Everyone, thanks for all the feedback! Using some of the suggestions, I got much better results today. I will post a follow-up video as soon as I am able.
I think the problem was the background! The em111 was looking at a much cleaner during the video test! The om1 was dealing with the leaves and trunk of the tree centered behind you! The em1 iii was dealing with a relatively clean background. Also different color grading might have thrown things off as the om1 looked flat and unsaturated! Just a few thoughts from the sidelines!
His testing situation is very typical shooting situation, not challenging; and something the others (Canon, Sony, Nikon) can easily do. So this is a concern. Enough for Rob is conclude "can't recommend". Hope this is an isolated case specific to this camera. OM-1 should do much better than M1iii. Regardless, as usual, Rob is giving us some really useful real world, no nonsense review.
@@dannygo4230 Yes but that is no serious argument. The easiest solution would have been to reverse the camera position. Not completely correct but would have eliminated the perspective factor provided e.g. the exposure did not change.
Interesting that the OM-1 seems very fast at finding your face and frame it, but can’t focus on it. Btw. I think you are a great teacher and I have learned a lot from watching your channel.
Thanks Rob, wow, that was a bad performance on the OM1. Can’t understand why it’s finds your face but not focusing on it? And also it did not find your eyes one time. Could you report this to OMDS? Hope they fix this soon
I suspect that it's because it's in LOG. LOG profiles in cameras have never been very good with AF until recently with Sony and Canon's advancements in tech. OM really has to modify the AF algorithm to properly read LOG or remake the LOG profile for 10-bit. I think this is also why the E-M1 III performed very well because it was in a more saturated picture profile.
@@charbax I can't say for sure until I get mine. But if I were to assume, HLG would perform much better and would most likely have better highlight retention than LOG. HLG is a Hybrid Log profile which work much better with 8bit and 10 bit/4-2-0. On my GH5s the AF is much better in HLG than LOG. It's nearly as accurate and responsive as using Cine D/Natural.
Hi Rob, This is the video I have been waiting for. I have an issue with the OM1 C-AF, face detect and the either (evil😀) eye if available, shooting relatively stationary interview subjects. Focus will lock, lose and hunt, rendering the footage unusable. I have been testing a number of different settings and the last attempt was the most successful: turning sensitivity -1 and speed to +1. Of course this is scenario specific and I am not yet convinced I have had a breakthrough. I have tried S-AF but I really do prefer a workable tracking solution. Really hope you can continue to research this (I have also raised this with Emilie).
Very interested to see how this finally works out. For me, AIF is The most important feature I look for in a new Olympus/OMS body. Content with the, now outdated AF, in my 2 Pen Fs. The type of shots I take don't require very speedy AF or AF tracking. If these are needed, particularly for people, or young children shots, I have no hesitation in borrowing a friend's Sony a6400 and using the magic Sigma 56 1.5- simply stellar and, if purchased, mint second hand, within my price range. Sony in FF and apsc has had AIF nailed for what must be the better part of ? 5 years'. Canon and Nikon have now caught up. OMS have had forever to build totally competent IAF- to at least match the far cheaper Sony a6400 and other bodies and it would be a major disappointment if the new OM1 did not, at least, match what some may regard as a more lowly a6400, let alone a 6100 etc. Complex features, such as focus stacking and the like are of little interest to me (others will strongly disagree), however, sharp quick, locking focus is fundamentally important to me. Am waiting for the flack to arrive- we all differ.
I have found nothing to match the AF of Sony. I rented a Canon R6 for a basketball job which the EM1III and EM1X could not handle. It was good enough but not as good as the Sonys are for that. I understand that OM Systems uses the AI subject detect to do what DSLRs used to do so easily. They still have a way to go with tracking although continuous autofocus is getting better in most conditions. Just wish they would add people as a subject because sometimes in sports, you want to get the person not the ball.
Hi Rob, when you removed the ND filter you increased the shutter to 1/750 of a sec? I thought you would shoot at 1/120th being twice 60fps? What aperture and shutter speed did you use on the M1 mark iii?
