Discrepancy between focal point in EVF and recorded file when using adapted lens?

This strikes me as impossible, but is there any mechanism by which a Pentax 645 lens adapted to GFX can have an object in focus in the EVF, but the recorded file is slightly front focused?

I'm shooting wide open, using either full magnification or peaking (low) + magnification to focus on an object about 50m away, but when I open the file the sharpest area of the image is consistently a few meters in front of where I focus using the EVF.

The lens is stopped down to the taking aperture when focusing.

I'm not adjusting aperture or moving the camera in between focusing and taking.

The diopter is on the minimum setting and I have normal vision.

I'm not using my T/S adapter so there are no movements to worry about, and my understanding is that any issue with how the lens/adapter is interacting with the sensor would manifest in the EVF.

I've also been working on the assumption that the zone of acceptable focus is usually 1/3 in front of the focal point, 2/3 behind, which seems flipped around here.

What's going on? Could IBIS be playing a role? I haven't attempted to recreate with a tripod because I'm travelling without my tripods, perhaps should try again without IBIS.

Noddemix said:

This strikes me as impossible, but is there any mechanism by which a Pentax 645 lens adapted to GFX can have an object in focus in the EVF, but the recorded file is slightly front focused?

Click to expand...

No. The EVF image is the sensor image.

Noddemix said:

I'm shooting wide open, using either full magnification or peaking (low) + magnification to focus on an object about 50m away, but when I open the file the sharpest area of the image is consistently a few meters in front of where I focus using the EVF.

The lens is stopped down to the taking aperture when focusing.

Click to expand...

These last two paragraphs are contradictory. Is the lens wide open when shooting or are you stopped down from wide open? Not that it matters.

Noddemix said:

I'm not adjusting aperture or moving the camera in between focusing and taking.

The diopter is on the minimum setting and I have normal vision.

I'm not using my T/S adapter so there are no movements to worry about, and my understanding is that any issue with how the lens/adapter is interacting with the sensor would manifest in the EVF.

I've also been working on the assumption that the zone of acceptable focus is usually 1/3 in front of the focal point, 2/3 behind, which seems flipped around here.

What's going on? Could IBIS be playing a role? I haven't attempted to recreate with a tripod because I'm travelling without my tripods, perhaps should try again without IBIS.

Click to expand...

Rich42 said:

Noddemix said:

This strikes me as impossible, but is there any mechanism by which a Pentax 645 lens adapted to GFX can have an object in focus in the EVF, but the recorded file is slightly front focused?

Click to expand...

No. The EVF image is the sensor image.

Click to expand...

Exactly, it makes no sense. I'm certain I'm focusing accurately. The throw on this lens is a bit awkward but I'm adjusting it very carefully and taking the picture when the focal point is definitely the sharpest possible.

Rich42 said:

Rich42 said:

I'm shooting wide open, using either full magnification or peaking (low) + magnification to focus on an object about 50m away, but when I open the file the sharpest area of the image is consistently a few meters in front of where I focus using the EVF.

The lens is stopped down to the taking aperture when focusing.

Click to expand...

These last two paragraphs are contradictory. Is the lens wide open when shooting or are you stopped down from wide open? Not that it matters.

Click to expand...

Fair point - I've made a few frames at f5.6 where the issue just about persists. By 7.1/8 the zone of acceptable focus is such that I can't see an issue, very obvious at f4 however. Should have just said aperture is the same when focusing and taking.

Rich42 said:

Rich42 said:

I'm not adjusting aperture or moving the camera in between focusing and taking.

The diopter is on the minimum setting and I have normal vision.

I'm not using my T/S adapter so there are no movements to worry about, and my understanding is that any issue with how the lens/adapter is interacting with the sensor would manifest in the EVF.

I've also been working on the assumption that the zone of acceptable focus is usually 1/3 in front of the focal point, 2/3 behind, which seems flipped around here.

