Syncing PAL AUDIO with NTSC VIDEO? (I think)

Hi I have an NTSC Hour long 23.98fps MKV file which I want to edit. I converted the MKV file to Apple Pro Res Proxy (for hard disk reasons).

I used MPEGStreamclip to do this, I didn’t change the framerate to PAL or anything, as I plan on putting anything I do online. However the audio always cut out around 12 mins into the footage ( I tried this a few times).

To save time converting again I decided to just export the audio from the MKV file instead and sync it up in FCP. But after exporting the audio I have found out that the audio is slightly longer (not by much, but enough to go out of sync). I assume this is down to the framerate, I am from the UK so I am assuming MPEGStreamclip exported the audio with PAL time frame in mind, and not keeping the original NTSC time frame. I know audio doesn’t have a set frame rate, but it is possible that is what has happened?

Length of video and audio in FCP
Video Length: 01:05:20:21
Audio Length: 01:05:24:22

Anyway, I have heard audacity is the way to go about it. But I edit video, I don’t know a lot by audio and audacity. And would appreciate some help.

Cheers

Mac OS X 10.6.8
Audacity 2.0.0

Ignoring the frames for a second because they’re really close that’s 3924 seconds for the longer one. Converting using 24.000 to 23.976 (1000/1001) gives you 3920 which is your other number. So one of your systems was running at 24 and the other at 23.976.

I was an elf on the FCP Pro forum for years and I still have an account.

Koz

Audacity has an Effect called Change Tempo. You could use this to shorten the run time of the audio so that it matches the run time of the video. Whether this simple “fix” will be suitable is going to depend on your need for “precise” synchronisation because of sound effects and/or lip-synch.

Can Change Tempo work with that small an error? 1000/1001 is not a very big error and it produces a correction value that never comes out even. I think it’s far better to find out which tool is assuming the wrong framerate and fix it there. My first pick is FFMpeg. I seem to recall someone complaining that FFMpeg had no provision for working at the older NTSC framerates. Koz

“Change Speed” can.

Cheers for the replies. I used Change temp and input the difference as 1.024 (i think), I’m on another computer now so can’t see the exact number I input, but after doing that I imported it into FCP and the time is exactly the same now. It still seems slightly off but I think all it needs now is moving a frame or two either way.

I still don’t really understand why the audio was slightly longer in the first place though.

My guess is when FFMpeg split the sound off, it produced the sound file with the wrong speed in the header, or actually produced the 48000 sample rate incorrectly. I don’t know of any way to find out in post production other than knowing the duration error. The error is perfectly correct for getting those two framerates wrong. Koz

If the complainers are correct and FFMpeg doesn’t know about 23.976, it would have produced a sound file assuming 24.000. When imported into a 23.976 project, it would have played slow. Koz