GH5 Can anyone test and confirm these improved AF results with GH5?

m43user said:
I think the "bug" is there is some hidden step that takes up time. So if your shutter angle or shutter speed is equal to 1/2 the time of the cycle, with that extra hidden time, there is no time left to perform AF again during the cycle. the shutter speed must be faster (take LESS time) than the time for 1/2 a cycle. Then it can be performed again in the 2nd half of the cycle.

Am I wrong thinking that in 30p, 1/60 is you get the slower/bad AF not the "fast/better" AF? I was under the impression it does not matter if shutter speed or shutter angle is used. I could be wrong, so if I am can someone correct this?
Click to expand...

I tested it out with my GH5 and got the same results as the video I linked, which is that shooting 30p and setting 1/60 shutter speed will result in the faster AF. My theory was that when you're shooting with a shutter angle of 180, you're actually getting 1/59.94 instead of 1/60, which is why you're getting slow AF with in 180 shutter angle.

As you said, the shutter speed has to be faster than 1/59.94, but in shutter angle it's not faster, it's exactly 1/59.94. But when you're in shutter speed, 1/60 is faster than 1/59.94.
The original poster/YodaYeo of the video actually commented this over on 43rumors "...the fact that 1/60 is fast and 180 ° is slow is actually down to time and nothing else. 1/60sec is faster than 180°, 180° is 1/59.94 making the second af scan run."
So it seems like my idea was right.
This still doesn't explain what's going on with 25p though, as I get the fast AF in 1/50, and 25p isn't like 24.97fps or something.

But I really think a lot of people may be getting confused about this whole thing (I know I was). It just doesn't apply to you if you were shooting in shutter speed mode, because if you shot 25p 1/50 or 30p 1/60 you would already be getting this "fast" AF. It's definitely an important discovery for people that were shooting in shutter angle though, since they need to switch to 179 degrees instead to get the same "fast" AF performance as people in shutter speed mode were getting. This also applies to you if you lowered your shutter and were shooting like 25p 1/40 or 30p 1/50, as now you know you'll be getting slow AF.
 
So if I do the math here after watching his follow up video:

25p with a 50hz setup = 2hz per frame

180 degree = 1hz, with a 1hz doing nothing... so you would think.... except the actual execution of 180 is too close to the edge, so even bumping down to 179 degree shutter buys enough time for the second Hz in the clock cycle to do a second AF run.

So 50/25 = 2hz per frame, shutter open half the time = an even split, but need to bump to just under 50% of the time with shutter open (179) in order to maintain a whole spare Hz for an AF run.

However with 24p at 60hz... 60/24 = 2.5hz per frame, shutter open 50% of that time = 1.25 hz per frame, leaving 75% of a clock cycle doing nothing when using a 180 degree shutter (50% of the time open).

In 24p, you would need to get that ratio, per my understanding, UNDER 1hz per frame. So if you set 150 degree shutter (i.e. shutter open 41.6% of the time per VS 50% of the time with 180 degree): 2.5hz * 41.6 = 1.04 hz per frame... STILL OVER 1hz! This is why you won't see improved AF performance at 24p at 180 degree, 179 degree, or 150 degree.

However, at that math, a 140 degree shutter angle in 24p should result in the faster AF performance because 140 degree shutter = 39% of the time spent open to capture the frame. Thus:

60hz dividded by 24p = 2.5hz per frame
2.5hz per frame * .39 (39%) of the time utilized for frame capture (ie 140 degree shutter) = .975hz per frame. This should leave 1.5hz or so "Extra" for an extra AF run.

Now, this might not be as effective as 25p/50hz BECAUSE 24p is operating @ 60hz. So what you will have at 140 degree shutter with 24p 60hz is:

1hz Image Acquistion + AF
1hz AF
.5hz camera does nothing
1hz Image Acquistion + AF
1hz AF
.5hz camera does nothing
1hz Image Acquistion + AF
1hz AF
.5hz camera does nothing


Vs. the scenario in which you have 25p 50hz and 179 shutter is:

1hz Image Acquistion + AF
1hz AF
1hz Image Acquistion + AF
1hz AF
1hz Image Acquistion + AF
1hz AF

But I am not sure in practice how that performance will play out.

But it would seem getting the camera down to 140 degree shutter in 24p 60hz would see an AF improvement, and only 179 degree in 25p 50hz is just enough to free up an extra clock cycle for AF in PAL - without affecting motion cadence very much - so that is definitely ideal.

