What's causing viewfinder focus to be slightly off?

[StewartR]

I bought a used DSLR the other day for use as a lens testing workhorse. But after a couple of days of using it, I've realised that it has a problem with the viewfinder. I could send it back for a refund, but before I do that I'd like to know whether it's anything I can fix myself. So I'd appreciate advice on what's causing it.

Here's the issue. When I use autofocus, the image in the viewfinder is not perfectly sharp. It's pretty good, but not perfect. I find that tweaking the focus manually, ever so slightly, makes the image noticeably sharper. It's quite consistent in that I need to focus the lens slightly closer than the autofocus does, which implies the autofocus is back-focusing slightly.

However, when I examine the images, it's clear that the autofocus image is perfectly sharp, and the manually focused image isn't (even though it looks better through the viewfinder).

So there's something slightly wrong with the viewfinder image. What might it be, and how might I fix it?

 

[T_J_G]

Has the camera got an adjustable eye piece for those who wear glasses etc?

 

[Mark twiglet]

What camera is it? My nikon’s have a dioper that’s adjustable

 

[PhilH04]

Dioptre adjustment?

 

[Mark twiglet]

What camera is it? My nikon has adjustment on it to adjust to my eyesight

 

[f/1.2]

As others have said probably the diopter used to run into this all the time as my missus wears glasses and would always forget to reset it and I would always forget to check.

Thankfully she has her own gear now.

 

[StewartR]

Has the camera got an adjustable eye piece for those who wear glasses etc?

What camera is it? My nikon’s have a dioper that’s adjustable

Dioptre adjustment?

What camera is it? My nikon has adjustment on it to adjust to my eyesight

As others have said probably the diopter used to run into this all the time as my missus wears glasses and would always forget to reset it and I would always forget to check.

No. It is not the eyepiece dioptre adjustment.

Sorry, I had hoped that would be obvious. When I tweak the focus manually, the image in the viewfinder is perfectly sharp.

 

[john.margetts]

Absolutely not the viewfinder dioptre adjustment. The camera lens produces an image on the focus screen which is either in focus or not. If it is not in focus, the viewfinder dioptre will have no effect. The viewfinder eye-piece can affect the screen being in focus, not the image on it.


If the image on the sensor is in focus but the image on the focus screen is not, the only possible reason is that the focus screen is in the wrong position. If the screen is a replaceable one, it is not seated properly. If it is not a replaceable one, it has been knocked out of position.

 

[john.margetts]

No. It is not the eyepiece dioptre adjustment.

Sorry, I had hoped that would be obvious. When I tweak the focus manually, the image in the viewfinder is perfectly sharp.

People do not like reading when they can be guessing!

 

[HoppyUK]

Absolutely not the viewfinder dioptre adjustment. The camera lens produces an image on the focus screen which is either in focus or not. If it is not in focus, the viewfinder dioptre will have no effect. The viewfinder eye-piece can affect the screen being in focus, not the image on it.


If the image on the sensor is in focus but the image on the focus screen is not, the only possible reason is that the focus screen is in the wrong position. If the screen is a replaceable one, it is not seated properly. If it is not a replaceable one, it has been knocked out of position.

^^^ This.

Edit: as you know Stewart, I've done a lot of this. I would never use the viewfinder image for critically accurate focusing. Use live view, magnified, and focus manually. Even live view auto-focus is not always 100% reliable.

And it's not easy! You'll often find that infinitesimally small movements of the ring can shift the critical point. The only way to be 100% accurate is to incorporate a focusing wedge into the image, something like the Spyder LensCal kind of thing, so you can see both the focused point and out-of-focus areas in front and behind, and check the final image.

Keep shooting distances realistic, and don't forget about field curvature and focus shift

 

[Mark twiglet]

Yes but what you are seeing in the viewfinder as sharp but on pc screen is out when you manualfocus!

Am I guessing correctly ?

 

[StewartR]

People do not like reading when they can be guessing!

Yes but what you are seeing in the viewfinder as sharp but on pc screen is out when you manualfocus!
Am I guessing correctly ?

Oh, the irony....

Mark, I hope you're not guessing. There's no need to guess.

Manual focus => viewfinder image sharp => photograph not quite sharp
Auto focus => viewfinder image not quite sharp => photograph sharp

 

[StewartR]

Absolutely not the viewfinder dioptre adjustment. The camera lens produces an image on the focus screen which is either in focus or not. If it is not in focus, the viewfinder dioptre will have no effect. The viewfinder eye-piece can affect the screen being in focus, not the image on it.

If the image on the sensor is in focus but the image on the focus screen is not, the only possible reason is that the focus screen is in the wrong position. If the screen is a replaceable one, it is not seated properly. If it is not a replaceable one, it has been knocked out of position.

^^^ This.

Thanks guys.

 

[HoppyUK]

Thanks guys.

See my edited comment above Stewart - crossed post

 

[Canon Bob]

I've had a couple of bodies where I've had to shim the focusing screen to get it equidistant with the sensor. The shims are the same shape as the periphery of the screen and come in various thicknesses.

Making sure that the screen is seated correctly would be the first thing to check.

Bob

 

[Mikesphotaes]

This happens when I use one of my lenses on my 30D but although OOF in the viewfinder, the photo is perfectly sharp.

Rest of my lenses are totally fine with the same camera.

 

[HoppyUK]

It could also be the sensor that's out of whack.

What you're seeing is the sensor and the focusing screen that are not perfectly aligned, so if it's a used camera that's taken a knock it could be either (or both!) though the focusing screen is the rather more likely culprit. Sensors are not solidly fixed and there is tiny movement with things like sensor-cleaning mechanisms.

The error manifests itself as a shift in depth of focus (not field) and that is miniscule - like small fractions of a mm with say a fast wide-angle. Longer lenses and higher f/numbers are much more tolerant. If you're able to examine the whole image area in detail, you'll probably find that it's sharp in some areas and out in others, in an irregular pattern (if it's a regular pattern, then that's probably field curvature).

 

[sk66]

If the image on the sensor is in focus but the image on the focus screen is not, the only possible reason is that the focus screen is in the wrong position. If the screen is a replaceable one, it is not seated properly. If it is not a replaceable one, it has been knocked out of position.

Yes and no... If the image is in focus on the sensor it will (should) be in focus on the focus screen.

But that does not mean it will be in focus for your eye(s). That is exactly the reason there is a diopter adjustment for the eyepiece in the first place. With less than perfect vision, and even with corrected vision, the distance to the focus screen may not work ideally for you. I wear glasses (corrected to 20/20 at distance/near) but I also need to adjust the diopter. I adjust the diopter for the non lens dependent items (meter/display).

If the diopter is adjusted and the recorded image/viewed image disagree, then there is an issue with the viewfinder/screen. I suppose it could also be a mirror alignment/distance issue as the mirror for the viewfinder is (effectively) separate from the mirror for the AF module.

 

[sk66]

It could also be the sensor that's out of whack.

IDTS... if it were, then the PDAF wouldn't agree with the resulting image.

 

[HoppyUK]

Yes and no... If the image is in focus on the sensor it will (should) be in focus on the focus screen.

But that does not mean it will be in focus for your eye(s). That is exactly the reason there is a diopter adjustment for the eyepiece in the first place. With less than perfect vision, and even with corrected vision, the distance to the focus screen may not work ideally for you. I wear glasses (corrected to 20/20 at distance/near) but I also need to adjust the diopter. I adjust the diopter for the non lens dependent items (meter/display).

If the diopter is adjusted and the recorded image/viewed image disagree, then there is an issue with the viewfinder/screen. I suppose it could also be a mirror alignment/distance issue as the mirror for the viewfinder is (effectively) separate from the mirror for the AF module.

If the eyepiece dioptre adjustment is out, then the viewfinder image will not look sharp at any time.

Good point about the mirror though, if that's not seating correctly it'd have the same effect as the focusing screen being out of true. Just thinking about that and the hammering it takes, probably fairly high on the list of possibles, but confusingly, quite likely intermittent.

 

[StewartR]

If the eyepiece dioptre adjustment is out, then the viewfinder image will not look sharp at any time.

Agreed. See post #7. I know you've read it but I suspect not everyone has.

 

[Canon Bob]

Good point about the mirror though, if that's not seating correctly it'd have the same effect as the focusing screen being out of true..

But then the sub-mirror would also be out of true and the PDAF would suffer similarly?

 

[StewartR]

It could also be the sensor that's out of whack.

IDTS... if it were, then the PDAF wouldn't agree with the resulting image.

If the diopter is adjusted and the recorded image/viewed image disagree, then there is an issue with the viewfinder/screen. I suppose it could also be a mirror alignment/distance issue as the mirror for the viewfinder is (effectively) separate from the mirror for the AF module.

OK, let's see if I understand.

We have three possible optical paths within the camera:

(1) Main mirror, focusing screen, viewfinder
(2) Sub mirror, autofocus sensor
(3) No mirror, imaging sensor​

(2) and (3) seem to be perfectly matched, so that suggests all the components in those paths are correctly aligned. If the sub mirror was out of whack, or the autofocus sensor, or the main sensor, then images which were deemed by the autofocus to be sharp would not be sharp. But they are.

