GH5 How interested are you in a GH6 anymore?

The AF still depends on a number of the AF points. Sigma fpl has phase detect (only) and its AF still sucks.

You have to add more context to that statement...fp L is some kind of sick experiment with 49 focus points. GH5 already has 200+.

Half as reliable is worthless. I'd like it to be more reliable than it already is

Nah, half as reliable means 70D-type DPAF which is many times more reliable than the current Panasonic standard. (Only the new RF cameras have deep learning.)

If Panasonic AF can stop bokeh twitching and pass a mild patience test then that is a major improvement.
 
I dont know about 70D af. I only know the A7SIII and if its AF were half as reliable then I wouldn't be using it at all
 
I don't know about that...I think you would still give it a try because you didn't have continuous video AF until last year for your entire life.

Reliability can mean different things. Part of it is a hit/miss ratio but the other part is just simply holding focus.

What makes AF unreliable is when it's twitching 90% of the time. But if it can hold focus perfectly even for 50% of the time than that is more usable than having AF that twitches 90% of the time every few seconds.
 
I was aware of video AF capabilities and chose not to seek it out. And if your AF is blowing focus at regular intervals or shifting subjects to the wrong thing regularly, then it's useless. The footage won't be usable
 
It's only useless if a bride walks down the aisle once and you miss the shot...but not useless if you're hired to capture random b-roll for the entire day.

Yeah, you might miss some shots but you'll have plenty of usable footage, including many wide-open tracking shots that the machine will most likely do a better job with.

Nevertheless...this conversation is more useless because I don't think you'd be interested in a GH6 anyway, haha.
 
True, different shooting scenarios have different needs. But even just for b-roll, you need the error frequency to be lower than the minimal shot length, be it 3 ir 5 seconds or whatever. And using AF ideally won't make your day take longer due to extra blown takes

My personal need for a gh6 is irrelevant to the abstract discussion of capabilities. But if it had 8K and I needed 8K, that could be useful. I still have 3 GH cameras and accessories so it would be an easy add-on
 
You've been thinking about the FX6 a lot lately...I can hear it...oh, the possibilities with that electronic-ND...
 
You have to add more context to that statement...fp L is some kind of sick experiment with 49 focus points. GH5 already has 200+ ...

I posted the data from all Sony made sensors on some other thread. The thing is no one knows how much is needed. Olympus M1 MK III has 121 PD and 121 CD and its AF performance is OK. GH-5 has 225 CD only and it's also OK (but not good). Fujifilm XT-4 has 425 PD only. And Phil Bloom says it's fine.
 
You make a lot of quirky jokes (which I like) along with all of the spec talk, but PB didn't say that unless he changed his mind or it was for photography...he actually made a great video about it.

 
You've been thinking about the FX6 a lot lately...I can hear it...oh, the possibilities with that electronic-ND...

I'd love an FX6 to replace my fs7 but the fs7 is still in demand. And my A7 S3 it's probably better Suited to gimbal work which is what I primarily use it for than the fx6

But I've been talking myself out of buying another A7 S3 and being patient enough to see what the A7iv looks like
 
Last edited:
I watched half of it. Then I had to answer a phone call. So I had to turn Phil off. The half I saw was fine (or reasonably fine).

Still, A6600, which has the same sensor size, has 425 PD + 425 CD and is more than fine. MFT is 2/3rd of the APS-C area, so something in the mid-500's, with PD and CD sharing the load, should be the goal. Olympus's 242 isn't enough. The question is whether the "equal share" is also the goal.
 
Actually what you originally said truly makes the most sense...we'll just never know what or how much is needed regardless if there is some sort of optimum because they can code the camera to work well or not work well.

That's really the critical-thinking bottom line as they control everything since no other country makes photography cameras.

And the Americans, Germans and Australians don't have continuous video AF (we'll see how RED does but they are still using Canon's mount).
 
Forgot the Chinese...they are a major part of camera-making but don't know where they stand with video AF these days (Z CAM, Kinefinity).
 
Technically speaking, a LiDAR based AF could serve as an alternative. It is center frame oriented, however. Samsung put it on their 108 MPX sensor because CD twitched and they couldn't add PD. It works reasonably well, with the previously mentioned caveats. The Japanese could, obviously, do it too. They could also have the Sony type CD+PD system ... but only if they bought it from someone else.

PS. Canon has its own system since makes its own sensors. And its dual pixel AF works very well on all its models, which gives the company a tremendous advantage. Unless those PD cross-points is what makes R5 shut down after fifteen minutes of operation (is that quirky enough?).
 
they're using stereoscopic depth-mapping AF on high-end focus puller systems according to brawley. i imagine that the advantage over lidar is that you can select an item in the frame rather than just blasting out a laser and hoping that it hits the right thing (or whatever is in the center of your frame. i always laugh a bit seeing these guys test lidar on an object/person that they scrupulously keep in the crosshairs)
 
Watched quite a few DJI LiDAR AF clips on YouTube. The system has three limitations - it's largely confined to the center-of-the-frame, it has a limited practical distance of about 20 feet (6 meters) and it always hits the closest subject in line of sight first.

Those caveats aside, it tends to nail the focus each and every time.
 
Back
Top