Splitting/Editing Multi-Track Audio

[SweetWilliamNow]

I have a DVD with 2 audio tracks, English and a French. In the middle of the DVD the English track goes out of sync for about 15 minutes. The beginning and end of the movie are fine, it's just the middle section that is off. I would like to edit the audio track, preferably lossless. I demuxed and have an ac3 file. I also downloaded Audacity, and opened the file in there, but couldn't figure out how to only edit the English track. Is what I'm trying to do possible? Any help or guides would be appreciated.

Also, I know the quality will take a hit once I convert, but is it possible to split an ac3 file with 2 different audio tracks into 2 mp3/wave files (one file for the English track, one for the French)? I have a good (well, I'm happy with it) mp3/wave editor and I think I'll have an easier time editing in this. Again any help/suggestions would be appreciated. Thanks.

[Baldrick]

How did you demultiplex? You should be able to get both ac3 tracks. Try pgcdemux or dvd decrypter (click on them for guides).

[Cornucopia]

Does that DVD have 2 separate STREAMS of audio (1 for English, 1 for French) or does it have 1 stream composed of 2 channels (English on L, French on R, for example)?

Your 1st sentence leads me to think the 1st option, but a later sentence leads me to think that the 2nd option is what's really going on.

AFA quality: ANY modification you make to one of those audio streams will require re-encoding/converting, and re-encoding entails loss of quality (EVERY TIME).

Yes, it's possible to split the tracks, but how it is done depends on which of those 2 options is actually occurring.

Scott

[SweetWilliamNow]

Thanks for the replies! I ripped it with DVDFab and demuxed with pgcdemux. I looked over the original pre-muxed file and there was only the English track on that (?). I re-ripped it and got the 2 audio tracks on the second rip (first rip was "main movie" only, second was "full disc"), demuxed and ended up with the 2 ac3s. Putting them side by side in my audio mixer shows they are out of sync everywhere! And when you line them up, 30 seconds down the line they are out of sync again??? I downloaded a English dialog VHS rip and now have 3 audio tracks that don't sync up at all! About ready to throw in the towel.

So what you're saying, Scott, is any audio modification with any editor/software will leave you with a quality loss?

[poisondeathray]

So what you're saying, Scott, is any audio modification with any editor/software will leave you with a quality loss?

When you're working in an audio editor, you're working in the uncompressed domain (the AC3 audio is decompressed to wave)

If you left it uncompressed wave, then there would be no quality loss

When you recompress with a lossy codec like AC3, you will get quality loss

LPCM is compatible on DVD, but you would have to redo the DVD (it's going to be much larger than the old AC3 stream)

[Evil_Burrito]

By multi-track you mean multi-lingual, correct? The dvd has 2 stereo tracks?

To re-sync using audacity, create a splits near where the out of sync begins and ends in a spot where the volume is almost silent. Then use the time shift tool to adjust the new slice. Make sure to save an audacity project so you don't lose your progress. Before exporting, to maintain the new edits, you must insert silence in blank areas. To do that, highlight a blank space, select generate then click silence. Remember NOT to merge/join/flatten your new regions, so you can come back to them; you can also create a text track to help make markers on the time line.

Since audacity is an audio without video editor it can be a little difficult to fix using this method, but it is still very possible.

If you are worried about quality loss set your project at the native frequency (typically 48k) and save it in .wav (lossless) format before you begin editing.

[Cornucopia]

I would go back to square-one and use AnyDVD to decrypt, and DVDDecrypter (if it will work, should unless the title is new & popular & using Sector scrambling crap) and rip to elementary streams (MPEG2 video and AC3 or LPCM audio). Load the audio streams into a multi-track audio editor and, using the French track as your guide (since as you say it is still in sync with the video), match the timing of the English to the French.
However, if this is an English-native movie, you would really want to not just use an audio app but rather a video nle app so you can check the lip sync.

Doing it that way has always been much less prone to glitches & skips for me. Maybe it's more steps for you than you want to do, but I'd rather spend a little bit of extra time up front to get it done right than have to spend a bunch of time later on having to fix more things...

Scott

edit: BTW, quality loss in these instances is almost unavoidable, but with careful choice of high-enough bitrate and good tools, you can minimize the loss.

