HawkEye stats, Part II
jzfredricks
any mistakes, and I�ll fix them up.
How far does the ball travel in each TV replay frame?
Ball travel per frame @ 100kmp/h (27.77m per second);
150fps = 18.5cm per frame
200fps = 13.8cm
250fps = 11.1cm
500fps = 5.5cm
FPS required to see 2.6mm movement = about 10700 fps
Ball travel per frame @ 150kmp/h (41.66m per second);
150fps = 27.7cm per frame
200fps = 20.8cm
250fps = 16.6cm
500fps = 8.33cm
FPS required to see 2.6mm movement = about 16000 fps
So at 150kmp/h and 500fps*, the ball travels 8.33cm between frames.
If the HawkEye operators manually select frames, as they can do, and
they�re out by a *single* frame� what does this 83.3mm error do to the
2.6mm margin of error?
* didn't someone mention 500fps cameras?
cricketrulez
> Again, I put these here for peer review. Please let me know if you see
>
> How far does the ball travel in each TV replay frame?
> Ball travel per frame @ 100kmp/h (27.77m per second);
> 150fps = 18.5cm per frame
> 200fps = 13.8cm
> 250fps = 11.1cm
> 500fps = 5.5cm
> FPS required to see 2.6mm movement = about 10700 fps
>
> Ball travel per frame @ 150kmp/h (41.66m per second);
> 150fps = 27.7cm per frame
> 200fps = 20.8cm
> 250fps = 16.6cm
> 500fps = 8.33cm
> FPS required to see 2.6mm movement = about 16000 fps
>
> So at 150kmp/h and 500fps*, the ball travels 8.33cm between frames.
> If the HawkEye operators manually select frames, as they can do, and
> 2.6mm margin of error?
>
> * didn't someone mention 500fps cameras?
Doesn't add up. Unless they are talking about the error in the lateral
position. This would mean that they are making a few assumptions (eg.
ball movement is strictly from wicket to wicket with some fudge factor
for lateral movement) and not strictly based on measurement alone. Can
of worms ..........
Dodo Twokay
>
> - Show quoted text -
If they have 1024x768 pixels or more lilely they have HD which is
1920x1080, and their accuracy of tracking is within 2 or 3 pixels,
then you probably get their numbers
jzfredricks
> On Jul 20, 4:50 am, jzfredricks<jzfredri...@gmail.com> wrote:
>> Again, I put these here for peer review. Please let me know if you see
>> any mistakes, and I ll fix them up.
>>
>> How far does the ball travel in each TV replay frame?
>> Ball travel per frame @ 100kmp/h (27.77m per second);
>> 150fps = 18.5cm per frame
>> 200fps = 13.8cm
>> 250fps = 11.1cm
>> 500fps = 5.5cm
>> FPS required to see 2.6mm movement = about 10700 fps
>>
>> Ball travel per frame @ 150kmp/h (41.66m per second);
>> 150fps = 27.7cm per frame
>> 200fps = 20.8cm
>> 250fps = 16.6cm
>> 500fps = 8.33cm
>> FPS required to see 2.6mm movement = about 16000 fps
>>
>> So at 150kmp/h and 500fps*, the ball travels 8.33cm between frames.
>> If the HawkEye operators manually select frames, as they can do, and
>> they re out by a *single* frame what does this 83.3mm error do to the
>> 2.6mm margin of error?
>>
>> * didn't someone mention 500fps cameras?
>
> Doesn't add up.
It certainly adds up - the maths are sound. What is important, as
always, is how the data is interpreted.
You are right - forward movement isn't anywhere near as important as
lateral movement. But unless the ball is dead straight there is an
element of both.
100kmp/h, 200fps = 13.8cm per frame.
I'd like to know how the HawkEye operator is supposed to select a
pitching spot when they've only got 1 image per 13.8cm!
cricketrulez
JZF, I was talking about the 2.5mm accuracy quoted by HE!
jzfredricks
> JZF, I was talking about the 2.5mm accuracy quoted by HE!
2.6mm.
Accuracy is important :)
Phil.
> Again, I put these here for peer review. Please let me know if you see
>
> How far does the ball travel in each TV replay frame?
> Ball travel per frame @ 100kmp/h (27.77m per second);
> 150fps = 18.5cm per frame
> 200fps = 13.8cm
> 250fps = 11.1cm
> 500fps = 5.5cm
> FPS required to see 2.6mm movement = about 10700 fps
>
> Ball travel per frame @ 150kmp/h (41.66m per second);
> 150fps = 27.7cm per frame
> 200fps = 20.8cm
> 250fps = 16.6cm
> 500fps = 8.33cm
> FPS required to see 2.6mm movement = about 16000 fps
>
> So at 150kmp/h and 500fps*, the ball travels 8.33cm between frames.
> If the HawkEye operators manually select frames, as they can do, and
> 2.6mm margin of error?
>
> * didn't someone mention 500fps cameras?
Yes I did, they're used in tennis because of the higher velocity
involved.
You have some misconceptions about how the system works however.
Using the high speed cameras at the ground every frame records the
position of the ball to the accuracy specified.
So you have 5 simultaneous images of the ball enabling the position in
three dimensions to be recorded at a series of precisely known times.
Determining the trajectories from that series of points is routine
ballistics, I explained how this could be done here about 13 years
ago! The key time for the accuracy of the position is not the time
between frames (which is important though) but the exposure time
of the camera which is much shorter thus giving high spatial
resolution on each frame.
Phil.
jzfredricks
> The key time for the accuracy of the position is not the time
> between frames (which is important though) but the exposure time
> of the camera which is much shorter thus giving high spatial
> resolution on each frame.
I'd not thought about "blur" with in each frame.
I assumed a perfect picture was taken!
I guess blur would just make it harder for the operators to succesfully
and manually select the point of impact.