(1) is the optical path which gives results conflicting with both (2) and (3). So that suggests the problem is either in the focusing screen or the main mirror. But if the main mirror is misaligned, the sub mirror is nevertheless fine, because (2) agrees with (3). I don't know how likely that is, how closely linked the two mirrors are.

My plan:

Firstly, check the focusing screen, remove and re-seat.

Secondly, if that doesn't work, check the main mirror, see whether there's anything interfering with it, maybe causing it to sit in not quite the right position, etc.

I'll report back. Thanks for all the help.

 

[StewartR]

But then the sub-mirror would also be out of true and the PDAF would suffer similarly?

I was writing my long post while you wrote this short one.

Is it possible for the main mirror to be out of true but the sub mirror to be OK? I don't know how they're connected / linked.

 

[sk66]

If the eyepiece dioptre adjustment is out, then the viewfinder image will not look sharp at any time.

Good point about the mirror though, if that's not seating correctly it'd have the same effect as the focusing screen being out of true. Just thinking about that and the hammering it takes, probably fairly high on the list of possibles, but confusingly, quite likely intermittent.

Hmm... I can adjust the diopter so that the text/display doesn't quite match the image sharpness, but that is probably more of an alignment issue with my glasses.

I have seen the mirror come loose before... it was only held on by double sided adhesive strips. The mirrors are *separately adjustable, but since this is a keep/return question I would be more inclined towards return.

*The adjustments might be to main/sub mirror as a unit and then to the PDAF module for agreement.

 

[HoppyUK]

Hmm... I can adjust the diopter so that the text/display doesn't quite match the image sharpness, but that is probably more of an alignment issue with my glasses.

I have seen the mirror come loose before... it was only held on by double sided adhesive strips. The mirrors are *separately adjustable, but since this is a keep/return question I would be more inclined towards return.

*The adjustments might be to main/sub mirror as a unit and then to the PDAF module for agreement.

Agree. Unless re-seating the focusing screen turns out to be a simple and conclusive fix, and they're non easily removed on most cameras, I'd return it.

 

[simonbarker]

People do not like reading when they can be guessing!

If you want to solve problems the best approach is to start with the basics and go from there, things like dioptre adjustment for example.

 

[StewartR]

If you want to solve problems the best approach is to start with the basics and go from there, things like dioptre adjustment for example.

If you want to solve problems the best approach is to start by understanding the problem. Please read post #7 and post #21.

 

[simonbarker]

If you want to solve problems the best approach is to start by understanding the problem. Please read post #7 and post #21.

I already did, you could try reading post #27 again as you seem to have missed my point entirely. Yes they gave you wrong advice but if you assume that certain things have been checked that's a good way to miss an easy solution.

 

[Canon Bob]

Is it possible for the main mirror to be out of true but the sub mirror to be OK? I don't know how they're connected / linked.

I guess it depends how the manufacturer has designed it. On the Canon bodies where I've look into this then the sub mirror is mounted on the main mirror frame and sits perpendicular until the main mirror raises....it then folds flat against the main mirror to clear the optical path to the sensor.

I believe that the famous 1DmkIII focusing issues were due to the main mirror (and hence the sub-mirror) not returning to a stable position during bursts.

 

[Box Brownie]

What body?

If it has a Pentamirror not a Pentaprism it is possible one (or more?) of the mirrors are slightly dislodged?

 

[john.margetts]

If you want to solve problems the best approach is to start with the basics and go from there, things like dioptre adjustment for example.

If manual focus will give a focused image on the focus screen, it cannot be the user's eyes or the viewfinder eye-piece that is causing the auto focused image to be out of focus. That was clearly stated in the opening post hence my comment about guessing rather than reading. Logic is the best approach if you want to solve problems

 

[Jim_Tod]

If I understand this correctly it's an issue with the mirror/ penta/ viewfinder pathway.

If you AF then the mirror/ submirror and sensor are actually OK but not looking that way in the vf- that probably rules out the phase detection route.

IF you MF then the mirror/ prism/ vf looks fine but produces an oof image on the sensor.

That would suggest to me that the penta/ vf end of the path might be out fractionally rather than the mirror.

 

[PhilH04]

No. It is not the eyepiece dioptre adjustment.

Sorry, I had hoped that would be obvious. When I tweak the focus manually, the image in the viewfinder is perfectly sharp.

People do not like reading when they can be guessing!

Guilty... apologies, simply asked to eliminate just in case, but then had to rush out so was unable to follow up...

 

[Jase]

Did you buy the camera new or is it used?

 

[Nod]

Used. (4th word in the OP...)