I don’t shoot video so I can’t wait for the photo examples to see if there is any improvement over my em1 mk3. If not I can stop worrying about buying om1 and stick with what I have. Thanks Rob. Loved your video.
I have also had problems with the eye detect autofocus in the picture mode on the OM-1 too. The green focus point was all over the face rather than the eye when reviewing the images in camera. It is definitely not as good as the M!X
Were both cameras using the same color profile and adjustments (contrast mainly) as for some reason this affected focus on the EM1s. Although with video you have less chances to improve focusing this way.
Interesting video but the settings seemed to be very different on the two cameras, please let the settings also be visible on the EM1 Miii. The camera position is very different, the OM1 has got much more light of the sky than the EM1. What a wonderful chair in the background.
This video made me do a video test with the E-M1X wearing a hat and glasses. Thankfully the focus locked on correctly. This behavior is surprising from the OM-1 considering it does recognize your face.
wow that is very interesting. Surly face detect would be easier solution compared to all the other detection modes.. Maybe a firmware update needed... Anyway it is what it is.. the main thing is we got T.J Hooker (William Shatner) on the case 13:25 🙂
You are the second professional that uses OMD-SYSTEM that does the continuous focus test that has the same focus problem, the other professional did it by taking pictures of a train and the focus was also lost, it was done by a Spaniard.
Sad AF. The only reason I was considering OM1 was for its PDAF which I thought would be much superior to the GH6 CAF. But if this is what one can expect, then it's unusable. And I want to film all videos handheld with the IBIS in 4K60 HLG 10bit HEVC to dual SD, I really need the AF to be absolutely 100% reliable. Like a Sony but in a micro43. Anyone know of any other OM1 reviews showing handheld IBIS AF performance in 4K60?
This is why you should be an official representative with OM. I’m sure they watch your blog (they best should) but we’d all be better served if you were hooked in directly to the company so you could give direct feedback and get these use case problems resolved.
Interested! But Rob i just wonder if that Oly 12~40mm f2.8 lens is not the problem.I recall you left it outside in snie etc and it could be suspicious Just my 2cent.
Thank you for this interesting but very unsatisfying result. Is it only in video-mode so bad and better in fotography-mode? The strangest thing to me, ist that the rectangle shows the detection of your face very well, but it remains unsharp. It also newer showed the recognition of an eye.
Interesting video. I set up my OM-1 this morning and tried to replicate your results. I was not able to do so. C-AF Video with straight face detection worked fine. C-AF Video with face and eye detection also worked fine. I tried with and without a ball cap much like what you had and it had no issues at all focusing on my face. I moved all around, forward / back / side to side / turned around, hid my face with the cap, etc. Each time AF locked on my face without issue. No idea what is the difference between your tests and mine. I had not even tried it until this AM. I also tried C-AF+TR with subject detection (I had it in the animal mode) in video and it had no trouble following my dog around the yard.
Closer to the lens, the face detection was no problem. Further away from the lens, the face detection did not perform at all. So, maybe he has set focus distance limit on the lens ?
thanx, Rob, for this comeback on the autofocus in video. i think in general log/flat profiles are worse in af, maybe because the hybrid system also uses the contrast-detect in addition to phase, and so contrast in log/flat profiles is harder to detect. but overall, yeah, at least according to this test, it does seem that om1's af in video is not reliable. i thought they really stepped up the game, but it seems although it's reliable in photos, it's not for videos.. sony works well even in flat profiles (and also in 24fps, which is harder on the system than 60fps, as far as i know, because in 60 the system has more data to work with, or something.. but if om1 does so bad in 60, i imagine its even worse in the cinematic 24)
When I borrowed an em1x I found the train tracking a mixed bag. Sometimes ok, other times it didn’t lock the face of locomotive. I’ve heard it was developed for bullet trains in Japan so may not be optimal for North American locomotives.
