Panasonic LUMIX S1 – Rolling Shutter Test

August 1st, 2019
Panasonic LUMIX S1 - Rolling Shutter Test

The Panasonic LUMIX S1 full frame camera has fared very well in our dynamic range and latitude tests and we were eagerly waiting to see if it will do equally well in our rolling shutter test. Just as a reminder, this dreaded phenomenon is negatively affecting CMOS based cameras. 

Panasonic LUMIX S1

We have already published several articles on the new Panasonic S1, and the latest one which shows exceptionally good dynamic range and latitude (underexposure) results of the new VLOG profile can be found here.

Now it is time to have a look at the dreaded rolling shutter, which is a negative side effect of CMOS based cameras. For fast moving subjects across the frame, rolling shutter causes distortions, like for example vertical lines that start to tilt. Another example is the “wobbly” feel of handheld footage shot without image stabilization.

Rolling Shutter sample – When a straight object becomes not so straight anymore….

What is causing the Rolling Shutter Phenomenon?

Rolling shutter occurs because the sensor read out happens from the top to the bottom of a frame with a certain finite readout speed (the rolling shutter value in milliseconds [ms]) as opposed to global shutter where the whole frame is being read at once.

This basically means that line per line from top to bottom is exposed and read out. As it takes a while to do this, a change in the frame may have happened in this rolling shutter read out time window going from top to bottom. Hence, objects that move in the frame during that time window get distorted.

See a nice example here by cmglee (licensed under creative commons CC BY-SA 3.0):

 

For us filmmakers it’s simple: the lower the rolling shutter value is, the less distortions happen, the better the final image representation.

Rolling Shutter of the Panasonic S1 at Various Modes

We are using a new test method based on a high frequency, quartz & microcontroller based strobe light at 300Hz. As the sensor readout happens from top to bottom, we are getting a sequence of black and white bars from the strobe.

Now it is rather simple, we just have to count the pairs of black & white bars and multiply them by 1000/300 = 3.333[ms]. Hence, the lower the number of black & white bars, the lower the rolling shutter result, the better for us filmmakers.

Let’s start with UHD (3840×2160), 25 frames per seconds in full frame mode, see below. We are getting a good value of 22[ms] – As a comparison, the Sony a7 series cameras all have values higher than 25 [ms], for example.

Panasonic S1 UHD 25p full frame rolling shutter of 22[ms]

Next up is UHD 25p in APS-C mode. Big surprise: the rolling shutter more than halves, to 10.4[ms]. In UHD 50p (APS-C) the value stays the same, at 10.4[ms].

Panasonic S1 UHD 25p APS-C rolling shutter of 10.4[ms]

For comparison, other APS-C cameras like the FUJIFILM X-T3 show 19.6 [ms] for UHD 25p, micro four third cameras like the Blackmagic Pocket Cinema Camera 4K show a readout speed of 16 [ms], the GH5 shows 15 [ms].

Panasonic S1 full HD 25p APS-C rolling shutter of 7.4[ms]

The biggest surprise is now at the end: just for fun I played around with full HD in APS-C mode, and measured a stunning 7.4[ms] (full HD in 50p APS-C is the same 7.4[ms])! This is a fantastic value!

Conclusion

The more I test and more I shoot with the new Panasonic S1 full frame camera, the more I’m loving it. If it weren’t for the upcoming Panasonic S1H camera, I would have pulled the trigger already and purchased it.

The good to very good rolling shutter results in the various modes just round off what we knew already – the Panasonic S1 is a really capable filmmaking device! With the various deals that are around (including free VLOG upgrades) it is also becoming a relatively affordable one.

Are you a Panasonic S1 user? If yes, what is your experience using it so far? Please share your thoughts in the comments section below.

14
Leave a Reply

guest
Filter:
all
Sort by:
latest
Member
August 4th, 2019

The irony here is you really don’t want a lower (faster) value for lower resolution modes than higher resolution modes as it means it is changing readout modes and likely lineskipping or binning which will could mean those modes have worse aliasing and noise. This worries me about shooting HD on the S1 and warrants further testing of aliasing and noise in HD modes as they may be inferior enough so as to make them not worth using which would be a problem for anyone who plans to shoot HD on some projects.

Jürgen
Jürgen
Guest
November 2nd, 2019
Reply to  Gunther Machu

Yes, because not everyone wants 4K and more. Nobody needs it for small Webvideos.

 Charles Hull
Charles Hull
Member
August 2nd, 2019

I really like the bar-strobe test approach!

But how do you have VLOG? My upgrade key is still on backorder from B&H? I spoke with them again today and they said to expect it “sometime in August”. Has this shipped in some parts of the world? Or to other US distributors?

 Artyom Kornikov
Member
August 1st, 2019

Sony A7rIII:
– 4k FF – 22.8 ms
– 4k APS-C – 16 ms
– 1080p FF – 7.7 ms
– 1080p APS-C – 6.1 ms
source: dvxuser.com forum

Member
August 1st, 2019

cant wait to see the same test for GFX100

 Daniel Rozsnyo
Daniel Rozsnyo
Member
August 1st, 2019

Could it be measured more precisely? I do not really trust those blurry bars :)
Also – the rolling shutter time is harder to understand in msec, so when will somebody get back to the proper units for showing the scan speed – which is Hz. From that, it would be obvious, that the sensor supports 1080p in 120Hz+ scan.

What is unexpected to me, is the UHD crop scan going at 96 Hz (especially for the vlog HDR capture). Really wondering which sensor is in the S1 then.

 Walter Vargas
Walter Vargas
Member
August 2nd, 2019
Reply to  Daniel Rozsnyo

I prefer milliseconds, because this tells you the time gap from the top of the frame to the bottom. Why would you want to do that math from Hertz?

 Luke Wen
Luke Wen
Member
August 1st, 2019

The rolling shutter in FHD mode is low because it’s not full pixel readout, FHD is sub-sampled using a combination of horizontal pixel-binning and vertical line-skipping. It’s the same thing on most mirrorless cameras including X-T3 and GH5S.

Jürgen
Jürgen
Guest
November 2nd, 2019
Reply to  Luke Wen

Uhm, is there any proof for this thesis?

Filter:
all
Sort by:
latest

Take part in the CineD community experience