NX500 4K Rolling Shutter?

[mnewxcv]

I remember reading that the 4K of the NX500 had superior rolling shutter to the NX1, and I believe to other cameras as well. I cannot find a thread or the source of this information though. Can anyone cite a scientific source (controlled test on a camera site) that reveals this? I know the 1080p and 2.5k footage is marginally worse than the NX1 1080p. 

[PannySVHS]

thanks for the question! I second it. Also, how is the 2.5k quality, bitrate and codec quality for grading? love the robust panasonic codecs. the 8bit 100mbit hd codec of the gh4 is really cool.

interested, if nx500 would offer this, paired with an awesome readout speed, tiny rolling shutter. give me gh3 quality with a 7ms readout. Thank you!

[KnightsFan]

I haven't seen anything specifically about rolling shutter on the NX500. So take this with a grain of salt...

The NX500 4K is a 1:1 crop, unlike the NX1's full pixel readout. Rolling shutter is likely proportional to the crop size, since it's the same sensor. That would estimate to be ~19.3ms for UHD. (NX1 reads 3650 vertical lines in 32.6ms, NX500 reads 2160 vertical lines. 2160/3650 * 32.6 = 19.3).

[mnewxcv]

42 minutes ago, KnightsFan said:

I haven't seen anything specifically about rolling shutter on the NX500. So take this with a grain of salt...

The NX500 4K is a 1:1 crop, unlike the NX1's full pixel readout. Rolling shutter is likely proportional to the crop size, since it's the same sensor. That would estimate to be ~19.3ms for UHD. (NX1 reads 3650 vertical lines in 32.6ms, NX500 reads 2160 vertical lines. 2160/3650 * 32.6 = 19.3).

user Otto K on another forum posted the following:

 

Quote

In light of recent popularity of talking about the rolling shutter, I did some measurements on my NX500 (crude measurements, FWIW)

1080p30 20ms

1080p60 20ms

UHD@30 17ms

DC4k@24 20ms

DC4k@24 - hacked to 2.5k@24 8ms

Still capture 33ms

Two interesting things are that next frame in FHD@60 begins before the previous one ends and very fast 2.5k (suggests line skipping or similar).

I'm heavily modifying my camera so I'm interested in what others can find out

so it would seem you were pretty much on the money with your estimate. That is the only info I have been able to find on nx500 rolling shutter, though interesting that the 2.5k is so low, and also that the 1080p is significantly worse than the nx1 (7.9ms according to this thread: http://www.dvxuser.com/V6/showthread.php?303559-Measuring-rolling-shutter-put-a-number-on-this-issue! )

[KnightsFan]

@mnewxcv

Sensors usually have a number of different readout modes with different resolutions, noise levels, and RS values. Samsung might be using a different readout method for HD on the NX1 and the NX500. 120fps  requires a readout under 8.333ms, and Samsung might have just made all HD framerates on the NX1 use that same 7.9ms mode, for whatever reason. The NX500 can't do 120, so there's no reason to use that readout mode at all.

BUT...

Tbh I am not sure whether the DVXuser test measured the NX1 HD in all frame rates. We'd have to dig through the thread to find out. It's possible that the NX1 has 7.9ms in 1080p120, and 20ms in 1080p30.

[mnewxcv]

1 hour ago, KnightsFan said:

@mnewxcv

Sensors usually have a number of different readout modes with different resolutions, noise levels, and RS values. Samsung might be using a different readout method for HD on the NX1 and the NX500. 120fps  requires a readout under 8.333ms, and Samsung might have just made all HD framerates on the NX1 use that same 7.9ms mode, for whatever reason. The NX500 can't do 120, so there's no reason to use that readout mode at all.

BUT...

Tbh I am not sure whether the DVXuser test measured the NX1 HD in all frame rates. We'd have to dig through the thread to find out. It's possible that the NX1 has 7.9ms in 1080p120, and 20ms in 1080p30.

Good point. Why is 8.33ms the magic number and not 16.66? Unless I'm doing my math wrong. 

 

Edit: and wouldn't you know I did my math wrong. 8.33ms it is ?

[Mokara]

On 8/6/2019 at 1:06 PM, mnewxcv said:

user Otto K on another forum posted the following:

 

so it would seem you were pretty much on the money with your estimate. That is the only info I have been able to find on nx500 rolling shutter, though interesting that the 2.5k is so low, and also that the 1080p is significantly worse than the nx1 (7.9ms according to this thread: http://www.dvxuser.com/V6/showthread.php?303559-Measuring-rolling-shutter-put-a-number-on-this-issue! )

60p can't be worse than 16.7ms because that would require you to start the next exposure before reading the last one, so their measurement technique is likely flawed. If it was flawed for that measurement, chances are it was flawed for the other frame rate/resolution combos as well.

[KnightsFan]

23 minutes ago, Mokara said:

60p can't be worse than 16.7ms because that would require you to start the next exposure before reading the last one, so their measurement technique is likely flawed. If it was flawed for that measurement, chances are it was flawed for the other frame rate/resolution combos as well.

Otto mentions this in his post. It is very curious. If anyone has a hacked NX500, it would be fun to investigate further.

[mnewxcv]

On 9/23/2019 at 8:33 PM, KnightsFan said:

Otto mentions this in his post. It is very curious. If anyone has a hacked NX500, it would be fun to investigate further.

I just bought an nx500. How can I test?

[KnightsFan]

55 minutes ago, mnewxcv said:

I just bought an nx500. How can I test?

The OP here has an easy method: http://www.dvxuser.com/V6/showthread.php?303559-Measuring-rolling-shutter-put-a-number-on-this-issue!

There's some error of course, but his measurements line up with manufacturer claims in the rare cases that manufacturers publish their numbers, so it should be reasonably accurate. I'm too busy this week, but if you get the footage I'd be happy to try to do the measurements next weekend.