Can the Z6 autofocus become as good as the A7m3 autofocus with the upcoming firmware update ? Locked

Started 5 months ago | Discussions thread
This thread is locked.
Alex Permit
Alex Permit Senior Member • Posts: 2,338
Re: First try with skaters: I'm impressed

Matsu wrote:

Marianne Oelund wrote:

Matsu wrote:

Marianne Oelund wrote:

Gabriele Sartori wrote:

michaeladawson wrote:

No. It won’t.

but I don’t care about low light AF. What it won’t have is a revamped AF-C tracking system.

Mike, why are you so sure about that?

The sensor is virtually the same, the rest is all math.

Mike isn't confident that Nikon engineers do math as well as Sony engineers - or that they have as much budget.

Hi Marianne, I’m curious if you’ve been using a z camera, and if so, how you’re making out with it, particularly in the case continuous AF and perhaps as compared to Nikon’s 51 pt PDAF DSLRs in that regard.

I was photographing a skating competition today, so I brought the Z7 along, with a 70-200/2.8 VR II mounted, to try out during the warmup periods.

I started off with AF-C single point, and I have to say it was just a joke. Couldn't stay with my subject and I was spending more time fighting to keep my subject in focus, than watching for moments to photograph.

But then I switched to Dynamic Area, and that made all the difference. Snappy acquisition and no problems tracking skaters without jumping off to the background. I was frankly surprised at how well it worked, and I would rate it good enough to use for customers' photos. It's still not quite a D5, but it could certainly compete with a D3.

I was going to try the AF wide area modes also, but it was a short competition and I ran out of opportunities. In the second half of May I have several major competitions to photograph, so I'll bring the Z7 along again for more practice.

Though I have a rudimentary understanding of human perception, I lack the scientific background/lingo to explain the magic of engineering sensors and/or AF. Still it seems that if a single point (pixel) from the image plane is used, what is gained in precision is sacrificed in sensitivity? Compared to discrete PDAF modules in DSLRs, individual AF pixels on MILC would appear significantly smaller, so to achieve the same apparent responsiveness to the user behind the camera, different mathematics is required, perhaps aided with as info from adjacent pixels as possible - creating by some form of summation (spatial and temporal) a larger sensing area? Pardon the imprecision of my wording here. Certainly, even a small cluster of pixels is a very small area from which to judge focus. I wonder how many pixels are recruited to create the focus point on a sensor? Just one? Some small group representing a fixed area of the sensor but larger than just a single pixel?

It sort of superficially makes sense that wider area modes should work better at tracking. It will be interesting to see how much the math can be improved via firmware to get to results quicker more reliably constrained by the age at which data can both be read from the sensor, and the DSP in the camera.

I guess we could know theoretically how much either Nikon or Sony performance could be changed relative to the other if we had information about the throughput of the sensors and processing power of each cameras chipset?

"Single point"  on a MILC uses many on sensor autofocus pixels that are aggregated to a single virtual point.  This is different from DSLR's, which use one relatively large sensor per autofocus point.

 Alex Permit's gear list:Alex Permit's gear list
Sony RX100 V Fujifilm X100F Sony a7R II Nikon D500 Nikon D5 +24 more
Post (hide subjects) Posted by
MOD Mako2011
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow