no_connection

You can frame-step through some parts and still see the problem frame by frame.

Having a p frame codec would not result in dropped frames when writing to card but corrupt data and generally weird stuff.

If you framestep through the part at 1:42 going under the bridge it seems to speed up then slow down, taking small jumps then a big jump then small jump then two big jumps etc. Scene after shows the same type of thing. Does not seem to be OIS ether as even traveling large amount by boat in a line causes it.

It looks like the capture timing have huge jitter to it.  Someone need to put a counter in front if the camera and check. Longer shutter may lessen the effect and look better but not remove it. We need evidence in a controlled environment  Even a big rotating thing like a record player then map out the points of rotation and see if gaps are different distance apart would work.
 

youtube video seems to be 29.97fps

I want to see a battery grip with one in it.
And also full sized HDMI or SDI out.

4k60p means less rolling shutter as it can only have 16.6ms to scan worst case.
Hopefully the yello era is over.

It does say "on sensor", and "correction". IBIS would be more like "prevention". But who knows.

No. IBIS would mean moving the sensor relative to the body. At least that is how I would define it as it's built in to the body to provide a "stationary" sensor.

On sensor would mean that there is no physical movement involved. (unless there is a sensor available that somehow moved internally).

Note the huge difference. Same as rolling vs global shutter. Not exactly interchangeable even tho both blacks out the sensor and provides the ability to read out data.

Wonder how "ON sensor 5-axis camera shake correction" compared to real IBIS. And how it works with long shutter speeds.

I was hoping you would get one. Now I don't have to look for any other review anywhere for video.
Please get some samples to this guy http://www.dvxuser.com/V6/showthread.php?303559-Measuring-rolling-shutter-put-a-number-on-this-issue!

Any 1080p 24/25 and 50/60?
 Any way you could link some short native clips out of the camera?

Good to see some material appearing. Anyone found any 1080p samples?
Rolling shutter is obviously gonna be a problem at times, so it needs a good 1080p for those situations to be worth it. If not it's just a a6300 all over again.

That is some horrible framerate conversion, and post image stabilization.
Why is is some shots converted to 30fps, and shot in who knows what framerate while others are shot in 30fps. At least some of the shots that was shot in 30fps looks decent apart from high shutter speed.

Rolling shutter is a compromise you have to work around, not throw it in to the worst situation and try fixing it in post. Upscaled 1080p would have been a better choice for some of the shots.

4 minutes ago, Sten said:

CGI? no 4K video on D5s

 

It's a RED Epic,

http://www.nasa.gov/mission_pages/station/research/news/red_epic_dragon_camera

Not timelapse but 4k from space.

With some samples from production firmware.
From what I get the 10min limit without grip is with HDMI out too. But who knows until proper tested.

You get what I mean, as high as it have to go to keep exposure in sunshine. Although shutter speed would actually be referring to the rolling shutter time, and should be called shutter duration. Or exposure time if you want. But then speed would be ISO.
 Or maybe how fast the sensor travels relative to the speed of light.

I think that anything with 0* shutter will have bad motion cadence.

However it does feel like some/all of it is frame rate converted from 30p or something like that as it kind of skipps ahead in motion like it drops frames. Or maybe have the speed manipulated in post.

You can also see some shots have been stabilized in post. Notably first one after water.

No idea on settings used. It's not horrible but not perfect ether.
https://***URL removed***/reviews/image-comparison/fullscreen?attr29_0=panasonic_dmcgx85&attr29_1=fujifilm_xt2&attr72_0=4k&attr72_1=4k&normalization=full&widget=383&x=0.20203099946552644&y=-0.5807423640597177

I so hope they fixed the freezing problem.

This is the tool I have used:
http://www.videohelp.com/software/recover-mp4-to-h264

There is also a linux only tool, command line that was a bit if work to get running. Ill link it if I can find it again.

Did you try enable brute force in options? Was a while ago I used it but it might spew out more files to try. I have not found any better than Photorec tho.

*edit* changing extension to .mp4 don't work? (guessing no)

*edit2* Can't remember if it was this tool I used, have to look closer, http://vcg.isti.cnr.it/~ponchio/untrunc.php

but it could possibly fix truncated files that have been written over partially.

"Noticeable aliasing/moiré in some situations while shooting in 4K. In full HD/60p, it is more evident."

I think that line worries me. And even worse in 1080p. I could live with the lower resolution for things needing fast movement if rolling shutter is well behaved.
Although I don't see any problems where I would expect them in the video linked.
We need proper out of camera 4k and HD samples.

And this guy to give a number on rolling shutter:http://www.dvxuser.com/V6/showthread.php?303559-Measuring-rolling-shutter-put-a-number-on-this-issue!

No idea how to do it on a mac but check the SMART log on the drive. Maybe it's doing a short test, not that they should be done that often anyway.. You can also see if it is spinning down or not.

Something feels wrong with that clip. Apart from being clipped, less detailed than the out of camera and what looks like alasing. But could be bad compression too.
Maybe it's soft because it hit diffraction limit?
Noise could be there to reduce banding, but without more info it's hard to know.
Also notice the color banding/blocks/rainbows on the left chair and floor tiles. Without a lossless look at the signal coming from the port it's hard to judge the camera properly. But at least the feature is there for those situations that need it.

I am a lot more impressed by the potential than the demo videos. I know they are probably meant to show off the film simulation and that's fine, it even looks good at times.
But they seem to lack that cinematic planning, like careful shutter speed to give smooth motion. and ND to keep with that and still get desired DOF. Unless you want the sharp shutter creatively, I think there is a famous movie for that.

From my limited experience it's easier to remove some saturation than add it. For the Hobbit they painted Mirkwood Forest in psychedelic colors then toned them down in post to have more options.

I think all of them are reading the file correctly, but the two top apps think they are sending the signal to a TV and adjusting the output to match. If they read the file as 16-255 it would ether be left alone or being clipped/expanded to 0-255.
If you are using nVidia card, check settings and see if you can force output to be 0-255 range, it might be that and hardware acceleration that causes it.
I use MPC-HC so I can't help you with quicktime or VLC.

Looks like the right version is remapped to 16-235 range.