OM-1 Autofocus Performance

[HicHic]

Is the OM-1ii significantly better than this or is it just as bad?

He says it's only a little bit better, but some others say it's a lot better. Who's wrong? I don't really trust his word for it because if you take a close look at the shot at 3:58, he is using those shots to show how the OM-1ii is able to get focus in that situation, but if you look closely, the eye is actually out of focus (that shot is back focused to the back of the wing), so maybe he is not that great at judging whether a shot is in focus or not. Maybe he only views the photos from a small size and doesn't zoom in to check critical focus.


Would love to hear those who have both cameras to chime in, or if someone with the OM-1ii can provide a burst series of photos showing the hit rate.

 

[HicHic]

This review too -

At 10:03, she says the AF "gets it 100% of the time", and shows a burst sequence of a puffin flying sideways to demonstrate this. If you slow down the video of those burst shots and view in 4K, you can see a significant amount of those shots are actually missing critical focus.



Is this the kind of result other OM-1ii users are getting too?




fail




fail




fail




fail




fail




fail




slightly off




critical focus achieved


fail




fail




fail




fail




fail


fail




fail




critical focus achieved

 

[sherman_levine]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

--

Sherm

Sherms flickr page

P950 album

P900 album RX10iv album
OM1.2 150-600 album

 

[HicHic]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

--

Sherm
Sherms flickr page

P950 album

P900 album RX10iv album
OM1.2 150-600 album

That is very alarming.

 

[3dpan]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

I'm in agreement with Sherman,
I have been wondering if the OM-1 (or others) use some form of predictive auto focus to allow for the (miniscule) time lag between focus command, focus motor response, and actual exposure.

Maybe the OM-1 has some form of predictive AF, but also maybe it is only effective on the PRO lenses that support 50fps tracking.

From memory (and correct me if I'm wrong) the Puffins were shot with the 150-600mm lens, which is only good for 25fps.

 

[HicHic]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

I'm in agreement with Sherman,
I have been wondering if the OM-1 (or others) use some form of predictive auto focus to allow for the (miniscule) time lag between focus command, focus motor response, and actual exposure.

Maybe the OM-1 has some form of predictive AF, but also maybe it is only effective on the PRO lenses that support 50fps tracking.

From memory (and correct me if I'm wrong) the Puffins were shot with the 150-600mm lens, which is only good for 25fps.

I only use pro lenses. 300/4, 150-400.

 

[3dpan]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

I'm in agreement with Sherman,
I have been wondering if the OM-1 (or others) use some form of predictive auto focus to allow for the (miniscule) time lag between focus command, focus motor response, and actual exposure.

Maybe the OM-1 has some form of predictive AF, but also maybe it is only effective on the PRO lenses that support 50fps tracking.

From memory (and correct me if I'm wrong) the Puffins were shot with the 150-600mm lens, which is only good for 25fps.

I only use pro lenses. 300/4, 150-400.

Correcting my previous post, in your reference video, Emily seems to have been shooting with "Big White".
In a later video she uses the 150-600mm.

So, my theory of Pro lens or not does not seem to be the answer.

 

[Jeroen1969]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

I'm in agreement with Sherman,
I have been wondering if the OM-1 (or others) use some form of predictive auto focus to allow for the (miniscule) time lag between focus command, focus motor response, and actual exposure.

Maybe the OM-1 has some form of predictive AF, but also maybe it is only effective on the PRO lenses that support 50fps tracking.

From memory (and correct me if I'm wrong) the Puffins were shot with the 150-600mm lens, which is only good for 25fps.

She did use the 150-600mm for that shot. She made another video using the MK 2 and The Big White where she shot puffins again.

Given the movement of the flying puffins, I think not getting a 100% perfect critical focus is acceptable. The nice thing about the burst mode (in raw) is one can probably harvest multiple perfect focussed shots, but not the full 100%, because if that was the case, why would one need a burst function shooting so many frames per second? To catch sudden movement?

 

[Jeroen1969]

Is the OM-1ii significantly better than this or is it just as bad?

He says it's only a little bit better, but some others say it's a lot better. Who's wrong? I don't really trust his word for it because if you take a close look at the shot at 3:58, he is using those shots to show how the OM-1ii is able to get focus in that situation, but if you look closely, the eye is actually out of focus (that shot is back focused to the back of the wing), so maybe he is not that great at judging whether a shot is in focus or not. Maybe he only views the photos from a small size and doesn't zoom in to check critical focus.

Would love to hear those who have both cameras to chime in, or if someone with the OM-1ii can provide a burst series of photos showing the hit rate.

I have both. Non structured investigation gives me the impression the MK II is slightly better. Bird AI is surely a lot better with the MK II.

 

[drj3]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

All the green focus point indicates is where in the frame the camera last focused when it focused before exposure. If the target is not in exactly the same position at exposure, then the point will not be correctly placed on the target. It is a location on the frame, not a location on the target.