What's going on? Could IBIS be playing a role? I haven't attempted to recreate with a tripod because I'm travelling without my tripods, perhaps should try again without IBIS.

Click to expand...

Click to expand...

I'm not a GFX owner but I've been adapting lenses to mirrorless cameras since 2013 and have never come across the kind of focusing discrepancy you describe. Across multiple brands/models/formats I've found EVF magnification (+ focus peaking when available) to be very accurate.

I'd defintely try turning off IBIS and seeing if your issue persists.

-Dave-

You say you're shooting wide open, but also that you close down to taking aperture.

Did you mean you are focusing wide open? I ask because I wonder if there's focus shift as you close down.

Rob de Loe said:

You say you're shooting wide open, but also that you close down to taking aperture.

Did you mean you are focusing wide open? I ask because I wonder if there's focus shift as you close down.

Click to expand...

Apologies - I clarified later. I also shot at f5.6/7.1 and meant that I was stopping down to the taking aperture when focusing in those scenarios, but obviously at f4 there was no need to stop down.

Noddemix said:

This strikes me as impossible, but is there any mechanism by which a Pentax 645 lens adapted to GFX can have an object in focus in the EVF, but the recorded file is slightly front focused?

I'm shooting wide open, using either full magnification or peaking (low) + magnification to focus on an object about 50m away, but when I open the file the sharpest area of the image is consistently a few meters in front of where I focus using the EVF.

The lens is stopped down to the taking aperture when focusing.

I'm not adjusting aperture or moving the camera in between focusing and taking.

The diopter is on the minimum setting and I have normal vision.

I'm not using my T/S adapter so there are no movements to worry about, and my understanding is that any issue with how the lens/adapter is interacting with the sensor would manifest in the EVF.

I've also been working on the assumption that the zone of acceptable focus is usually 1/3 in front of the focal point, 2/3 behind, which seems flipped around here.

What's going on? Could IBIS be playing a role? I haven't attempted to recreate with a tripod because I'm travelling without my tripods, perhaps should try again without IBIS.

Click to expand...

Disabling IBIS made no difference.

Outcome: it's the lens. If I focus meticulously on the centre of the frame, then use the joystick to look around, there are other areas of the frame that are sharper than the centre where I've focused. I'm used to dealing with field curvature but this is something different, and the Pentax-A 645 120mm f4 is known to have a fairly flat field in any case. This particular lens also performs very well when digitising negatives, so I'm not exactly sure what's going on here to make it perform so badly away from MFD/toward infinity and when open wide (improves beyond f5.6).

My copy of the lens performs noticeably worse at these sorts of distances than the 645 75mm wide open, and at f3.5 the 75mm is superb. What a great little lens that is.

I know Rob de Loe is familiar with both these lenses. Any ideas Rob? If there was an issue with this copy, doesn't it seem odd that it does so well at MFD on the copy stand?

I have the 120/4.

I haven’t seen such behavior.

Rich42 said:

Noddemix said:

This strikes me as impossible, but is there any mechanism by which a Pentax 645 lens adapted to GFX can have an object in focus in the EVF, but the recorded file is slightly front focused?

Click to expand...

No. The EVF image is the sensor image.

Rich42 said:

I'm shooting wide open, using either full magnification or peaking (low) + magnification to focus on an object about 50m away, but when I open the file the sharpest area of the image is consistently a few meters in front of where I focus using the EVF.

The lens is stopped down to the taking aperture when focusing.

Click to expand...

These last two paragraphs are contradictory. Is the lens wide open when shooting or are you stopped down from wide open? Not that it matters.

Click to expand...

If the focusing is done wide open, we could be looking at focus shift here.

David Kieltyka said:

I'd defintely try turning off IBIS and seeing if your issue persists.

Click to expand...

You can't completely turn off IBIS. There is no mechanical lock on this camera.

Noddemix said:

Disabling IBIS made no difference.

Outcome: it's the lens. If I focus meticulously on the centre of the frame, then use the joystick to look around, there are other areas of the frame that are sharper than the centre where I've focused.

