In my case a Red-Tailed Hawk filling about a quarter of the frame in sun with catchlights on its big yellow eyes completely failed to activate BEAF at all, so that's an epic fail on Sony's part!
What's interesting is that there are clearly two types of experience being reported with v1.31 firmware -- those who see no difference or even an improvement, and those who see a noticeable (often dramatic) degradation of eye AF capabilities. I'm in the latter camp, but the stark contrast in peoples' experiences makes me wonder if there's a bug in the firmware that's triggered by some other in-camera setting.
wordfool wrote:
In my case a Red-Tailed Hawk filling about a quarter of the frame in sun with catchlights on its big yellow eyes completely failed to activate BEAF at all, so that's an epic fail on Sony's part!
What's interesting is that there are clearly two types of experience being reported with v1.31 firmware -- those who see no difference or even an improvement, and those who see a noticeable (often dramatic) degradation of eye AF capabilities. I'm in the latter camp, but the stark contrast in peoples' experiences makes me wonder if there's a bug in the firmware that's triggered by some other in-camera setting. ...Show more →
I have been thinking that could be the issue as well. There are several focus-related parameters, it is difficult to re-create/test every combination.
This has been my understanding from the beginning. When I posted upthread that I didn't appear to have any problems, I was taking this into account. BEAF doesn't detect eyes very often and in my view never did. You have to be quite close, in acceptable conditions and with the right kind of bird. Ditto for animals. If BEAF doesn't detect, I just accept that one of those conditions hasn't been met and adapt accordingly.
So when I said that I hadn't experienced any problems with the A1, 200-600 and firmware 1.31, I was refering to AF accuracy issues and not that eye detect wasn't working when I though it should have.
So people seem to have two issues here:
1. BEAF didn't detect an eye when the photographer though it should.
2. The image was not in perfect focus.
a. in situations where an eye was detected
b. in situations where an eye wasn't detected.
Perhaps we should be clear about which issues we think have got worse with 1.31?
wordfool wrote:
In my case a Red-Tailed Hawk filling about a quarter of the frame in sun with catchlights on its big yellow eyes completely failed to activate BEAF at all, so that's an epic fail on Sony's part!
What's interesting is that there are clearly two types of experience being reported with v1.31 firmware -- those who see no difference or even an improvement, and those who see a noticeable (often dramatic) degradation of eye AF capabilities. I'm in the latter camp, but the stark contrast in peoples' experiences makes me wonder if there's a bug in the firmware that's triggered by some other in-camera setting. ...Show more →
I do recall that, roughly after the firmware update (but not sure), BEAF would not engage at all. I thought I might have accidentally pressed the button I use to toggle through the eye types, but noticed while cycling that it had already been set to birds as birds was the last to show up. But. Afterwards it worked. Huh. So I thought maybe it forgets the setting during power cycle? No it doesn't. For me that was a glitch occurring exactly once. Or I might have been seeing things as, of course, I can't reproduce the effect.
I don't shoot birds but I've had spotty performance with the A1 and 200-600mm. I've tried a few different copies of the lens with the same results. I need to try my A9II's with the 200-600mm to see if the results are the same.
It's insane to me that there are so many AF woes being reported against Sony's flagship cameras with no resolution in sight. I feel for you guys not being able to at least downgrade to a FW that seemed to work better for you.
RoamingScott wrote:
It's insane to me that there are so many AF woes being reported against Sony's flagship cameras with no resolution in sight. I feel for you guys not being able to at least downgrade to a FW that seemed to work better for you.
It's definitely frustrating not being able to use the flagship $6k camera with what should be a great lens. I rarely need the full resolution when I'm shooting the 200-600 so being able to punch in with the A1 really opens up possibilities for the combo. Unfortunately my hit rate is so low I can't use it on assignments.
That response is missing the step of resetting the camera before performing the update (as well as after). Unfortunately one cannot re-install Sony firmware so if the previous installation went awry in some way you’re SOL.
I reinitialized after installing firmware (like the person at DPR) and it made no difference to my BEAF issue.
I could not discern any difference in BEAF performance when I upgraded my current A1 (from 1.0 to 1.31).
In this part of the world, I can imagine that a lot of the perceived "issues" might just be due to the generally poorer light conditions we typically have this time of the year... since low-light AF is one area where Sony still lags all of their competitors.
wordfool wrote:
In my case a Red-Tailed Hawk filling about a quarter of the frame in sun with catchlights on its big yellow eyes completely failed to activate BEAF at all, so that's an epic fail on Sony's part!
a target that's only 1/4 of the frame is going to have a small face and small eyes, so it's a too-short focal length problem, there isn't enough picture data for the a.i. to work on, and if you are shooting stopped-down it's harder on the af.
even with the worst-case long neck beaf situation, if you put the sony real-time tracking focus box on the head this latest firmware will detect the face and trigger beaf, see the a1/200-600/1.4x example below.
the orange box is tracking/eyeaf location, every frame in the burst was on or very close to the eye.
osv2 wrote:
a target that's only 1/4 of the frame is going to have a small face and small eyes, so it's a too-short focal length problem, there isn't enough picture data for the a.i. to work on.
even with the worst-case long neck beaf situation, if you put the sony real-time tracking focus box on the head this latest firmware will detect the face and trigger beaf, see the a1/200-600/1.4x example below.
the orange box is tracking/eyeaf location, every frame in the burst was on or very close to the eye.