Hi Rob. Interesting. Maybe it's a lens. I have a completely different experience. Om1 is in my opinion more effective than m1miii. I did a similar test with a 17mm 1.2 wide open and the om-1 was definitely better in low light. In good light, i didn't see much of a difference. The OM1 recognizes the face a little faster, but also not every time. With 12-40 2.8 i didn't have any problems either. There is one more thing that affects autofocus, that is picture mode. With H265 10 bit om-log, h264 flat and om-log af is slower because because the contrast is reduced. It's the same with the M1miii (the autofocus with the natural profile works better than with flat and log). The difference, however, is not as big as in your test. I had some problems with tracking+caf with 12-40 2.8 and om-1 and m1. I had to turn off the camera and remove the lens. Sometimes 12-40 works louder on one camera than on another, and sometimes it loses autofocus. Also 8mm 1.8 works louder on m1mii than on m1miii and om1. 🤔 Ps.I hate variable filters. I use cpl and nd but never vnd. CPl and VND can reduce the image quality noticeably (not only the colors but the contrast-vnd) I have 3 cpl filters and one is clearly better than the rest. I have one vnd from k&f and...😤 Ps2 For om-log use CAF sensitivity +1.
I noticed on an old camera (EM1II) that for stills, the focusing was better with vivid mode and contrast turned up. I am not quite sure why the camera does not directly use the RAW. It did not seem right that the profile chosen would affect focus, but it did and I used that since I processes the RAW anyway so it made no difference in the final photo. Now it seems you are saying it is the same for video focusing and there the settings to affect the final result. I wonder what and why the processing flow is that this makes a difference. An R6 the focusing is the same no matter what video or still profile I use.
Surprising results -- did you reset both cameras back to factory settings before you started? Maybe sth was off on your OM-1? Cuz like Ben said, the box was there all the time, so it does not seem to be about lack of detection, rather lack of activating AF.
Very interesting Rob. I don't do much video but I can report a similar experience in a quick test on still photos. I thought I must have the settings wrong, but maybe not. I almost never shoot people so it could be me.
Hi Rob, thanks for the video. I haven't had any trouble like you just did but I always have tracking on. It looked like you did not. I also use the +1 +1 settings. Also, I did notice in the beginning it lost you a lot as you moved to the edge of the frame but acted better as you were closer to the center. Interesting results.
Apparently OM1 reps say face/eye detect works fine - except when using tracking. Check out DP review on TH-cam (OM1 vs R5 sports photography shootout).
Hi Rob, I am really interested if you could do a video AF accuracy test, I used to have a em1 ii, and I am generally fine with the speed and response. but what I find really annoying is that a lot of time, it just doesn't react to small movement. for example if your head moved closer a little bit, the focus will not change, and you will get a slightly out of focus footage, sometime you need to zoom in to see, but sometime it is pretty obvious. Also the camera doesn't always focus to the point you set, especially during panning, it tend to jump to the next subject before it move into the focus point. Since I never used a mark iii, I am curious how they compare. I already sold my mark ii, and will definitely get the om-1 in a few months, since I think this is finally a worthy successor to the mark ii. The hardware platform is there, and there is always possibilities they may improve the software at some points.
Did you have zones selected at all? I usually have it set to select all zones though I have not used it on the OM1 yet. Seams like the ND filter made a big difference but face detect box was right on your face the entire time. The tree is beautiful. :)
Interesting. Are you sure this is only video? This sunday I was out shooting at a private party. I had several occasions where the white square was there, but the face was not in focus. Then I let the camera refocus and the second time the face was in focus. It seems like it sometimes pick the backround even with a white square on the face in foreground. I was a bit worried when I saw the results back at the computer. The where many photos where I am certain, the camera showed it picked up the face, but it had not. I am suspecting that I need to have the the camera AF set the whole frame and not just a single point. Maybe your problem is the same. I have never experienced this on my m1 ii.
When I pause the video at a couple of locations where the problem occurs, I can see the focus appears to be on a pair leaves with a lot of reflected light, these leaves aren't in the frame with the E-M1. I'm thinking the algorithm isn't limiting the eye detection to the area detected as a face. It would be useful to see another test for comparison with the camera locations and lenses swapped.
