k-h.a.w wrote:
Thanks Gary. Many thanks for the great shots. My understanding of Eye-AF is that during shooting a white/green wireframe rectangle/square should be visible around an eye. Is this what you saw? TIA.
K-H.
Actually, I haven't seen that with birds, now that you mention it. I have seen it occasionally with humans, though.
The single focus point rectangle on the E-M1 i is quite a bit larger than was the one in the mark 1 and so it often covers quite a bit of the bird at longer distances, like the second image. Maybe the eye detect isn't the.only reason I get better eye shots with the mark 2, but it seems to help, so I leave it on.
gary0319 wrote:
Actually, I haven't seen that with birds, now that you mention it. I have seen it occasionally with humans, though.
The single focus point rectangle on the E-M1 i is quite a bit larger than was the one in the mark 1 and so it often covers quite a bit of the bird at longer distances, like the second image. Maybe the eye detect isn't the.only reason I get better eye shots with the mark 2, but it seems to help, so I leave it on.
Thanks Gary for your candid feedback. As long as it doesn't interfere with or slow down the E-M1.2, I agree, one might as well leave Eye-AF on with the off-chance that I might be actually helpful. I have done that as well and always pay attention to whether a face or eye framing wire mesh rectangle appears. I have noticed those white/green framing rectangles only with human faces so far. I don't think there is a law of nature that would prohibit Olympus from offering face/eye detection in the future for non-human faces/eyes - maybe even through a firmware update. We can always hope. But I don't think the E-M1.2 is quite there yet.
I also agree with you that the focus system of the E-M1.2 is not only faster, but also more accurate in picking out what I had intended to focus on. An eye, like the one in your first example image, certainly offers a high micro-contrast area that I would think the focus system shouldn't ignore provided it is in its focus rectangle.
Of course, one can always point the focus rectangle on an eye without invoking Eye-AF.
Took a few shots of my cat tonight and I can definitely state that face and eye focus works on my cat! Not every time or very consistently though. The big white box will frame his head and the small green box with fix on his eye. And sometimes the single AF box would take over. And then the white + inner green box would take over the face-plus-eye focus tracking again.
You have to try it yourself - I promise you that it works, albeit not every time - but approx 50% of the time.
Thanks Bobby for the info, much appreciated.
This is the first definitive statement I have seen.
Which lens, aperture, and camera settings did you use?
In particular, which focus area, which AF, and which shutter? TIA.
I didn't find it all that useful to rely on it, because as you stated the face/eye has to full the frame pretty substantially in order for it to pick up. It's better than the Fuji Eye detect in my opinion but still behind Sony.
TMaG82 wrote:
I didn't find it all that useful to rely on it, because as you stated the face/eye has to full the frame pretty substantially in order for it to pick up. It's better than the Fuji Eye detect in my opinion but still behind Sony.
I have to agree with you. Sony has the best face/eye detection and tracking. Having said that AF performance in low light is very poor compared to the E-M1.2.
k-h.a.w wrote:
Thanks Bobby for the info, much appreciated.
This is the first definitive statement I have seen.
Which lens, aperture, and camera settings did you use?
In particular, which focus area, which AF, and which shutter? TIA.
K-H.
Single AF with face/eye detection. E-M1.2 with the Nocticron.
As a pet/rescue shooter and looking to get the Nocticron, Eye detection working on pets would be a HUGE benefit for me.
From what I have been reading from poeple in different forums etc, Face/Eye detection effectiveness tends to go Sony<Panasonic<Olympus<Fuji at least of the major mirrorless systems.
Wilvoeka wrote:
As a pet/rescue shooter and looking to get the Nocticron, Eye detection working on pets would be a HUGE benefit for me.
From what I have been reading from poeple in different forums etc, Face/Eye detection effectiveness tends to go Sony<Panasonic<Olympus<Fuji at least of the major mirrorless systems.
Face/eye detect works on the E-M1.2 but it's not reliable or dependable i.e. it comes on some of the time but not every time. Having said that, I think a fast lens like the Nocticron works better than slower lenses, but it's only a hunch.
I have done quite a bit of pet rescue work as a photographer. Here's a couple of images I shot with my old Canon 7D.
Well one thing is for sure, it doesn't work on bees (although CAF + tracking seemed to do a good job keeping a bees head in shot on a wind blown shrub today)
I was shooting some macro tests and found it almost impossible to get my mk2 + 60mm macro to focus on the head of my bugs and bees which I finally tracked down to having accidently switched eye detect on.
Switched it off and my problems were solved so I could get on with trying SAF vs CAF vs CAF+ tracking. In the windy conditions prevailing CAF or CAF + tracking worked best
Wilvoeka wrote:
As a pet/rescue shooter and looking to get the Nocticron, Eye detection working on pets would be a HUGE benefit for me.
From what I have been reading from poeple in different forums etc, Face/Eye detection effectiveness tends to go Sony<Panasonic<Olympus<Fuji at least of the major mirrorless systems.
I would tend to agree with you there on the order. Fujis is basically there for fluff.
So Bobby many thanks, I displayed your image https://www.fredmiranda.com/forum/ufiles/56/1438956.jpg
and Imagemaster's image https://www.fredmiranda.com/forum/ufiles/18/1438818.jpg
on my 4K monitor, took my E-M1.2 + Nocticron 42.5/1.2, selected Face & Eye Priority On, and chose all focus points.
The camera put indeed an unstable (oscillating or briefly appearing) white/green wireframe mesh around the cat's and hummingbird's face.
It also briefly put a green wireframe mesh square/rectangle around an eye and and took some pictures.
In the EVF the focus is indeed on the eye although I used all focus points.
I find the same if I choose the central focus point but don't point directly at an eye.
So Eye-AF indeed sporadically seems to work but the algorithm seems unstable.
Olympus should be able to correct the unstable Eye-AF focus algorithm.
This is very encouraging indeed.
K-H.
PS: To the admins: Please delete an image if the owner objects. TIA.
