R5 with RF100-500 - focus performance for BIF

[Jonathan Huyer]

As you probably saw with my Image of the Week submission, I played with my new RF100-500 on the R5 yesterday and had quite a bit of fun. I really wanted to test the auto focus capability with birds and especially birds in flight. My subject was Great Grey Owls, and we had plenty of action. At this point I would rate the focus performance as "pretty good", and I'm especially hoping that I can get some feedback as to what might have been going wrong in some cases. Have a look at the sequence below and please share your thoughts.
My focus settings were:
- Servo, animals, eye-detect enabled
- AF mode 1, with tracking sensitivity set to -2 and accel/decel tracking set to +2 (I heard this was supposed to make things better)
- All focus points activated, with the initial point chosen automatically
The focal length was at 500 mm, aperture f/7.1 (wide open). Images below are unprocessed and not cropped. I used electronic shutter at 20 fps.
The conditions were obviously challenging for focus. A long-distance shot, a busy background, and plenty of snow falling. The camera had the owl in focus before it took off, but could then only hold focus for a couple more shots before it lost it completely. The images after this sequence are progressively worse even though I held the owl close to the middle of the frame.


#1: Sharp



#2: Still good.



#3: Still good, but really hasn't moved much yet.



#4: A bit of softness



#5: Completely soft





#6: Lost it completely, focus shifting to the background



So what do you think? Is it the distance, the tough conditions, wrong settings? Or some combo of all that? I'd appreciate your feedback.

You can click on any shot to see a larger version.
Thanks!

[Busted Knuckles]

grey owl on a grey back ground low contrast small eye target.

in clear air I have little difficulty once it locks on. I use the default settings. It does miss once in awhile in acquisition. Once acquired it seems pretty darn good.

my 2 cents.

[Kayaker72]

I'll be curious if others have any suggestions, but that is pretty darn close to my set up (I think I am -1 and +2).

Looking at the images, it looks like it grabbed onto the snow directly behind the initial subject. Not sure why. The -2 on sensitivity should be making it stay on subject, and it could have grabbed the tree. But it does seem like it is the snow behind the tree that it grabbed onto.

So, I would chalk it up to a very difficult subject. The falling snow against the background is a bit too much like the owl, IMO.

[Joel Eade]

I would agree.....it's probably more the effect of tough conditions with little contrast between the subject and things around it.

I have not had a ton of birds in flight experience with the R5 + RF 100-500mm but in clear conditions I have seen it track birds quite well even if they are very small in the frame.

I don't think your settings are to blame.

[Fast Glass]

I agree that it is tough situation for AF, most notably because of the small AF target. This is something I am finding a weakness on my 5Ds R as well and I just learn to be very precise where I place the AF point.

Now even on my 1D III and 1Ds III has a significantly better time.

1Dx III, childs play. It handles situations like this like a boss.

Now for the obvious, a longer lens will fill up the frame more and make a tremendous difference and even though some of the older big whites are dated in some ways. It will be a tremendous upgrade from the 100-500mm, and depending on what you get maybe not even that much more money. A 500mm f/4 IS can be had for pretty reasonable prices or a 600mm f/4 IS for roughly $4500. And of course the veritable 800mm f/5.6 IS, which used can be had for $8000/$10000 depending on condition, age ect.

Pretty much the AF performance what you are seeing is what the R5 is capable of. You might get better hit rate with a fixed AF location, because it doens't hop around, but even then the subject is very small in the frame and as you or the subject moves is liable to lock onto the background. But it probably be better, so I would try that. As I did shoot it a fair amount R5. And it's nowhere near what it is cracked up to be. It's a 5-series, if you are coming from that perspective, then it is not so surprising. Coming from a 1-series, it's a night and day difference in more challenging situations.

[Jonathan Huyer]

Thanks for the feedback. Yah it's probably just a real tough focus situation. I'm also wondering if the AI algorithm is as well tuned towards owls as it is towards other birds. It seems to lock quickly onto birds that have a proper beak, so I'm guessing that the AI is programmed that way. Owls are kind of lumpy and funny looking, which might fall outside the pre-programmed focus algorithms.

Meanwhile I do have to say that the R5 with 100-500 is sure nice for hand holding! So incredibly light. When you can get by with f/7.1, it's a game changer.

[HDNitehawk]

Bird takeoffs are tough. They explode so fast it is hard for the AF to kick in. If those shots were sequential in five frames a whole 1/3 of a second elapsed +/-. That's pretty demanding. Also here is way to much going on otherwise in that setting that would undermine you. I haven't gotten a chance to test the R5 with a take off or landing like this one, but those type of shots are one of my favorites and I have been looking for a chance to test it.

Typically though with the DSLR's the bird might fall out of focus for an instant but once the camera caught up it would track. You saying it progressively got worse is concerning.

One thing I have noticed with the R5 when using just a single point for focus it has a harder time picking up a subject in the bush than some of the DSLR bodies. It seems like the single point is fat. Maybe there is a setting I am overlooking, I will just keep playing with it until I figure out if it is my settings or the camera.

[Jonathan Huyer]

It's true, I'm asking quite a lot of the technology and I suppose to be able to identify the bird on the treetop at all is pretty good. I was hoping that the focus would catch up to the owl when it was flying, but it just wouldn't reacquire it at all. Ironically this might be due to the settings I chose, which enhanced the lock-on tracking. Once the focus decided that I really wanted to take a picture of the background trees, it ignored the bird completely. So it's a setting that cuts both ways, I suppose. Maybe if I had set the focus mode to quickly jump to the nearest subject, it would have found it again.

With regards to the single-point focus, yes the standard point is rather large. The spot-AF point is about half the size, which should work better in the bush. However you probably give up some accuracy when you make the point smaller. One thing that I do like about the 1DXIII versus the R5 is that the focus points are very small. I don't even bother activating the spot-AF mode on that camera. The R5 single-point focus boxes are quite a bit larger. But then, you get coverage over nearly the full screen which is quite an advantage.

[HDNitehawk]

The spot AF seems like it will be a negative for me. It sounds like you are doing the same thing I am doing, trying to find the balance in the camera for the right settings. Last weekend was the first real outing with the RF 600MM F/4 and the R5 and I came away with a list of things I need to sort out for the AF.

One other thing that is puzzling is that the settings in the body seem to randomly change. For instance I was in AV mode I have it set in Auto ISO and suddenly it switches to a set IS0. Another time I am shooting and the picture seems washed out and I find that the camera is set to Kelvin. I still believe it is an ergonomic issue and I just haven't gotten used to the body.

[Jonathan Huyer]

Yes that switch from Auto ISO to set ISO is a problem I've had too! You have to go back to page 2 of menu 1 and choose ISO speed settings, then select Auto once again. I don't know why that happens. I haven't had the white balance change like you describe though.