Shutter angle is a video technique where you use double the shutter speed vs. the frame rate of the recorded video. So 120th/sec when recording 60p video.
LOG profile (low contrast) versus Natural profile (normal contrast)? If you want your test to be relevant, you need to set both cameras to the same profile and same aperture for both lenses with identical filters (or no filter).
Thank you Rob for test. That’s interesting. I always liked CAF on Olympus cameras tracking face - eye on mine EM1 m3 never was an issue. I think VND and log format plays big role in this situation. I wonder would results be difficult if you would have changed AF settings, e.g. sensitivity.
As much as it looks like you've uncovered a problem, knowing what happened to that 12-40 gives me pause. In addition to trying with / without eye detect, you may want to try with the lens that was on the MIII (or some other lens) as well. The water spots in the 12-40 in conjunction with the much smaller focus points in the OM1 may not be playing well together.
What was your iso set to on the OM-1? I find lowering to 100 or 64 is a good alternative to ND...I mean if it cuts enough stops for your 180 shutter...edit: i see you had it at 400 iso probably because of OM Log requirements; maybe try the flat profile and use the iso to cut out the extra light, wonder if that would yield better results...although it seems that a firmware update might be the only solution (assuming all of your settings were correct).
the log profile usually requires higher iso. on olympus, its 400 (also, its called omlog400 or so). curiosly, i found that even in the natural profile, which can use lower iso, 400 produces (at least in my observations) a sharper image (i used 400iso in natural in daylight with nd and it seemed to me sharper than a lower iso in that profile, which is not a log profile)
Did you say you were using the 12-40 lense on the OM-1? I'd be interested in seeing the lenses switched, maybe it was the lense since it was left in the rain and snow long ago 🤷. The pictures looked better I think... It went so fast that was hard to tell. I really want the OM-1 but not if this is a real issue. Thanks for testing and posting Rob!
Most cinematographers use ND filters when recording to achieve the proper shutter angle in video. I'm going to do a separate video for ND and Polarizer filters.
OK, now ya did it! Your new camera only recognizes birds! 😁. You did so much bird shooting, you will have to put some feathers on your head for it to work. 😵💫 Your new camera appears to be running on bank robbers detect, where the robbers look like a blob.
Rob, I'm curious why you didn't follow your own No.1 rule and begin by resetting both cameras to their factory defaults 🤔? From this flawed test, you can't really say anything other than it's possible to confuse the OM-1's face detect autofocus but, unfortunately, you've made it very difficult to determine why. Rick
You should try it before sending it back. It may work well for you, especially if you're already invested in m4/3 lenses. Otherwise many have found the Sony pretty reliable. Unfortunately, I don't have any Sony gear to compare.
Everyone, thanks for all the feedback! Using some of the suggestions, I got much better results today. I will post a follow-up video as soon as I am able.
Thank you Rob
I think the problem was the background! The em111 was looking at a much cleaner during the video test! The om1 was dealing with the leaves and trunk of the tree centered behind you! The em1 iii was dealing with a relatively clean background. Also different color grading might have thrown things off as the om1 looked flat and unsaturated! Just a few thoughts from the sidelines!
For me quite the opposite!
His testing situation is very typical shooting situation, not challenging; and something the others (Canon, Sony, Nikon) can easily do. So this is a concern. Enough for Rob is conclude "can't recommend". Hope this is an isolated case specific to this camera. OM-1 should do much better than M1iii. Regardless, as usual, Rob is giving us some really useful real world, no nonsense review.
@@dannygo4230 Yes but that is no serious argument. The easiest solution would have been to reverse the camera position. Not completely correct but would have eliminated the perspective factor provided e.g. the exposure did not change.
Interesting that the OM-1 seems very fast at finding your face and frame it, but can’t focus on it. Btw. I think you are a great teacher and I have learned a lot from watching your channel.