Click to expand...

Aha!

Noddemix said:

I'm used to dealing with field curvature but this is something different, and the Pentax-A 645 120mm f4 is known to have a fairly flat field in any case. This particular lens also performs very well when digitising negatives, so I'm not exactly sure what's going on here to make it perform so badly away from MFD/toward infinity and when open wide (improves beyond f5.6).

My copy of the lens performs noticeably worse at these sorts of distances than the 645 75mm wide open, and at f3.5 the 75mm is superb. What a great little lens that is.

I know Rob de Loe is familiar with both these lenses. Any ideas Rob? If there was an issue with this copy, doesn't it seem odd that it does so well at MFD on the copy stand?

Click to expand...

It's hard to know without having the lens ad adapter in hand, so I'm now taking a guess...

When you're photographing the negatives, is the camera pointing straight down? And when you're photographing a distance subject, I presume the camera is horizontal?

If that's correct, one possibility is that there's a bit of play on the adapter, or possibly even the lens barrel itself. You're not seeing it when the lens is vertically oriented (photographing negative) because the lens is pointing straight down. However, for the distance subject, the lens is cantilevered off the mount. Sometimes that accounts for differences across the frame.

Have you tried the diagnostic routine of shooting the scene with the camera oriented one way, and then flipping the camera over to shoot the scene again -- then comparing to see if the problem moves around the frame?

Rob de Loe said:

It's hard to know without having the lens ad adapter in hand, so I'm now taking a guess...

When you're photographing the negatives, is the camera pointing straight down? And when you're photographing a distance subject, I presume the camera is horizontal?

If that's correct, one possibility is that there's a bit of play on the adapter, or possibly even the lens barrel itself. You're not seeing it when the lens is vertically oriented (photographing negative) because the lens is pointing straight down. However, for the distance subject, the lens is cantilevered off the mount. Sometimes that accounts for differences across the frame.

Have you tried the diagnostic routine of shooting the scene with the camera oriented one way, and then flipping the camera over to shoot the scene again -- then comparing to see if the problem moves around the frame?

Click to expand...

Thanks Rob, that's an interesting theory, could well be something like that. I'll give it a go tomorrow. I've tried a couple different adapters so the issue will presumably be with the lens itself.

It's a novelty to have the centre of the frame as sharp as is possible and see other portions of the frame significantly sharper. In the few decades I've been using cameras I've never seen anything quite like it. I'll post an example tomorrow as well.

Noddemix said:

Rob de Loe said:

It's hard to know without having the lens ad adapter in hand, so I'm now taking a guess...

When you're photographing the negatives, is the camera pointing straight down? And when you're photographing a distance subject, I presume the camera is horizontal?

If that's correct, one possibility is that there's a bit of play on the adapter, or possibly even the lens barrel itself. You're not seeing it when the lens is vertically oriented (photographing negative) because the lens is pointing straight down. However, for the distance subject, the lens is cantilevered off the mount. Sometimes that accounts for differences across the frame.

Have you tried the diagnostic routine of shooting the scene with the camera oriented one way, and then flipping the camera over to shoot the scene again -- then comparing to see if the problem moves around the frame?

Click to expand...

Thanks Rob, that's an interesting theory, could well be something like that. I'll give it a go tomorrow. I've tried a couple different adapters so the issue will presumably be with the lens itself.

It's a novelty to have the centre of the frame as sharp as is possible and see other portions of the frame significantly sharper. In the few decades I've been using cameras I've never seen anything quite like it. I'll post an example tomorrow as well.

Click to expand...

What you're describing really does sound like field curvature. Concave field curvature is typical, but convex is possible. Both can give you what you're describing.

I have one lens in particular that was especially vexing because the field curvature across the whole image circle was not concave. Instead, it had a wavy form, so as I shifted across the scene, it would improve and then degrade as I approached the edge of the image circle.