You ignore the fact that I've been using the A1 for almost two years taking photos of birds, including plenty of hawks in the same area at the same times of year in the same light, so I know how it used to perform and what to expect and I can say with certainty that it is no longer performing at the same level when it comes to BEAF. I also might have exaggerated the "filling quarter of the frame" factor... it was relatively close but maybe not quarter of the frame As far as long-necked water birds go, I've found the A1 always had problems finding the eye so that's nothing new.
I also don't generally use tracking focus when shooting birds but I wonder now if that might be a factor in the relative failure of BEAF since v1.31. I generally just use AF-C non-tracking (small spot or zone focus) as do a lot of photographers AFAIK.
wordfool wrote:
You ignore the fact that I've been using the A1 for almost two years taking photos of birds, including plenty of hawks in the same area at the same times of year in the same light, so I know how it used to perform and what to expect and I can say with certainty that it is no longer performing at the same level when it comes to BEAF.
i noticed that something had changed in the af as well, but i fixed it for the most part by no longer using the [1(Locked On)] setting, i put it on "2", i did not reload the previous settings off of an sd card... i also did a factory reset first, fwiw.
as someone posted earlier, this is a difficult thread because nobody is talking about what settings they are using.
wordfool wrote:
As far as long-necked water birds go, I've found the A1 always had problems finding the eye so that's nothing new.
i just demonstrated that it finds the eye without any problem, just not in the wide-area mode that most bif people out here prefer.
wordfool wrote:
I also don't generally use tracking focus when shooting birds but I wonder now if that might be a factor in the relative failure of BEAF since v1.31. I generally just use AF-C non-tracking (small spot or zone focus) as do a lot of photographers AFAIK.
i don't see that methodology with other bif shooters, they tend to use wide-area mode, depending on the background.
i don't know how what you are calling "non-tracking" is different than tracking, but for moving targets, tracking with af-c is what sony recommends, i can't imagine doing it any other way:
"Even if you usually shoot images using the tracking or Face/Eye AF function, you can pause the tracking and face/eye detection functions in the following cases, where you want to switch a subject to another one during shooting or the camera cannot recognize a face because the subject is wearing a helmet.
When the tracking function is paused, the focusing area will be switched to the focusing area where tracking was started. When tracking was started by selecting [Spot], you can specify a subject you want to shoot at the spot you set in the frame indicating the focusing area.
After an intended subject is in focus, you can restart tracking by canceling [Tracking-Off Hold]." https://support.d-imaging.sony.co.jp/support/ilc/custom/04/en/practice1.html
arbitrage wrote:
Seems the ship has sailed on this one....
I'm holding off getting that 600mm f4 GM so will stick with the A1 200-600 for now . I'll keep building the pot up so it gives me more options what to get . Z9's are a little easier to get in the UK , just the lenses that's a PIA . I'm just using the A1 same as I used to use the A9ii , all eye priority off .
osv2 wrote:
i noticed that something had changed in the af as well, but i fixed it for the most part by no longer using the [1(Locked On)] setting, i put it on "2", i did not reload the previous settings off of an sd card... i also did a factory reset first, fwiw.
as someone posted earlier, this is a difficult thread because nobody is talking about what settings they are using.
i just demonstrated that it finds the eye without any problem, just not in the wide-area mode that most bif people out here prefer.
i don't see that methodology with other bif shooters, they tend to use wide-area mode, depending on the background.
i don't know how what you are calling "non-tracking" is different than tracking, but for moving targets, tracking with af-c is what sony recommends, i can't imagine doing it any other way:
"Even if you usually shoot images using the tracking or Face/Eye AF function, you can pause the tracking and face/eye detection functions in the following cases, where you want to switch a subject to another one during shooting or the camera cannot recognize a face because the subject is wearing a helmet.
When the tracking function is paused, the focusing area will be switched to the focusing area where tracking was started. When tracking was started by selecting [Spot], you can specify a subject you want to shoot at the spot you set in the frame indicating the focusing area.
After an intended subject is in focus, you can restart tracking by canceling [Tracking-Off Hold]." https://support.d-imaging.sony.co.jp/support/ilc/custom/04/en/practice1.html ...Show more →
When I first started birding I used tracking, but quickly discovered that for any fast moving bird in trees or against other busy backgrounds the tracking proved to be a detriment and so with experimentation and advice of some folks here I switched to using basic, AF-C non-tracking small spot and zone focus (and occasionally wide for BiF). Me tracking the bird with the focus area (and assistance from BEAF) is far more reliable than allowing the camera move the focus area around, regardless of how sticky I make the tracking sensitivity. I don't think this is an unusual way of doing things from what I've read and discussed with other photogs in the field over the years, but I wonder if Sony has now optimized BEAF for tracking focus.
My problem with water birds is probably user error as much as anything else because I don't often photograph them. But if an Egret, which already has a small pale eye in a sea of white, bends its neck I find BEAF goes a bit haywire, often finding an "eye" in the dark space between the bend of the neck and something in the background or, in one extreme example, it just kept wanting to focus on the dark space created when the bird bent its neck enough to give the illusion (at an angle) of two parts of the neck meeting (despite the small spot focus area being right over the bird's head).
As far as long-necked water birds go, I've found the A1 always had problems finding the eye so that's nothing new. 