Thanks Rob, wow, that was a bad performance on the OM1. Can’t understand why it’s finds your face but not focusing on it? And also it did not find your eyes one time. Could you report this to OMDS? Hope they fix this soon
I suspect that it's because it's in LOG. LOG profiles in cameras have never been very good with AF until recently with Sony and Canon's advancements in tech. OM really has to modify the AF algorithm to properly read LOG or remake the LOG profile for 10-bit. I think this is also why the E-M1 III performed very well because it was in a more saturated picture profile.
So.. don't try to use AF with LOG? Is HLG profile considered LOG? All I want to do is film 4K60 10bit HLG but this AF performance looks unusable.
@@charbax I can't say for sure until I get mine. But if I were to assume, HLG would perform much better and would most likely have better highlight retention than LOG. HLG is a Hybrid Log profile which work much better with 8bit and 10 bit/4-2-0. On my GH5s the AF is much better in HLG than LOG. It's nearly as accurate and responsive as using Cine D/Natural.
Hi Rob, This is the video I have been waiting for. I have an issue with the OM1 C-AF, face detect and the either (evil😀) eye if available, shooting relatively stationary interview subjects. Focus will lock, lose and hunt, rendering the footage unusable. I have been testing a number of different settings and the last attempt was the most successful: turning sensitivity -1 and speed to +1. Of course this is scenario specific and I am not yet convinced I have had a breakthrough.
I have tried S-AF but I really do prefer a workable tracking solution. Really hope you can continue to research this (I have also raised this with Emilie).
Very interested to see how this finally works out.
For me, AIF is The most important feature I look for in a new Olympus/OMS body.
Content with the, now outdated AF, in my 2 Pen Fs. The type of shots I take don't require very speedy AF or AF tracking.
If these are needed, particularly for people, or young children shots, I have no hesitation in borrowing a friend's Sony a6400 and using the magic Sigma 56 1.5- simply stellar and, if purchased, mint second hand, within my price range.
Sony in FF and apsc has had AIF nailed for what must be the better part of ? 5 years'.
Canon and Nikon have now caught up.
OMS have had forever to build totally competent IAF- to at least match the far cheaper Sony a6400 and other bodies and it would be a major disappointment if the new OM1 did not, at least, match what some may regard as a more lowly a6400, let alone a 6100 etc.
Complex features, such as focus stacking and the like are of little interest to me (others will strongly disagree), however, sharp quick, locking focus is fundamentally important to me.
Am waiting for the flack to arrive- we all differ.
I have found nothing to match the AF of Sony. I rented a Canon R6 for a basketball job which the EM1III and EM1X could not handle. It was good enough but not as good as the Sonys are for that. I understand that OM Systems uses the AI subject detect to do what DSLRs used to do so easily. They still have a way to go with tracking although continuous autofocus is getting better in most conditions. Just wish they would add people as a subject because sometimes in sports, you want to get the person not the ball.
Hi Rob, when you removed the ND filter you increased the shutter to 1/750 of a sec? I thought you would shoot at 1/120th being twice 60fps? What aperture and shutter speed did you use on the M1 mark iii?
I don’t shoot video so I can’t wait for the photo examples to see if there is any improvement over my em1 mk3. If not I can stop worrying about buying om1 and stick with what I have. Thanks Rob. Loved your video.
I have also had problems with the eye detect autofocus in the picture mode on the OM-1 too. The green focus point was all over the face rather than the eye when reviewing the images in camera. It is definitely not as good as the M!X
Super useful information. unfortunately, it confirms my fears. Hope they fix it asap!
Were both cameras using the same color profile and adjustments (contrast mainly) as for some reason this affected focus on the EM1s. Although with video you have less chances to improve focusing this way.
Interesting video but the settings seemed to be very different on the two cameras, please let the settings also be visible on the EM1 Miii. The camera position is very different, the OM1 has got much more light of the sky than the EM1.
What a wonderful chair in the background.
