Jared Polin new video tests eye tracking in production R5/R6

Bernie_King said:

RDM5546 said:

Bernie_King said:

I don't know. I'm trying to find that out as well. There's apparently some kind of cut off.

Click to expand...

I was thinking there may be some 1.2 or 1.3X cropping mode you are forced into when using than lens. May an alternative there is they cut off the edges sensor because the work substand with the craziness of that DO 800mm lens. There clearly is a story of deliberate intervention of some type. It will come out with more reviewing going on.

Click to expand...

I found the answer in Gordon Laing's review of the 800 and 600 f/11 lenses.

https://www.cameralabs.com/canon-rf-600mm-800mm-f11-review/

"There are some restrictions: at f11 and smaller, the AF area is reduced from the full frame to a smaller region"

Looks like I'm good at 1200mm with my 600f4+2xIII converter. At least I hope so ;-)

Click to expand...

Vunite said:

gavin said:

OK for almost static subject. The video from the safari eye AF is much more interesting and also much tougher lighting.

Click to expand...

Wouldn't necessarily say so. The tracking seems to work very well in movement, as long as the eye is a proper rounded eye. It seems to have a few difficulties with the flamingo at sharper angles. I guess the software looks for the circle. I wonder how it works on slit snake eyes and other non humanlike eyes.

Click to expand...

ericbowles said:

Vunite said:

gavin said:

OK for almost static subject. The video from the safari eye AF is much more interesting and also much tougher lighting.

Click to expand...

Wouldn't necessarily say so. The tracking seems to work very well in movement, as long as the eye is a proper rounded eye. It seems to have a few difficulties with the flamingo at sharper angles. I guess the software looks for the circle. I wonder how it works on slit snake eyes and other non humanlike eyes.

Click to expand...

The specific animal, it's color, and the shape of the eyes make a big difference in the Nikon version of animal eye detect. I found some subjects were great, but others - particularly with a brindle coat or solid black eye where challenging. I've done a lot of testing, and found that some subjects had near perfect eye-detect, while others missed 100% of the time.

This is a function that will require a lot of subject specific testing. I've seen cases where the Nikon version performs amazingly well, but other subjects in the same session won't work at all. The rule of thumb is if I can focus on the eye, it's better than delegating to the camera.

Click to expand...

Vunite said:

ericbowles said:

Vunite said:

gavin said:

OK for almost static subject. The video from the safari eye AF is much more interesting and also much tougher lighting.

Click to expand...

Wouldn't necessarily say so. The tracking seems to work very well in movement, as long as the eye is a proper rounded eye. It seems to have a few difficulties with the flamingo at sharper angles. I guess the software looks for the circle. I wonder how it works on slit snake eyes and other non humanlike eyes.

Click to expand...

The specific animal, it's color, and the shape of the eyes make a big difference in the Nikon version of animal eye detect. I found some subjects were great, but others - particularly with a brindle coat or solid black eye where challenging. I've done a lot of testing, and found that some subjects had near perfect eye-detect, while others missed 100% of the time.

This is a function that will require a lot of subject specific testing. I've seen cases where the Nikon version performs amazingly well, but other subjects in the same session won't work at all. The rule of thumb is if I can focus on the eye, it's better than delegating to the camera.

Click to expand...

It's either just the angle, but I thought it had a bit more trouble with the flamingo.

Click to expand...