see if focus is right. I'll have to shoot some film and develop it to check.
As for local dealers... here in Portugal that's a bit difficult. I'll try to see if I can contact Tokina to see if there is something that can be done. Either that or www.foto-schneider.de where I bought the lens (via eBay).
Do you think I should upgrade my *ist DS' firmware? How can you see wich firmware mine has (didn't find anything on the net after a quick search).
Thanks!
Posted 17/11/2005 - 04:02
Link
I forgot you had an SFXn
This is quite fascinating, I'm not sure about the firmware (I'd have to look into it myself) but I would imagine that your DS is working perfectly because it's ok with Pentax lenses and you have the same results when focusing manually. By that token, I would expect the same results with your SFX - so the fault must lie with the Tokina!
It's a fairly wide ranging zoom and although you say it's remarkably sharp at 28mm wide open, I suspect that there must be some serious abberations occuring in order to throw the phase detection out (if that's what's happening).
I look forward to any further developments - interesting stuff!
This is quite fascinating, I'm not sure about the firmware (I'd have to look into it myself) but I would imagine that your DS is working perfectly because it's ok with Pentax lenses and you have the same results when focusing manually. By that token, I would expect the same results with your SFX - so the fault must lie with the Tokina!
It's a fairly wide ranging zoom and although you say it's remarkably sharp at 28mm wide open, I suspect that there must be some serious abberations occuring in order to throw the phase detection out (if that's what's happening).
I look forward to any further developments - interesting stuff!
Die my dear doctor, that's the last thing I shall do!
Posted 17/11/2005 - 08:14
Link
Ok, take a look at these pics; they're cropped and scaled, but you'll be able to see the effect even at these scaled sizes (otherwise the pics would be gigantic to post).
From top to bottom, I tried to focus on the 28-80's numbers, then on the water bottle's barcode and then on the TV's remote "Sanyo" logo. Here's the 35-80's pics:
And now the Tokina's pics, focused on exactly the same places:
Those are all wide open, 1/30sec at 35 F/4 and 28 F/3.5 respectively. The conditions are not the best to test, I know. But if the differences can be seen so easily here, on better condittions it would be even more evident.
I've read some Sigmas have focusing problems. I've read about a Canon user who says Canon's AF system is... different in his own words, and the Sigma 18-200 does a lot of front-focus. He also says he can solve it by re-focusing on the same spot, but my DS doesn't move an inch when already "focused". In fact, it does this with every lens, and the Canon's autofocus was something I saw in a store while trying the 300D: the focus would hunt around if focusing on the same spot several times...
I'm still thinking that the DS' body should have full control over the focus... and I still don't understand where exactly is that the lens comes into the equation (unless Matt is right and the body calculates (erroneosly) the hyperfocal distance)...
I might try to update the firmware... where exactly can I see wich one I have?
Thanks!
From top to bottom, I tried to focus on the 28-80's numbers, then on the water bottle's barcode and then on the TV's remote "Sanyo" logo. Here's the 35-80's pics:
And now the Tokina's pics, focused on exactly the same places:
Those are all wide open, 1/30sec at 35 F/4 and 28 F/3.5 respectively. The conditions are not the best to test, I know. But if the differences can be seen so easily here, on better condittions it would be even more evident.
I've read some Sigmas have focusing problems. I've read about a Canon user who says Canon's AF system is... different in his own words, and the Sigma 18-200 does a lot of front-focus. He also says he can solve it by re-focusing on the same spot, but my DS doesn't move an inch when already "focused". In fact, it does this with every lens, and the Canon's autofocus was something I saw in a store while trying the 300D: the focus would hunt around if focusing on the same spot several times...
I'm still thinking that the DS' body should have full control over the focus... and I still don't understand where exactly is that the lens comes into the equation (unless Matt is right and the body calculates (erroneosly) the hyperfocal distance)...
I might try to update the firmware... where exactly can I see wich one I have?
Thanks!
Posted 17/11/2005 - 14:55
Link
Blimey!
Matt is correct! The body is trying to focus on the hyperfocal distance!
I gave it a quick try during lunch; I set the body on Aperture priority, wide open (F/3.5), point at some infinity (distant mountains) and my DS wasn't able to focus (it moved to infinity and made that sound of wanting to go further, stopping and blinking the autofocus confirmation hexagon). Then I changed the aperture to F/11, and the body focused to about 5 meters... and I was still pointing the lens at infinity.
The body is actually trying to focus on the hyperfocal distance, and failing miserably. When wide open, it probably is trying to back-focus again, but since there is no more room to go... it stops with the warning.
How did I never realize that the focus point was connected with the aperture before?! I don't remember reading about this anywhere...
I'll give it another go later in the evening when I arrive home.
P.S.: I was/am enamoured with a Tokina 28-70 F/2.6-2.8 , and I purchased the 28-105 as it was a cheap (59Euro) one to see the Tokina characteristics... now I'm afraid the same will happen with the more expensive lens...
Matt is correct! The body is trying to focus on the hyperfocal distance!
I gave it a quick try during lunch; I set the body on Aperture priority, wide open (F/3.5), point at some infinity (distant mountains) and my DS wasn't able to focus (it moved to infinity and made that sound of wanting to go further, stopping and blinking the autofocus confirmation hexagon). Then I changed the aperture to F/11, and the body focused to about 5 meters... and I was still pointing the lens at infinity.