This video made me do a video test with the E-M1X wearing a hat and glasses. Thankfully the focus locked on correctly. This behavior is surprising from the OM-1 considering it does recognize your face.
wow that is very interesting. Surly face detect would be easier solution compared to all the other detection modes.. Maybe a firmware update needed... Anyway it is what it is.. the main thing is we got T.J Hooker (William Shatner) on the case 13:25 🙂
You are the second professional that uses OMD-SYSTEM that does the continuous focus test that has the same focus problem, the other professional did it by taking pictures of a train and the focus was also lost, it was done by a Spaniard.
I think it's really close. Most of the time it's perfectly fine. Hopefully they can improve the af with some firmware updates.
Sad AF. The only reason I was considering OM1 was for its PDAF which I thought would be much superior to the GH6 CAF. But if this is what one can expect, then it's unusable. And I want to film all videos handheld with the IBIS in 4K60 HLG 10bit HEVC to dual SD, I really need the AF to be absolutely 100% reliable. Like a Sony but in a micro43. Anyone know of any other OM1 reviews showing handheld IBIS AF performance in 4K60?
This is why you should be an official representative with OM. I’m sure they watch your blog (they best should) but we’d all be better served if you were hooked in directly to the company so you could give direct feedback and get these use case problems resolved.
Interested! But Rob i just wonder if that Oly 12~40mm f2.8 lens is not the problem.I recall you left it outside in snie etc and it could be suspicious Just my 2cent.
Thank you for this interesting but very unsatisfying result. Is it only in video-mode so bad and better in fotography-mode?
The strangest thing to me, ist that the rectangle shows the detection of your face very well, but it remains unsharp. It also newer showed the recognition of an eye.
Interesting video. I set up my OM-1 this morning and tried to replicate your results. I was not able to do so. C-AF Video with straight face detection worked fine. C-AF Video with face and eye detection also worked fine. I tried with and without a ball cap much like what you had and it had no issues at all focusing on my face. I moved all around, forward / back / side to side / turned around, hid my face with the cap, etc. Each time AF locked on my face without issue. No idea what is the difference between your tests and mine. I had not even tried it until this AM. I also tried C-AF+TR with subject detection (I had it in the animal mode) in video and it had no trouble following my dog around the yard.
Closer to the lens, the face detection was no problem. Further away from the lens, the face detection did not perform at all. So, maybe he has set focus distance limit on the lens ?
thanx, Rob, for this comeback on the autofocus in video. i think in general log/flat profiles are worse in af, maybe because the hybrid system also uses the contrast-detect in addition to phase, and so contrast in log/flat profiles is harder to detect. but overall, yeah, at least according to this test, it does seem that om1's af in video is not reliable. i thought they really stepped up the game, but it seems although it's reliable in photos, it's not for videos.. sony works well even in flat profiles (and also in 24fps, which is harder on the system than 60fps, as far as i know, because in 60 the system has more data to work with, or something.. but if om1 does so bad in 60, i imagine its even worse in the cinematic 24)
When I borrowed an em1x I found the train tracking a mixed bag. Sometimes ok, other times it didn’t lock the face of locomotive. I’ve heard it was developed for bullet trains in Japan so may not be optimal for North American locomotives.
Hi Rob. Interesting.
Maybe it's a lens. I have a completely different experience. Om1 is in my opinion more effective than m1miii. I did a similar test with a 17mm 1.2 wide open and the om-1 was definitely better in low light. In good light, i didn't see much of a difference. The OM1 recognizes the face a little faster, but also not every time. With 12-40 2.8 i didn't have any problems either. There is one more thing that affects autofocus, that is picture mode. With H265 10 bit om-log, h264 flat and om-log af is slower because because the contrast is reduced. It's the same with the M1miii (the autofocus with the natural profile works better than with flat and log). The difference, however, is not as big as in your test.
I had some problems with tracking+caf with 12-40 2.8 and om-1 and m1. I had to turn off the camera and remove the lens. Sometimes 12-40 works louder on one camera than on another, and sometimes it loses autofocus. Also 8mm 1.8 works louder on m1mii than on m1miii and om1. 🤔
Ps.I hate variable filters. I use cpl and nd but never vnd. CPl and VND can reduce the image quality noticeably (not only the colors but the contrast-vnd) I have 3 cpl filters and one is clearly better than the rest. I have one vnd from k&f and...😤
Ps2 For om-log use CAF sensitivity +1.
