Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Video not smooth (it's jerky) when I move the camera ..

L

Leonardo Sergiani

Member
Joined
May 26, 2012
Messages
15
Reaction score
0
Points
1
Location
Rimini-Italy
Hello everybody
A premise, I just subscribed to the forum, for the first time, as well as recently in possession of this fantastic Scarlet.
I come from a past use of the camcorder so I'm not very expert ... I try to learn how to optimize this jewel in my hands.
Sorry for the questions that will I do if I seem stupid, even for the mistakes in my English, I'm Italian!
Coming to my question.
I noted that, the shots I'll do with the Scarlet at 24 fps (1/48th sec) are quite jerky, not smooth when I move the camera, for example, from right to left(even at 25 and 30, but less at 25 and more less at 30).
Someone could tell me if I'm wrong settings, or may depend on what this annoying thing?
Thanks in advance for any clarification
Leonardo
 
When you pan the camera fast at 24fps , jerk is normal. Please see some youtube 24fps videos. they also look same. you need to take care to move the camera slow if you want to shoot at 24fps.
 
Frame jitter can occur with progressive frame rates, like the Scarlet uses, during quick pans. (which isn't an issue with interlaced frame rates of video cameras that you may be used to.)
The "jerky" frame will be much less apparent if your following a moving subject through the frame while panning.

See the link below for the ASC recommend panning speeds. For example... shooting 24fps on a 50mm lens should take no less then 23 seconds when panning the camera 90 degrees.
http://4.bp.blogspot.com/_P_2wATt-I1U/Syz7yx4iPHI/AAAAAAAAABs/-xuMAvRwRHo/s1600-h/panning.JPG
 
Old digital cinema issue someone trying to solve pushing to 'standardize' 48p (ehm... Peter...) or different shutter angles. Btw the blurring/strobo/etc can be compensate 1) with slow pans 2) with shutter as 1/120 etc 3) trying not to include in framing very edged and contrasted scenes (like black square on white etc)...
 
Last edited:
Thanks to all for the precious suggestions.
I started to understand a little more about this issue that I believe will already be well known!
I will continue to text to get the best results with my Scarlet, and waiting for 48p, ok !!
 
Mark Wherley said:
Frame jitter can occur with progressive frame rates, like the Scarlet uses, during quick pans. (which isn't an issue with interlaced frame rates of video cameras that you may be used to.)
The "jerky" frame will be much less apparent if your following a moving subject through the frame while panning.

See the link below for the ASC recommend panning speeds. For example... shooting 24fps on a 50mm lens should take no less then 23 seconds when panning the camera 90 degrees.
http://4.bp.blogspot.com/_P_2wATt-I1U/Syz7yx4iPHI/AAAAAAAAABs/-xuMAvRwRHo/s1600-h/panning.JPG
Click to expand...

Mark, Thanks for that input and link to panning speeds. I''ll find that very useful
 
Mike Lary said:
Do you have solid data to back those numbers or are you pulling them exclusively from Eric Pascarelli's independent test results?
Click to expand...
Those tests were solid. The math was solid.

I ran the results by the developer of Synth Eyes (used in Dragon Tattoo, Mission Impossible Ghost Protocol, etc.) looking for errors.

The clock is 3x slower in the Scarlet.

That's why you can't record as fast, and why it takes longer to pull a frame off the sensor, thus the 3x increased read/reset rate.

That's why the Scarlet will need a lot of upgrades to run Dragon.
 
Guys, just freakin pan your cameras and move on! If you don't like the look of 24p that's not Scarlet's problem.
 
Les C. said:
Those tests were solid. The math was solid.
Click to expand...

Maybe so, but you're touting the results of a single independent test as fact. I'd like to see more to back that up, preferably from someone at Red.
 
Jerky playback of 24fps or 23.976fps footage can be due to the monitor refresh not being a multiple of the camera frame rate, such as 48fps, 72fps, 96fps, 120fps.

Even if the graphics card playing back the image is going near a multiple, the computer may not be re-refreshing the graphics card buffer in sync with the Vblank all the time, and if you are playing back from a harddrive you may be dropping frames from time to time.

Also the monitor may have its own buffer that may not be in sync all the time with the graphics card or the computer, so jutter may occur and if the camera is panning you see the jutter more than when the camera is not panning.

Just about the only playback that you can be sure of for non-stop 100% frame accurate display is a 35mm film print shot stop motion in a stop motion film recorder, like taken from DPX frames that have a 1:1 frame relationship to the camera 'raw' frames.

If you are watching a compressed playback that used keyframe compression methods the whole screen may NOT refresh at full resolution and at the shooting frame rate, the screen may only refresh at 8fps and so when the camera pans the image can drop to lower resolution or break into noisy blocks, or the motion estimation may fail and the image areas slide around in an unnatural way.
 
Mike Lary said:


Maybe so, but you're touting the results of a single independent test as fact. I'd like to see more to back that up, preferably from someone at Red.
Click to expand...
The methods and calculations are clearly detailed in the thread.

The thread I quoted had 7,734 views. RED has read the thread.

RED has a tendency to chime in on incorrect information about a RED camera.

They are not publicly disclosing the rolling shutter rate.

You tend to brag about figures that are good, so ... come to your own conclusions.
 
Les and others,

I do think my test is correct, and the Scarlet has about 14ms read/reset time, which is disappointingly slow. I've repeated it several times and performed it on several cameras. And I haven't heard from RED about it being inaccurate.

But rolling shutter should not noticeably contribute to judder which is what this thread is about.

Better explanations are mentioned above: 24fps played on a 60Hz monitor can create a garish 3-2 pulldown (because it's frames not fields that are being repeated) and 24fps in general has judder which needs to be prevented through judicious camera operation. These camera techniques have been published in the ASC Manual for decades.

Also, digital cameras in general (whether rolling or global shutter) at 24p tend to show a bit more judder than film cameras for a couple of reasons.

One of the more interesting reasons is the "sharpness" of the motion blur's edge. Yes, film and digital have equal amount of motion blur at 180° shutters, but the blur itself looks slightly different. On film the blur has a soft edge because the shutter is moving across the film somewhat above the film plane (so it is defocused) and in digital the sensor is being scanned with absolute sharpness. The motion blur starts and ends with a hard edge and this contributes to judder.

But regarding the OP, I would look at mismatches in your monitoring frequencies and also your camera technique. Rolling shutter should have nothing to do with it, although it's undesirable for other reasons.
 
Last edited:
You must log in or register to reply here.
Back
Top