For us, that means if we do a gimbal shot that doesn't require perfect audio sync we can now just shoot 25p at 179d for Continuous AF to track and then do a simple interpret footage to 24p in post. And even if we needed audio sync, we could manually fix it in post (a PIA, but not too bad unless a really long shot).

Now, I haven't done the math here and maybe I'm wrong to begin with (but I think the above is accurate?), but it would seem 30p in 60hz would work for a subtle slow motion footage interpret and the math should shake down roughly the same at 179 degree shutter?
 
filmguy123 said:
Now, this might not be as effective as 25p/50hz BECAUSE 24p is operating @ 60hz. So what you will have at 140 degree shutter with 24p 60hz is:
Click to expand...
*IF* his theory as to why the behavior is present is correct, and *IF* you are not shooting at true 24P (24hz system frequency.)
 
joe12south said:
and *IF* you are not shooting at true 24P (24hz system frequency.)
Click to expand...

No, but under his theory, 24p @ 24hz does not help. Here's why:

If his theory is correct, the faster AF comes from using a full extra Hz cycle for AF only. Hence, 25p @ 50hz and 180 degree shutter is actually using up 2hz cycles. You would think with 180 degree shutter it would only use up 1, but 180 at 50% actually teeters over the limit and uses a full extra Hz cycle. 179 solves this, using 1hz or half the clock cycle for the frame acquistion, and the second hz for AF.

But at 24p @ 24hz, you don't have 2hz per frame. You have 1. No matter how much you speed up the shutter angle, you'll only ever have one hz. 179 degree shutter here would edge you under a half Hz, sure, at maybe .4999 hz usage per frame, but the AF cycle doesn't work on half hertz. It needs a full.

Hence 25p @ 50hz with 179d shutter being the sweet spot.

Now flip around to 24p @ 60hz, where you get 2.5 hz per frame. Well, as I understand 24p pull down, 2 hz per frame, 3 hz per frame, 2hz per frame, 3hz per frame, for an average of 2.5. So my math may break down here a tad and this might not be spot on since I am averaging and 24p pull does not actually average from my understanding. But when averaged, that's 2.5hz per frame or 1.25 herz at 180 degree shutter, meaning you need a 140 degree shutter to get a frame to take under <1 hz, in order to free up a new hertz cycle to dedicate to auto focus. So 1 for image acquisition, 1 for AF, and then a wasted half-hertz... meaning performance will never be as good as 25p 50hz, but probably better than 24p 180 degree shutter if you can get it down to 140. This should be tested.

Where the 24p/60hz gets trickier is it's not averaged, it's every other I think, which throws a wrench and raises questions...

I realize reading both of my posts twice and thinking about the math may be required to wrap head around.

DISCLAIMER: This is all *IF* his theory is correct, and *IF* I am understanding it correctly :)

DISCLAIMER 2: Sorry if this is all wrong, just brainstorming here so we can try and crack the code as to what is happening mathematically.
 
Yup, that's right, if this theory is correct, then there is no shutter angle that will help with cinema 24P. By that same logic, though, no angle should help with 60P @ 60hz or 50P @ 50hz, either.
 
i watched these vids this morning.

i guess i am a few days behind!

i wanted offer props to the youtuber who "discovered" this possible fix.

another positive example of the power within digital communities.

i am just sayin ;-)

be well.

rob
smalltalk.productions/nyc
 
joe12south said:
Yup, that's right, if this theory is correct, then there is no shutter angle that will help with cinema 24P. By that same logic, though, no angle should help with 60P @ 60hz or 50P @ 50hz, either.
Click to expand...

Agreed. But, I am confused as to what you're saying. I did not think his fix helped 60p at 60hz or 50p at 50hz either? I think he even explained in his video how it didn't, right?

This only works on 25p 50hz or 30p 60hz regarding an improvement via shutter.

But of course 60p 60hz is better because even at standard 180 degree shutter you are using 60 cycles for AF, vs 24p at 60hz, you are only using 24 (2.5 hz per frame with no spare hz cycles for an extra AF run unless at 140 degree shutter). 60p at 60hz at 180 degree nets 60 AF cycles. 24p at 60hz at 180 degree shutter nets 24. 24p at 60hz at 140 degree shutter should net 48ish I believe. 25p at 50hz at 179 degree should net 50 af cycles.

So to me this is all cohesive?
 
filmguy123 said:
Interesting, and, maybe it's time we all wrap our heads around the digitial era. Especially with a camera like this in a digital era. Why are we using 50hz and 60hz frequency standards anyway?
Click to expand...

Thats a valid question indeed. While I live in 50hz territory, I have decided a while ago to almost only shoot in 30/60hz ss it gives some more flexibility.
 
filmguy123 said:
Why are we using 50hz and 60hz frequency standards anyway?
Click to expand...

Please correct me if I'm wrong, but I was under the impression that NTSC countries used 60Hz because the AC electricity in those countries alternated at 60Hz, and likewise for the PAL countries and 50Hz AC electricity.

This might have been a constraint or a convenience during the analog era of media production/display but I would imagine that in the digital age there doesn't necessarily need to be a connection between the "mains" AC frequency and the frequency media devices (cameras and TVs).
 
filmguy123 said:
Now flip around to 24p @ 60hz, where you get 2.5 hz per frame. Well, as I understand 24p pull down, 2 hz per frame, 3 hz per frame, 2hz per frame, 3hz per frame, for an average of 2.5. So my math may break down here a tad and this might not be spot on since I am averaging and 24p pull does not actually average from my understanding. But when averaged, that's 2.5hz per frame or 1.25 herz at 180 degree shutter, meaning you need a 140 degree shutter to get a frame to take under <1 hz, in order to free up a new hertz cycle to dedicate to auto focus. So 1 for image acquisition, 1 for AF, and then a wasted half-hertz... meaning performance will never be as good as 25p 50hz, but probably better than 24p 180 degree shutter if you can get it down to 140. This should be tested.

Where the 24p/60hz gets trickier is it's not averaged, it's every other I think, which throws a wrench and raises questions...
Click to expand...

With all of your disclaimers seconded, it seems to me that a shutter angle of less than 120 degrees should result in "fast" AF even with 24p 3:2 pull down. i.e. 119 degrees (or 1/72 rounded to 1/80) should get your image capture period under 1hz even at 3hz/frame. (And another caveat: if you're used to shooting at 24p 1/40, 1/80 is a whole stop down.)

Benjobe said:
...I would imagine that in the digital age there doesn't necessarily need to be a connection between the "mains" AC frequency and the frequency media devices (cameras and TVs).
Click to expand...

As long as your light source(s) are also not AC powered. (i.e. DC LEDs, sunlight, etc.)
 
With all of your disclaimers seconded, it seems to me that a shutter angle of less than 120 degrees should result in "fast" AF even with 24p 3:2 pull down. i.e. 119 degrees (or 1/72 rounded to 1/80) should get your image capture period under 1hz even at 3hz/frame. (And another caveat: if you're used to shooting at 24p 1/40, 1/80 is a whole stop down.)
Click to expand...
According to this theory, no, as the autofocus cycle is tied to the frame cycle. The frame lasts the entire herz cycle, no matter how long the shutter is open.
 
Agreed. But, I am confused as to what you're saying. I did not think his fix helped 60p at 60hz or 50p at 50hz either? I think he even explained in his video how it didn't, right?
Click to expand...
You're probably right, I thought that it was said that it worked at 60/50P.
 
joe12south said:
According to this theory, no, as the autofocus cycle is tied to the frame cycle. The frame lasts the entire herz cycle, no matter how long the shutter is open.
Click to expand...
But AFAIK it hasn't been established that 24p = 24Hz (like 60p = 60Hz and 50p = 50Hz), so filmguy123 and I were discussing the possibility that 24p might be 60Hz with 3:2 pull down (similar to how 30p = 60Hz, not 30Hz and 25p = 50Hz, not 25Hz, both kinda sorta but not really with 2:1 "Hz pull down" when using a 179 degree shutter (or less) ).
 
So it seems there's a new GH5/S firmware coming with AF improvements - are we thinking it will incorporate some of the findings here and make the need to run this hack obsolete?

Also, regarding AF lenses this works with, does anyone know if it is native only, or will adapted lenses such as the EF Sigma Art 18-35 via speedbooster run this AF hack well, too?
 
What is your information about a firmware update? I was thinking they would act as it is pretty embarrassing if you ask me.
 
It seems like there will be image improvements as well. For me that means less processing (less sharpening and nr), but for Panasonic it usually means more lol...:undecided
 
You must log in or register to reply here.
Back
Top