Totally agree about variable ND filters.
I noticed on an old camera (EM1II) that for stills, the focusing was better with vivid mode and contrast turned up. I am not quite sure why the camera does not directly use the RAW. It did not seem right that the profile chosen would affect focus, but it did and I used that since I processes the RAW anyway so it made no difference in the final photo. Now it seems you are saying it is the same for video focusing and there the settings to affect the final result. I wonder what and why the processing flow is that this makes a difference. An R6 the focusing is the same no matter what video or still profile I use.
Surprising results -- did you reset both cameras back to factory settings before you started? Maybe sth was off on your OM-1? Cuz like Ben said, the box was there all the time, so it does not seem to be about lack of detection, rather lack of activating AF.
Very interesting Rob. I don't do much video but I can report a similar experience in a quick test on still photos. I thought I must have the settings wrong, but maybe not. I almost never shoot people so it could be me.
Hi Rob, thanks for the video. I haven't had any trouble like you just did but I always have tracking on. It looked like you did not. I also use the +1 +1 settings. Also, I did notice in the beginning it lost you a lot as you moved to the edge of the frame but acted better as you were closer to the center. Interesting results.
Yes I would love to be entered into give away. Harsh keep up telling it like it is in all cases. That's the only way we are ever going to get answers.
Apparently OM1 reps say face/eye detect works fine - except when using tracking. Check out DP review on TH-cam (OM1 vs R5 sports photography shootout).
Why doesn't the M1 III show the focus box?
I used a separate hdmi recorder on the OM-1. I don't have an extra one so I just showed the EM1.3 file.
any filter, ND or not, do affect the Af. E.g. the 300mm end of 100 300 lens, is badly influenced when the filter is on.
Could you use LiveND instead?
Hi Rob, I am really interested if you could do a video AF accuracy test, I used to have a em1 ii, and I am generally fine with the speed and response. but what I find really annoying is that a lot of time, it just doesn't react to small movement. for example if your head moved closer a little bit, the focus will not change, and you will get a slightly out of focus footage, sometime you need to zoom in to see, but sometime it is pretty obvious. Also the camera doesn't always focus to the point you set, especially during panning, it tend to jump to the next subject before it move into the focus point.
Since I never used a mark iii, I am curious how they compare.
I already sold my mark ii, and will definitely get the om-1 in a few months, since I think this is finally a worthy successor to the mark ii. The hardware platform is there, and there is always possibilities they may improve the software at some points.
OMG. I am looking to order one when stock becomes available and this is not good
Did you have zones selected at all? I usually have it set to select all zones though I have not used it on the OM1 yet. Seams like the ND filter made a big difference but face detect box was right on your face the entire time.
The tree is beautiful. :)
Interesting. Are you sure this is only video? This sunday I was out shooting at a private party. I had several occasions where the white square was there, but the face was not in focus. Then I let the camera refocus and the second time the face was in focus. It seems like it sometimes pick the backround even with a white square on the face in foreground. I was a bit worried when I saw the results back at the computer. The where many photos where I am certain, the camera showed it picked up the face, but it had not. I am suspecting that I need to have the the camera AF set the whole frame and not just a single point. Maybe your problem is the same. I have never experienced this on my m1 ii.
When I pause the video at a couple of locations where the problem occurs, I can see the focus appears to be on a pair leaves with a lot of reflected light, these leaves aren't in the frame with the E-M1. I'm thinking the algorithm isn't limiting the eye detection to the area detected as a face. It would be useful to see another test for comparison with the camera locations and lenses swapped.
What is shutter Angle? Do you mean shutter speed?
Shutter angle is a video technique where you use double the shutter speed vs. the frame rate of the recorded video. So 120th/sec when recording 60p video.