To make this all messier, field curvature does not have to be uniform. If the lens took a knock and something is slightly out of alignment, you could also be dealing with everyone's favourite problem: tilted elements. These lenses are many decades old, so who knows what happened to yours during its long life.

Rob de Loe said:

Noddemix said:

Rob de Loe said:

It's hard to know without having the lens ad adapter in hand, so I'm now taking a guess...

When you're photographing the negatives, is the camera pointing straight down? And when you're photographing a distance subject, I presume the camera is horizontal?

If that's correct, one possibility is that there's a bit of play on the adapter, or possibly even the lens barrel itself. You're not seeing it when the lens is vertically oriented (photographing negative) because the lens is pointing straight down. However, for the distance subject, the lens is cantilevered off the mount. Sometimes that accounts for differences across the frame.

Have you tried the diagnostic routine of shooting the scene with the camera oriented one way, and then flipping the camera over to shoot the scene again -- then comparing to see if the problem moves around the frame?

Click to expand...

Thanks Rob, that's an interesting theory, could well be something like that. I'll give it a go tomorrow. I've tried a couple different adapters so the issue will presumably be with the lens itself.

It's a novelty to have the centre of the frame as sharp as is possible and see other portions of the frame significantly sharper. In the few decades I've been using cameras I've never seen anything quite like it. I'll post an example tomorrow as well.

Click to expand...

What you're describing really does sound like field curvature. Concave field curvature is typical, but convex is possible. Both can give you what you're describing.

I have one lens in particular that was especially vexing because the field curvature across the whole image circle was not concave. Instead, it had a wavy form, so as I shifted across the scene, it would improve and then degrade as I approached the edge of the image circle.

To make this all messier, field curvature does not have to be uniform. If the lens took a knock and something is slightly out of alignment, you could also be dealing with everyone's favourite problem: tilted elements. These lenses are many decades old, so who knows what happened to yours during its long life.

Click to expand...

Here's what happens to the focal plane when field curvature and tilt are combined.









--

JimKasson said:

Rob de Loe said:

Noddemix said:

Rob de Loe said:

It's hard to know without having the lens ad adapter in hand, so I'm now taking a guess...

When you're photographing the negatives, is the camera pointing straight down? And when you're photographing a distance subject, I presume the camera is horizontal?

If that's correct, one possibility is that there's a bit of play on the adapter, or possibly even the lens barrel itself. You're not seeing it when the lens is vertically oriented (photographing negative) because the lens is pointing straight down. However, for the distance subject, the lens is cantilevered off the mount. Sometimes that accounts for differences across the frame.

Have you tried the diagnostic routine of shooting the scene with the camera oriented one way, and then flipping the camera over to shoot the scene again -- then comparing to see if the problem moves around the frame?

Click to expand...

Thanks Rob, that's an interesting theory, could well be something like that. I'll give it a go tomorrow. I've tried a couple different adapters so the issue will presumably be with the lens itself.

It's a novelty to have the centre of the frame as sharp as is possible and see other portions of the frame significantly sharper. In the few decades I've been using cameras I've never seen anything quite like it. I'll post an example tomorrow as well.

Click to expand...

What you're describing really does sound like field curvature. Concave field curvature is typical, but convex is possible. Both can give you what you're describing.

I have one lens in particular that was especially vexing because the field curvature across the whole image circle was not concave. Instead, it had a wavy form, so as I shifted across the scene, it would improve and then degrade as I approached the edge of the image circle.

To make this all messier, field curvature does not have to be uniform. If the lens took a knock and something is slightly out of alignment, you could also be dealing with everyone's favourite problem: tilted elements. These lenses are many decades old, so who knows what happened to yours during its long life.

Click to expand...

Here's what happens to the focal plane when field curvature and tilt are combined.



--
https://blog.kasson.com

Click to expand...