The body is actually trying to focus on the hyperfocal distance, and failing miserably. When wide open, it probably is trying to back-focus again, but since there is no more room to go... it stops with the warning.
How did I never realize that the focus point was connected with the aperture before?! I don't remember reading about this anywhere...
I'll give it another go later in the evening when I arrive home.
P.S.: I was/am enamoured with a Tokina 28-70 F/2.6-2.8 , and I purchased the 28-105 as it was a cheap (59Euro) one to see the Tokina characteristics... now I'm afraid the same will happen with the more expensive lens...
Posted 17/11/2005 - 15:14
Link
Interesting, I'll see if I can simulate the same thing with my DL and a Tamron (manual focus) lens + 1.7x AF adapter. The camera wont 'know' what lens I'm using and will only have the aperture and adapter readings to go by.
I must admit that I have so far only used my Tamrons in manual mode but the focus confirmation has been spot on.
Well, I've just tried my 70-150mm f3.5 and 135mm f2.5 in various modes at both distant and nearby objects and there is no apparent focus shift whatsoever between max. and min. aperture settings.
Either the DL firmware is significantly different to your DS or the Tokina is causing the problem. You could use the lens exclusively in manual mode but that would defeat the object of having an AF lens, perhaps a firmware update is the solution!
I must admit that I have so far only used my Tamrons in manual mode but the focus confirmation has been spot on.
Well, I've just tried my 70-150mm f3.5 and 135mm f2.5 in various modes at both distant and nearby objects and there is no apparent focus shift whatsoever between max. and min. aperture settings.
Either the DL firmware is significantly different to your DS or the Tokina is causing the problem. You could use the lens exclusively in manual mode but that would defeat the object of having an AF lens, perhaps a firmware update is the solution!
Die my dear doctor, that's the last thing I shall do!
Posted 22/11/2005 - 22:37
Link
Kimbo, thanks for your time.
Indeed, you're right. None of my Pentax lenses exhibit the Tokina's simptoms. Changing the aperture does not change the focus point.
But I have been doing more tests and, at 28mm, the Tokina really changes the focus point as I change from F3.5 to F8. At 105mm however, focus seems to be spot on. In manual focus, the focus confirmation reacts in exactly the same way.
If I set the aperture at F8 and focus I always have good focusing, so I will be using it like this, and opening up after focusing if needs be.
I'll also upgrade my firmware and see if that makes any difference.
Thanks again!
Indeed, you're right. None of my Pentax lenses exhibit the Tokina's simptoms. Changing the aperture does not change the focus point.
But I have been doing more tests and, at 28mm, the Tokina really changes the focus point as I change from F3.5 to F8. At 105mm however, focus seems to be spot on. In manual focus, the focus confirmation reacts in exactly the same way.
If I set the aperture at F8 and focus I always have good focusing, so I will be using it like this, and opening up after focusing if needs be.
I'll also upgrade my firmware and see if that makes any difference.
Thanks again!
Posted 23/11/2005 - 06:25
Link
Flink
Apparently this problem is not that uncommon. I remember a few months ago a huge go-round on Luminous Landscape about the same thing happening with a guy's brand new Canon 20D as well as his warranty relacement. It got pretty heated with several people refusing to believe Canon was shipping flawed equipment and it must be operator error. Turns out Canon did have flawed equipment, lots of them (go figure). I also remember one guy was having trouble with his Canon 10D but ONLY with his Tokina lens.
I found this, not sure if the methodology is correct but maybe it could help you.
http://md.co.za/d70/focus21.pdf
Drew Peabody
Apparently this problem is not that uncommon. I remember a few months ago a huge go-round on Luminous Landscape about the same thing happening with a guy's brand new Canon 20D as well as his warranty relacement. It got pretty heated with several people refusing to believe Canon was shipping flawed equipment and it must be operator error. Turns out Canon did have flawed equipment, lots of them (go figure). I also remember one guy was having trouble with his Canon 10D but ONLY with his Tokina lens.
I found this, not sure if the methodology is correct but maybe it could help you.
http://md.co.za/d70/focus21.pdf
Drew Peabody
Posted 23/11/2005 - 07:42
Link
Flink
Has anyone answered your question about how to tell your current firmware version?
Check your firmware version by turning the camera on while holding down the MENU button. Your firmware version will be displayed on the LCD screen. Latest version for the *ist DS is V1.02 and can be downloaded from http://www.pentaximaging.com/customer_care/show_firmware?firmId=3
Drew Peabody
Has anyone answered your question about how to tell your current firmware version?
Check your firmware version by turning the camera on while holding down the MENU button. Your firmware version will be displayed on the LCD screen. Latest version for the *ist DS is V1.02 and can be downloaded from http://www.pentaximaging.com/customer_care/show_firmware?firmId=3
Drew Peabody
Add Comment
To leave a comment - Log in to Pentax User or create a new account.
151 posts
21 years
Vale de Cambra,
PORTUGAL
Is it possible for a lens to have back-focus? I mean, with the body in autofocus, the lens focuses consistently further away than the point the body marks as focused.
I tried several lenses on the same body, and only this one exibits this characteristic... I thought the SAFOX system worked by phase-matching, so there was no way a lens could not be focused right... unless the system had a problem, in wich case ALL lenses would focus wrong... and that is not my case.
Does this make any sense? Did anyone have similar experiences? Thanks!