Audio conversion 25 fps to 29.97 fps

[vladp]

Hi,

Unfortunately, I cannot post this question in Audio, so I can post only here.

I have the U.S. DVD in NTFS format (29.97 fps) and audio track in my native language as a WAV file extracted from PAL MPEG. Now I want to overlay this audio track over the DVD movie. I have no problems with authoring, the only problem I have is the synchronisation of this file. In order to do this, I need first to convert this WAV file, or in other words to stretch it to match 29.97 fps. How can I do this?

Thanks in advance,
Vlad.

[steveryan]

BeSweetGUI

[vladp]

Thank you for this link. But could you give me step-to-step instructions how to convert the WAV file recorded in 25 fps to the NTSC (29.97 fps) DVD using BeSweet? Or do you know the tutorial on this issue?

[vladp]

Okay, I did it with BeSweet. However, it changed the pitch of the voice. Is there any solution for this problem?

[manono]

Probably 99 times out of 100 you'll want to convert your 25fps file to 23.976fps, and not 29.97fps (which, as you noticed, makes the audio too high pitched). The BeSweet GUI has a Preset for PAL->NTSC (25.000 to 23.976). As for a guide, it's so easy to use that you don't need one, but maybe try this one:

https://forum.videohelp.com/viewtopic.php?t=129419

Here's Doom9's guide using BeSweet with the even easier-to-use BeLight GUI:

http://www.doom9.org/belight-ac3toac3.htm

[vladp]

Thanks manono. Yes, it seems much better. I wonder why the DVD is 29.97 fps and audio conversion produces better results when converting 25.000 to 23.976. Still, there is a problem of asynchronisation after about 32 min of the movie. Should I fix it manually?

[guns1inger]

The DVD is usually 23.976 fps with 2:3 pulldown applied. The player, and many apps, will interpret this as 29.97 fps for playback. G-spot will confirm what it actually is. If you are having sync issues it may actually be 29.97 fps (most NTSC video is recorded and encoded at this speed), in which case you need a better converter.

The other thing to understand is that audio doesn't really have frames per second. This is purely a video and film reference. Audio has a running time. Any good audio editing application should be able to time stretch the audio to match a specific running time, and also pitch correct it at the same time so it still sounds right. You do need to be careful if you stretch it too far, as this can introduce sampling errors, however for PAL to NTSC conversion it should be OK if the source is clean.

GoldWave, Sound Forge and Audition can all do this, and do it well. I don't know if Audacity can.

[vladp]

What is the ratio of stretching should I use in this case? 25/23.976?

[manono]

Still, there is a problem of asynchronisation after about 32 min of the movie. Should I fix it manually?

Hehe, if you want it fixed you'll have to do something, I guess. Maybe there's a glitch in either the video or the audio that causes this to happen. Or maybe one of the videos has some frames added or deleted at that point. First I would try and find the exact place where this happens, try and figure out why it happened, and try and figure out the exact amount of the delay.

Assuming nothing is corrupt, if reencoding the video, you can add or remove frames to bring them back into synch. If you want to fix the audio, you can split it at the point where the problem begins (I use HeadAC3he to split AC3 audio) and either cut however much needs to be cut, or add silence to pad out the audio (I use DelayCut for this). Then you can rejoin the audio pieces at the authoring stage (I use Muxman for this).

These procedures can be a bit tricky, so I hope you've had some experience. Good luck.