Could this combination result in a scenario where it's not possible to achieve critical focus at the centre of the frame (in this case)? That's the thing I find unusual - in the past with similar issues, either curvature or a tilted element, I've always been able to get a sharp centre but at the expense of some other portion of the frame.

SMC Pentax-A 645 Macro f4 120mm at f4

Some files I made in my garden yesterday. Green is where I focused, red is where critical focus seems to have been achieved. All at f4, all carefully focused in the centre of the frame. There was no way to get that portion of the frame any sharper than it is here.

The compression on these uploads makes it difficult to see, but it's obvious if you click "original size" and view at 100%.


SMC Pentax-A 645 Macro f4 120mm at f4


SMC Pentax-A 645 Macro f4 120mm at f4


SMC Pentax-A 645 Macro f4 120mm at f4

In the last one, the sharpest area is way in front of the hedge, let alone the nesting box.

Full files here:

https://drive.google.com/drive/folders/1P2zqP8tNiGJFIDDcX4qiGWH7O2DiAMGR?usp=drive_link

Proof that I'm able to manually focus a camera (from the Pentax 75mm and GF 120mm):


With the Pentax-A 75mm at f3.5, using the same Kipon adapter, and a 1/8 black mist filter I forgot to remove!


With the GF 120mm at f4, manually focused

Noddemix said:

JimKasson said:

Here's what happens to the focal plane when field curvature and tilt are combined.

Click to expand...

Could this combination result in a scenario where it's not possible to achieve critical focus at the centre of the frame (in this case)? That's the thing I find unusual - in the past with similar issues, either curvature or a tilted element, I've always been able to get a sharp centre but at the expense of some other portion of the frame.

Click to expand...

In my experience, no. With 100+ adapted lenses on various bodies over the years, I have never encountered a case where it was impossible to get the centre in focus. The problems I've had to address always involve sharp centre but problematic edge(s) and corner(s).

I downloaded your files and explored them at full resolution. Assuming you focused properly, there's definitely something strange going on. It's not obvious to me what the issue is because you're using 3D targets.

The least expensive option in terms of time at this stage is to buy another one. However, if you want to explore the issue further, I'd recommend a different kind of target. Can you find something that has these characteristics? If you can square up against a scene like this, you can check a lot of problems quickly by focusing on the centre of the line of the subject (bushes and trees here) and then checking across the frame.

• If you focus on centre of the bushes, are the bushes at left and right (and all along) equally in focus?
• Is the grass/field in front of the subject falling out of focus evenly across the frame?
• Are parts of the scene beyond the point of focus in better focus than the centre?

For an indoor test of flatness of field, do you have a bookcase you can set up carefully? In this setup, all the items on the shelf are in the same plane as the front of the bookcase (including the shelves that I pulled forward where that was possible).





And for maximum fun... if you want to check for flatness of field across the frame, you can set up a wall of Siemens Stars. This is overkill for your situation; I use this for evaluating adjustments to cell spacing. If the field isn't flat, or if elements are tilted, you'll know. To your previous question about centres, I have never had a lens where I couldn't get focus on centre. The problem is always edges and corners.

Thanks for your time Rob - I could definitely have picked a better scene as you can't really see how sharpness is distributed throughout the frame. I'll try again today using your post as a reference and see what I can come up with.

I've been reluctant to buy another as I've had bad luck sourcing a good copy of this lens. I bought one last year that was immaculate, looked like it had never been removed from the box, and yet it was completely de-centred, but at least in a way that made sense to me (unlike this one). I know Rich and I think left-eye(??) have both said their copies are superb, but I just haven't been able to find one yet.

I may try the Mamiya equivalent, or in fact there's a camera shop near me selling the -FA version of this. I have a Fotodiox adapter that I believe allows me to change the aperture on the FA lenses. Do you know if there's anything about the FA version that would mean it's not suitable for digitising negatives? I understand the A and FA are optically identical with the 7 group floating element design, but I wonder how the helicoid is compared to the A. It's rare that I hear people using the FA lenses on GFX