The position of the target at exposure is a function of target movement, camera/lens movement, and panning accuracy.

So, if the target movement (speed/up/down/direction) does not change, no body parts (wings) change position and you as a photographer moves at exactly the same rate as the target (with no up/down/direction differences) panning perfectly, then the green dot will be in the correct position on the target.

In practice, the green dot will be on exactly the same point on the target only it the camera/lens in on a sturdy tripod focusing on a completely stationary target. In all other cases its accuracy depends on the target and photographers' movements.

 

[MEDISN]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

Happens with every camera. Yesterday with the A1...





It appeared to be tracking the player in the EVF but 16 out of a 20 frame burst were focused on the netting. By the time it focused properly the play was over.







Frustrating but no camera is perfect.

 

[sherman_levine]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

All the green focus point indicates is where in the frame the camera last focused when it focused before exposure. If the target is not in exactly the same position at exposure, then the point will not be correctly placed on the target. It is a location on the frame, not a location on the target.

The position of the target at exposure is a function of target movement, camera/lens movement, and panning accuracy.

So, if the target movement (speed/up/down/direction) does not change, no body parts (wings) change position and you as a photographer moves at exactly the same rate as the target (with no up/down/direction differences) panning perfectly, then the green dot will be in the correct position on the target.

In practice, the green dot will be on exactly the same point on the target only it the camera/lens in on a sturdy tripod focusing on a completely stationary target. In all other cases its accuracy depends on the target and photographers' movements.

The concern (at least my concern) is that if the subject-to-camera distance calculated at the time the bird was in the green square position is different from the actual distance at exposure time, then the bird will be out of focus.

If the bird was flying directly at the camera in the center of the capture field then the green square would be on the bird, but the bird would still be out of focus.

"Perfect" focus would occur only when the bird was flying across the frame at constant distance from the camera.

--

Sherm

Sherms flickr page

P950 album

P900 album RX10iv album
OM1.2 150-600 album

 

[drj3]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

All the green focus point indicates is where in the frame the camera last focused when it focused before exposure. If the target is not in exactly the same position at exposure, then the point will not be correctly placed on the target. It is a location on the frame, not a location on the target.

The position of the target at exposure is a function of target movement, camera/lens movement, and panning accuracy.

So, if the target movement (speed/up/down/direction) does not change, no body parts (wings) change position and you as a photographer moves at exactly the same rate as the target (with no up/down/direction differences) panning perfectly, then the green dot will be in the correct position on the target.

In practice, the green dot will be on exactly the same point on the target only it the camera/lens in on a sturdy tripod focusing on a completely stationary target. In all other cases its accuracy depends on the target and photographers' movements.

The concern (at least my concern) is that if the subject-to-camera distance calculated at the time the bird was in the green square position is different from the actual distance at exposure time, then the bird will be out of focus.

If the bird was flying directly at the camera in the center of the capture field then the green square would be on the bird, but the bird would still be out of focus.

"Perfect" focus would occur only when the bird was flying across the frame at constant distance from the camera.

If the distance traveled at exposure since the last focus in greater than the depth of field, then the image could be out of focus. This will depend at least partially on whether the camera simply focuses each frame without any predictive information and how far the target travels during the short interval between focus and exposure.

We do not know if the OM1s use a predictive focus algorithm (the E-M1s definitely did since Olympus tells us that) or simply uses focus of each frame. The advantage of predictive focus is that it can focus on a target moving in any direction as long as there is a constant distance change between frames.

The disadvantage of a predictive focus algorithm is that it will result in a significant number of unfocused frames when the distance traveled changes. My E-M1s would require 3-7 frames to correct focus for very fast-moving targets which changed direction/distance.

I think the OM1s use some type of predictive focus algorithm, but one that may be overridden by any actual large focus distance change between frames (based on the setting C=AF sensitivity).

The reason I think there is at least some predictive algorithm is that I have sequences of flying purple martins flying directly toward the camera at a decreasing speed while creating a reduced depth of field which remain in focus. However, when there is a quick change in distance with no prior distance changes (small bird takeoff flying directly at the camera) the camera has problems focusing on the initial images when there is no information on which to base predicted distance.

 

[plantdoc]

I do not shoot BIF but I have noticed CAF, E shutter and slower burst showing AF variability with even a still target. Used a Oly M1 mk3 and Oly OM1 mk1 with 75-300mm and 100-400mm and Lumix G95 with Lumix 100-300. Increasing SS to very high seems to make little difference. All the listing combos show similar behavior. Moving targets seem slightly worse. Do not understand the AF technology to comment, Maybe interaction with IS? I am not steady enough to frame the subject without IS. I just select the best pic out of the burst.

