Re-author DVD with different audio source???

[SpankSinatra]

I want to take a dvd and re-author it and add a different audio source onto the dvd. What is the best program? Do I just use the VOB from the dvd and the wav file to do it?

[Five]

its not as simple as that, a lot of work has to be put in to get the sync right. the details of that I'll leave to one of the dvd authors to explain

ps Brian

[SpankSinatra]

[saltman]

The basics are....

use a program such as tmpgenc xpress (mpeg tools) to demultiplex the .vobs to their elementary streams. (m2v and wav or ac3 or whatever it is.)

use a program such as sony vegas video to input your old wav and new wav. align them perfectly (assumes the original was synced properly. if not, you will need to input your video source and use it as a guide also).

trim the new wav file to be exactly the same size as the old one.

use a program such as sony dvd architect to create the dvd menus with your new files. export to a video_ts structure.

[SundayDriver]

The first question to answer would be if the new audio source runs at the same speed as the old audio source.

If the source is on DVD already, I would also suggest using DVD Decrypter to extract the vob information in File mode with no splitting.

[Author85]

What everybody has said above is correct, but I'll add some more information. I've done many DVD reauthor's with alternate audio sources. My big thing is taking DVD with crappy AC-3 sourced audio and re-synching them to either better sounding PCM sources or soundboards.

First off, I don't know if you realize what a HUGE job it is to resynch a show. If you don't have timecodes to use (which is probably a 99.9% chance you don't) you are going to have to do it the old fashioned way of visually synching up the peaks on the waveforms on some kind of NLE. I've found from experience that DAT recordings are much easier to synch up than analog or cassette recordings. The reason is that DAT seems to record at a much more constant speed. Even if it is fast or slow, as long as it's not variable, you can apply a linear correction across large portions of the soundtrack. If it's analog sourced, watch out! I've had some analog sourced re-synch's that have taken 200 to 300 corrections across an entire concert.

The reason, IMO, is that in older style cassette recordings, the "record" speed seems to be pretty variable, directly dependant upon the battery supplying a constant and steady voltage source. More modern recorders, especially DAT's, seem to be less prone to a variable speed recording, and are thus easier to make timing corrections to in post.

In an NLE, I basically use the original audio soundtrack as a "guide source", since I know that it is synched correctly to the video. I then place my "new source" directly below the original "guide source" and make timing corrections to it to exactly match the graphical waveform peaks and valleys. It's all visual at this point.

Oh, also, be sure to turn your "quantize to frames" feature OFF, if you are using a video NLE for this. If you don't, then the editor will only allow you to make cuts or corrections at each video "frame", which is about 1/30 of a second. With "quantize to frames" turned off, you can make cuts and/or timing corrections at any point on the timeline, not just at every 1/30 of a second. This is EXTREMELY important, as most timing corrections cannot be done at just 1/30 of a second intervals. If you want your audio perfectly synched you must do this. If you are using an audio editor for this project, then chances are your program will default to not making cuts every 1/30 of a second, so it won't be an issue at all.

Once you graphically have the waveforms for both the "guide source" and "new source" lined up visually, then you can play both audios on the timeline at the same time to fine tune your synch job. It will be very obvious when the synch starts drifting as drumbeats and other sounds will sound off-beat. When this happens, you know you have to go back to the "new source" timeline to make some kind of timing correction, to get the source back in synch with the "guide source". This is what you have to do, sometimes every 10 to 15 seconds, sometime every 10 to 15 minutes, depending on the playback speed of the "new source". As you can see, this can get quite repetitive and time-consuming.

Now, you can make a very half-assed synch job that will sound like crap, and be noticeably out of synch, by simply "slapping" the new source on the timeline and trimming the ends without doing the above mentioned work. But because the VAST majority of audio recordings weren't recorded at a constant speed and hence don't playback at a constant speed, you won't get anything you would want to trade. If you want an absolutely perfect synch job, you must use my method above.