[SweetWilliamNow]

Thanks again for help, folks! I had the same plan, Scott, side by side wavs in a audio editor syncing to the French track (DVD is unsubbed French with a 2nd English audio track, and French is OK). I don't know what I'm doing yet with Audacity, and it's easier to compare the 2 files side by side in the audio editor. It's been a royal pain in the A, though (most all of the last 40 minutes are out of sync on the English track, with some of the same music cues being different durations). I'm happy to say, I'm just about done with it and it's coming out nice.

One last problem: I had to put a piece from the French track into the English track (music theme) I'm re-syncing and the French track is lower in volume than the English track. Is there a program I can brighten the sound (French track is dull sounding compared to the more crisp English track) and raise the volume of a wave/mp3 without it being in the red? Should I start a new thread for that question?

[Cornucopia]

Well, you'll never get it to match exactly, but what you can do is look at the steady-state eq spectrum of both tracks (during similar music-only portions) and visually match the curve of one to the curve of the other. Also take note of dynamic range. If you have to, you can can compressed (I'm talking dynamic range compression here, not bitrate reduction) the levels so that what used to be a 60dB spread becomes a 40dB spread (or similar - I'm just using an example) and then you'd be able to bump up the overall level of that reduced spread by 10-20dB, which is noticeable, without even going above the original peak. Of course, this MUST be done in MODERATION (or it will sound bad).

Scott

edit: there are tools that can do similar things automatically, but they are $$$!

[SweetWilliamNow]

I'm not sure I understand correctly. It sounds like your telling me to compress 20 dB then raise the level 20 dB. Wouldn't that leave it the same?

[Evil_Burrito]

One simple way to fix your new problem in audacity is, select your audio and use the amplify effect. A better way (takes a little more effort) is to use the envelope tool. That way, you can visually match the waveforms.

Read this document to understand how to use envelope tool.
http://manual.audacityteam.org/man/Envelope_Tool

And no there is not a magic button to make audio brighter. Sony does have an "enhance" effect, but I assume you don't have acid pro or soundforge. And I don't recommend that effect anyways. There are a million ways to to achieve a "brighter" sound, the simplest way would be to use an equalizer and raise the frequency.

[Cornucopia]

I'm not sure I understand correctly. It sounds like your telling me to compress 20 dB then raise the level 20 dB. Wouldn't that leave it the same?

Let's say you compress the dynamic range of a signal by 20dB. That means the loudest part of the signal and the quietest part, which might have had a range of -60 to -0dB now has a range of -50 to -10. All signal strengths within that range are proportionally adjusted to maintain a similar relationship to each other. Then you raise by 10dB to get it to -40 to -0. Or you could just as easily say you went from -60 to -0 to a range of -60 to -20, then raised by 20dB to get a range of -40 to -0. Same thing, but depends on how the compressor acts on the signal (actually most work this 2nd way).

Btw, @evilburrito, eq does not raise the frequency (that would be pitch shifting), it raises the level/volume of a range of frequencies. That MIGHT help, but would probably be incomplete.

Edit: and the volume envelope tool, while also handy, won't totally match the envelope of one signal to an envelope of another either. For a more exacting effect, you would want to use "convolution". Even then, you only want SOME characteristics to be similar.


Scott

[Evil_Burrito]

Yea I know the difference between equalizing and pitch shifting. Cornucopia, I think you misunderstood me, but that doesn't matter, I am not the one who needs help. I am trying to keep things simple enough for SweetWilliamNow to understand because obviously he is an absolute beginner audio engineer. So I don't think it is helpful trying to explain compression or convolution. Also I don't think he has access to any friendly gui compressors, if he is only using audacity.

[Cornucopia]

I figured you knew, but I wanted it to be clear to the OP, being as green as he seems. Certainly don't want to start them off on the wrong foot with a wild goose chase. And I agree we should NOT go into convolution, but it doesn't hurt to mention that as an advanced possibility. That's why I didn't go any further with it.
But certainly compression is or ought to be a part of every audio engineer's basic toolkit. And Audacity does support VST plugins, which includes a number of fairly decent GUI compression tools.

Scott

[Evil_Burrito]

True.

SweetWilliamNow, you need to learn about using compressors (not lossy formats like .mp3) if you plan to do any more projects that involve editing audio. It can be a lot to learn, but it is crucial knowledge.