Greg,

 

[drj3]

I do not shoot BIF but I have noticed CAF, E shutter and slower burst showing AF variability with even a still target. Used a Oly M1 mk3 and Oly OM1 mk1 with 75-300mm and 100-400mm and Lumix G95 with Lumix 100-300. Increasing SS to very high seems to make little difference. All the listing combos show similar behavior. Moving targets seem slightly worse. Do not understand the AF technology to comment, Maybe interaction with IS? I am not steady enough to frame the subject without IS. I just select the best pic out of the burst.

Greg,

If you want to measure camera focus variability with a specific lens, then use a study tripod with remote release. If you want to measure total camera/lens variability plus photographer focus variability, hand hold the camera and shoot a stationary target. Shooting moving target includes focus variability produced by the camera/lens, photographer and subject movement.

Can IS affect focus accuracy?

Statement from OM Systems Tech on why the firmware 1.4 update of my OM1.1 would correct focus accuracy (and it did).

With the firmware update, the CAF and SAF anomalies were corrected.

This correction gives the lens more accurate information in terms of faster focusing and better positioning of the image stabilizer

--
drj3

 

[HicHic]

Given the movement of the flying puffins, I think not getting a 100% perfect critical focus is acceptable.

The flight movement was sideways. That's the easiest BIF situation because the focus distance is barely changing. The hardest is when the subject is quickly flying towards the camera, or having erratic changes in direction. Flying sideways in a straight line is by far the easiest situation for any camera.

"Not getting 100%" is an interesting way to describe a sub 20% success rate of critical focus.

 

[HicHic]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

Happens with every camera. Yesterday with the A1...



It appeared to be tracking the player in the EVF but 16 out of a 20 frame burst were focused on the netting. By the time it focused properly the play was over.



Frustrating but no camera is perfect.

I don't think anyone expects a perfect camera. Just something that has a decent hit rate, even with 75% I will be satisfied. I mean, 75% in critical focus. Look at the puffin shots, only 2/16. And it wasn't like your example where, when the A1 missed focus, it was because it focused on the net - at least it nailed focus on something. Look at the puffin shots, the ones that missed critical focus was focused on no man's land. Not on the puffin, not on the background - everything was fuzzy. So yes, while there is no "perfect camera", there is a huge difference between a hit rate that is closer to perfect, and one that's well below it.

 

[HicHic]

Is the OM-1ii significantly better than this or is it just as bad?

He says it's only a little bit better, but some others say it's a lot better. Who's wrong? I don't really trust his word for it because if you take a close look at the shot at 3:58, he is using those shots to show how the OM-1ii is able to get focus in that situation, but if you look closely, the eye is actually out of focus (that shot is back focused to the back of the wing), so maybe he is not that great at judging whether a shot is in focus or not. Maybe he only views the photos from a small size and doesn't zoom in to check critical focus.

Would love to hear those who have both cameras to chime in, or if someone with the OM-1ii can provide a burst series of photos showing the hit rate.

I have both. Non structured investigation gives me the impression the MK II is slightly better. Bird AI is surely a lot better with the MK II.

Interesting, Mike Lane says it is only marginally better, but you are saying "surely a lot better". Do you mind expanding more on this? In what way is it a lot better? By what metrics and how are you making this judgement? If possible, please share un-interrupted burst sequences to show the % hit rate of critically focused shots.

 

[MEDISN]

It appears that there can be significant subject movement between the autofocus calculation and the image exposure, if the focus point display can be taken seriously.

https://www.dpreview.com/forums/thread/4799519

That is very alarming.

Happens with every camera. Yesterday with the A1...



It appeared to be tracking the player in the EVF but 16 out of a 20 frame burst were focused on the netting. By the time it focused properly the play was over.



Frustrating but no camera is perfect.

I don't think anyone expects a perfect camera. Just something that has a decent hit rate, even with 75% I will be satisfied. I mean, 75% in critical focus. Look at the puffin shots, only 2/16. And it wasn't like your example where, when the A1 missed focus, it was because it focused on the net - at least it nailed focus on something. Look at the puffin shots, the ones that missed critical focus was focused on no man's land. Not on the puffin, not on the background - everything was fuzzy. So yes, while there is no "perfect camera", there is a huge difference between a hit rate that is closer to perfect, and one that's well below it.

Well the A1 didn't have a decent hit rate in that sequence. 4/20 were in focus, 2 of them I would say critical focus, the other two slightly front focused. My post was more about the disconnect between where a camera tells you it focused and where it actually focused. The A1 was telling me the subject was being tracked. The EXIF tells me it achieved focus confirmation (green box). In reality, it totally missed the goal shot in that sequence.

I see you posted heavily cropped images of a puffin above but you stripped the EXIF away so I can't see what settings you were using. That would be the first step to diagnosing any problem.

 

[HicHic]

I see you posted heavily cropped images of a puffin above but you stripped the EXIF away so I can't see what settings you were using. That would be the first step to diagnosing any problem.

Did you even read the post?