Certain 59.94 clips play jittery on 23.98 timeline but play smoothly in the source window

[rachelcenter]

One tv company I work with sends me a lot of MXF footage (59.94fps) and a handful of them play smooth on my 23.98 timeline but then a handful of them play really jittery on a 23.98 timeline, so when I match back to the jittery footage, it looks smooth in the source window. So why do some clips look jittery on the timeline and others dont? The company who is sending me the footage is using a proprietary software to restore footage from an archive. And I've brought this up with them before and they said sometimes restoring as xdcam can fix the issue. Wanted to see if there was some kind of pending update on Adobe's part to fix the jittery playback on the timeline?  When I pull up the "codec" column in premiere associated with the jittery clips it says xd59, Mpeg and dvn7.

I can't publically share the footage but I can send a private link to Adobe folk in the background.

I end up avoiding the jitter by nesting the 23.98 timeline and the problematic shots in a 59.94 sequence upon export, but it's odd that some of them play fine with no issues on the 23.98 timeline while others don't (while always looking great in source monitor)

|https://community.adobe.com/t5/premiere-pro-discussions/how-can-i-determine-if-this-frame-rate-issue...

[Richard M Knight]

As I understand it the source monitor plays back any clip at its native framerate and resolution, it's only when placed in a sequence that the framerate changes.

[rachelcenter]

https://community.adobe.com/t5/premiere-pro-discussions/how-can-i-determine-if-this-frame-rate-issue...

[Kevin-Monahan]

Please provide the team with more info, @rachelcenter. See, How do I write a bug report?

I hope a developer will comment shortly.

Thanks,

Kevin

[Warren Heaton]

@rachelcenter 

Any difference in temporal resolution (frame rate) is prone to not playing back well in a Timeline - even more so when dividing the Source Footage frame rate by the Timeline frame rate does not result in a whole number.

Try enabling Optical Flow for Time Interpolation in the Speed/Duration dialog box for each 59.94 clip in your 23.98 Timeline.  You’ll want to select the Clips afterward and then choose Sequence > Render Selection.  Then play the footage with Playback Resolution set to Full.  If that looks good, be sure that Optical Flow is also enabled whenever you export.

If After Effects is part of your subscription plan then another option is to roundtrip the 59.94 Clips to After Effects where you can use Timewarp to remaster the frame rate.

Topaz Labs Video AI is worth a try as well to remaster the frame rate if you are not liking the Optical Flow or Timewarp results.  Topaz Labs offers a trial version, but if you decide you like it the best, it is a separate purchase. 

Each of these options take some time to render. While revising rough cuts, I would use a Clip Label to easily identify the 59.94 clips and then take the time to remaster those based on what’s in your fine cut.

[rachelcenter]

None of these suggestions are reliastic for a 2-3 day turn around with close to 3.5 TB of footage being sent to me. You'd have to be crazy (and out of external hard drive space) to round trip those to AE or Topaz when you could just nest the 23.98 sequence into a 59.94 for output and put the problematic 59.94 shots on layer v2 within the nest.  I don't recommend sharing this idea with anyone else. Not practical.

[Warren Heaton]

It's very common in documentary work.  If faster remastering is needed, you could look into a hardware solution by Taranex (now owned by Blackmagic Design) or Panasonic.  That, or split the rendering acrcoss multiple, fast workststions.

I didn't mention this earlier, but if you already have distribution in place make sure that the approach you're taking passes their quality control as soon as possible.  The last thing you want is for a program to get rejected due to how the archival shots were handled.  

[rachelcenter]

None of this addresses the core issue at hand.

[Warren Heaton]

Timewarp in After Effects was formerly Khronos from the Foundry.  It's one of the top software-based options for exactly this issue before machine learning solutions become available. 

Your clips play well in the Source panel because the frame rate of the clip is being used.  These clips do not play well in the Program panel because a different frame rate is being used that is not a multiple of the source frame rate.

[rachelcenter]

I disagree with your assessment. Half the 59.94 clips play fine in my timeline. The other half look jittery. It's inconsistent and doesn't make sense.