So, that's how I do it, and I always make an absolutely perfect synch, down to the microsecond, for my audio re-authors. If anybody has an easier way, please let me know, as I haven't found a way to do it any better without the benefit of audio/video timecodes.

Good luck!

[musik_maniak]

Originally Posted by Author85

So, that's how I do it, and I always make an absolutely perfect synch, down to the microsecond, for my audio re-authors. If anybody has an easier way, please let me know, as I haven't found a way to do it any better without the benefit of audio/video timecodes.

Thanks Author85, for an interesting post. I learned a couple of things.
It seems like a very precise method, and probably the way it should be done.

IMO, I've had good success using a method, which is obviously not as much work as what you put into it:
First of all, I make sure the sync of the original audio track is right on. If it's not, I'll fix it with TMPGEnc or something similar.
Then I extract the audio track and open it up in Cool Edit. I pick two obvious downbeats in the music; one at the beginning of the recording, and one at the end. Using the waveform peaks, I measure the exact time between those those two downbeats.
Then I open up the new audio track in Cool Edit and find the same two (beginning and end) downbeats. I adjust the speed of the new audio track (by changing the pitch in Cool Edit) until the time between those two downbeats is exactly the same as it is on the original track. This ensures that the new audio runs at the same speed (pitch).
In order to get the new track to start at the same time as the old one, I measure the time between the start of the original audio track and the first downbeat that I had selected (the "beginning" one). I then trim the beginning of the new audio track so it matches this. (I also trim the end of the new audio track, but obviously, the end portion is not as critical as the beginning.)
So that usually gives me a new audio track that is identical to the original, as far as sync is concerned.
Granted, if the audio speed varies on either the new or the old track, it would present problems, but if not, this method seems to work fine.

[SpankSinatra]

ok so i have the original ac3 and m2v file that I got from the dvd vob file. I also have a single wav of the new audio. What do I use to listen to them side by side and sync them up? the ac3 is longer than the wav since more video was recorded than soundboard recording.

[rhinowing]

try Vegas Video....you'll want to fill the missing part of the soundboard with the camera audio as well, probably.

[rhinowing]

something like audacity or cool edit would work too

[AAR.oner]

Originally Posted by Brian_Barbour

ok so i have the original ac3 and m2v file that I got from the dvd vob file. I also have a single wav of the new audio. What do I use to listen to them side by side and sync them up? the ac3 is longer than the wav since more video was recorded than soundboard recording.

whatever program you are using for editing/authoring...most PC folks around here prefer Sony Vegas, but theres others as well

[KoolKat]

Originally Posted by saltman

The basics are....

use a program such as tmpgenc xpress (mpeg tools) to demultiplex the .vobs to their elementary streams. (m2v and wav or ac3 or whatever it is.)

That one is not needed & actually isn't that clever.

Here's a bit of nice basic for all to have fun with.

The seperate Tmpeg Mpeg Editor program is more frame accurate & precise and saves doing the demux operation.

Load the TS folder with source wizard,swap the audio,adjust the audio sync with your gap correction + or - input time ,then export lossless...........now author with whatever.
That'll save you a few hours of fooking about.

*NOTES:
Mpeg Editor will not accept AC3 ,however,"most" of what can be done in Mpeg Editor program can now be done in Tmpeg Author Pro at frame accurate level which will take AC3.

If you ever use Author Pro for editing frames out before final author and your source audio is AC3 i strongly recomend watching your finished ,burnt product before deleting files and wotnot.
AC3 is a funny bugger and it may wobble out of sync again from those edit positions.
(It's only a .05% chance but it has happened so be safe eh)

Don't forget audio for DVD is 48000 & not 44100 so upsample required if swapping(that answers most of your going out of sync after swapping questions)

Not a detailed walk through i know.But you can understand the writing on the tabs in the programs cant ya?

Have fun

K_K