LOG profile (low contrast) versus Natural profile (normal contrast)? If you want your test to be relevant, you need to set both cameras to the same profile and same aperture for both lenses with identical filters (or no filter).
Still it does not matter, the OM-1 did not find his face 90% of the time, I say if it is within 30 days return the damn thing could be a lemon.
Thank you Rob for test. That’s interesting. I always liked CAF on Olympus cameras tracking face - eye on mine EM1 m3 never was an issue. I think VND and log format plays big role in this situation. I wonder would results be difficult if you would have changed AF settings, e.g. sensitivity.
It is weird that the camera can detect your face (the rectangular box) but unable to control the lens focus on it correctly.
I think that instead of C-AF with face detect you should have had C-AF + Tracking with face detect when using video. Worth trying anyway.
As much as it looks like you've uncovered a problem, knowing what happened to that 12-40 gives me pause. In addition to trying with / without eye detect, you may want to try with the lens that was on the MIII (or some other lens) as well. The water spots in the 12-40 in conjunction with the much smaller focus points in the OM1 may not be playing well together.
What was your iso set to on the OM-1? I find lowering to 100 or 64 is a good alternative to ND...I mean if it cuts enough stops for your 180 shutter...edit: i see you had it at 400 iso probably because of OM Log requirements; maybe try the flat profile and use the iso to cut out the extra light, wonder if that would yield better results...although it seems that a firmware update might be the only solution (assuming all of your settings were correct).
the log profile usually requires higher iso. on olympus, its 400 (also, its called omlog400 or so). curiosly, i found that even in the natural profile, which can use lower iso, 400 produces (at least in my observations) a sharper image (i used 400iso in natural in daylight with nd and it seemed to me sharper than a lower iso in that profile, which is not a log profile)
Did you say you were using the 12-40 lense on the OM-1? I'd be interested in seeing the lenses switched, maybe it was the lense since it was left in the rain and snow long ago 🤷. The pictures looked better I think... It went so fast that was hard to tell. I really want the OM-1 but not if this is a real issue. Thanks for testing and posting Rob!
I’ll keep my E-m1 Mk2. Thanks.
not as fast as SONY, but quick enough. Cheeper EM1mk3 might be good buy.
Thanks for the feedback.
Why would you test AF with an ND filter on. A polarizer would be a better test with something on the lens.
Most cinematographers use ND filters when recording to achieve the proper shutter angle in video. I'm going to do a separate video for ND and Polarizer filters.
Om-1 software requires a repair package. Definitely.
Interesting.
OK, now ya did it! Your new camera only recognizes birds! 😁. You did so much bird shooting, you will have to put some feathers on your head for it to work. 😵💫 Your new camera appears to be running on bank robbers detect, where the robbers look like a blob.
Rob… wth is with the OM-1 or maybe just your camera? That’s bad
The only solution maybe you should put a Bird sticker logo in your face 😎
Very strange. It's clearly recognizing you because the box is really sticky. For some reason the focus is not moving though. Pretty bizarre.
Olympus are photo masters, not video 😉 you need Panasonic for that 😊 but they aren't that good on photography though 😬
Rob, I'm curious why you didn't follow your own No.1 rule and begin by resetting both cameras to their factory defaults 🤔? From this flawed test, you can't really say anything other than it's possible to confuse the OM-1's face detect autofocus but, unfortunately, you've made it very difficult to determine why.
Rick
it sees you because it puts the box on you …. so why isn’t it following through and focusing!!!?
Hard to say.
What a disappointment. I want my money back. Just paid for OM-1 and unfortunately it has been shipped. Will be returning unopened.
You should try it before sending it back. It may work well for you, especially if you're already invested in m4/3 lenses. Otherwise many have found the Sony pretty reliable. Unfortunately, I don't have any Sony gear to compare.
It does recognize your face, but still can't keep you in focus.
Skin tone sucks with OM-1. Glad I don’t have your issues.
Glad I don't shoot video..lol
Looks almost as bad as Nikon cameras with the Eye AF, terrible
I haven't tried the Nikon.
Snif…