[Warren Heaton]

All i-frame formats (also called intraframe formats) tend to be more forgiving when it comes to simply dropping a p59.94 or i59.94 clip into a p23.98 Timeline.

How does Optical Flow work on the problematic clips?

Ideally, if we have control over it, our Timeline frame rate matches our delivery frame rate, our source footage with sync sound matches your Timeline frame rate, and and our b-roll without sound can be at a higher frame rate that's a multiple of the Timeline frame rate (facilitating smooth slow motion as an option in post).

[rachelcenter]

kevin can your team develop a script (no not a movie script) that will allow your team to collect all of that data with the push of a button by me?

[rachelcenter]

• Issue - stated at the top of this page
  
• Adobe Premiere Pro version number:  24.6.1 (build 2)
  
• Operating system - macOS sonoma 14.6.1
• System Info: CPU, GPU, RAM, HD:

  • CPU - not listed in mac studio system report.

  • GPU (Intel-based Macs and PCs)  not listed in mac studio system report.

  • GPU driver (PC) not listed in mac studio system report.

  • RAM not listed in mac studio system report but it says 128gb memory

  • Hard Drive (and, if external, how the drive is connected: USB 3, Thunderbolt, NAS, etc.). - 4tb ssd sandisk

• Video format: listed above in description paragraph at top of page
  
• Workflow details: played clips in source monitor then on timeline.
  
• Steps to reproduce - (Very important!) Described above in top paragraph
  
• Expected result - Described above in top paragraph, all 59.94 video should play normally on timeline
  
• Actual result -  some 59.94 plays choppey on timeline, some doesnt.
  

[Stephen Barrante]

Old post - but maybe this might help someone else experiencing a similiar issue. We discovereed based on this post - https://community.adobe.com/t5/premiere-pro-discussions/interpret-footage-from-60fps-to-24fps-result... that the footage we were provided, although encoded from camera at 59.94 fps, has in fact duplicated frames. Very easily spotted by hitting "next frame", and seeing every frame repeat. This explains why when we drop the frame rate down by 50% we do not get nice smooth "slo mo" b-roll, but a jittery mess. 

Alas we will have to use an AI tool like Topaz for select shots in order to produce our "slo-motion" frame rate on these troubled clips. While I want to blame the DP for not understanding this about the camera, I'm even more frustrated that this feature even exists. Maybe someone can explain, but it provides no value whatsoever except to simply double the file size. 

[R Neil Haugen]

This at times can happen with especially interlaced media going through framerate interpretations/changes. As actually, 59.94i and 29.97i are really the same thing ... interlaced media has two half-frames per "frame" if counted in 29.97 format.

So it can simply end up with a double frame showing here and there at times. Years ago, when interlaced was more common, this was a very widely handled issue. And most users probably had an idea off the bat what the problem was.

"We" don't all work with interlaced on a daily basis anymore, so it's a surprise.

[Warren Heaton]

@Stephen Barrante 

If the video is p59.94, there should be no duplicate frames.

If the video is i59.94 but originally 23.976 having gone through 3:2 pulldown, then there will be duplicate frames.  If i59.94 is viewed as single field, it may appear progressive when it's actually interlaced.

With digital cameras that record to tape, the 3:2 pulldown can happen in camera (hence the 3:2 pulldown removal option available during captrue from tape).  I don't think there's a tapeless camera that does 3:2 pulldown internally as it's not necessary.

Ideally, a DP shoots at the frame rate needed for the project for anything that's sync sound.  B-roll without sync sound can be shot at twice the frame rate or higher, but should always be a multiple of the sync sound frame rate.

[Stephen Barrante]

@Warren Heaton  - thanks for the additonal details. Unfortuantely we don't have any details from the DP regarding the camera setup before we received this B-roll. I opened up the footage in Quicktime as well as Premiere and it simply displays the frame-rate as 59.94 with no indication if it's interlaced or progressive in the frame-rate field or properties (also tried VLC). 

[R Neil Haugen]

Check the media data in MediaInfo ... download that free utility from their annoying page, then after install, simply drag/drop a file onto the desktop icon.

When MediaInfo then opens, go into the "Tree view", and you can see the actual file data. Posting a screengrab by drag/dropping it onto your text reply area would be good.