Syncing spearate audio/video: Better to stretch/shrink video than audio?

[stantheman1976]

Someone was asking on another board about syncing up a music DVD he got with a better audio source. He was given the advice by a main member of that board to stretch/shrink the video instead of audio to make them match. The OP is using Vegas like myself. I've always adjusted the audio track when using separate sources and never had any problems. Opinions?

[manono]

I'd say the guy got good advice. A video's framerate can be adjusted without reencoding. To stretch the audio you have to reencode. Of course, the type of video might make a difference. You didn't say. And you don't need Vegas to do the job.

[stantheman1976]

Well we both happen to be using Vegas and I've always found it to be excellent when syncing separate sources. I've shrunk audio a few seconds on a 1 1/2 hour project before and didn't notice any artifacts or irregularities.

[manono]

OK, so why ask a question to which you think you already know the answer? You asked for opinions. I gave you mine. It's worth about what you paid for it.

[gadgetguy]

I think it depends on the final destination. If you're going to a format with a strict video framerate requirement and the video is already at that requirement, then it makes sense to adjust the audio. If going to a format that is more flexible, then it makes sense to avoid re-encoding both the video and audio by adjusting the video framerate.

[stantheman1976]

OK, so why ask a question to which you think you already know the answer? You asked for opinions. I gave you mine. It's worth about what you paid for it.

You said Vegas was not required for it. All I did was state that we both happened to be using Vegas. Other software is not required when you have something that can intelligently accomplish what you need to accomplish. I was just stating my opinions also.

[AlanHK]

To stretch the audio you have to reencode.

No you don't. Assuming you have AC3 audio, you can use,
AC3scale (http://thunder.eeap.cwru.edu/~zhuo/audioscale.html) to stretch or compress it.
This cuts or duplicates audio frames to make it fit the target length. Very useful for small adjustments.
To cut or add silence, use Delaycut.

[NukeyDoo]

Hi all!

I'm new to all of this, including Vegas itself. I have a separate mp3 with a song, and a video of a concert where the bang is playing the same song. Just as a goof, (and to teach myself all of this) I am trying to sync up the singer of the mp3 with the singer in the video so it looks as if he's singing it.

Does anyone have any newbi-ish links to show me how to stretch and shrink the audio and how to make it fit the singer in the video exactly? From all of my reading so far, this all seems really complicated.

What would be REALLY nice (don't know if this tool exists) is one like one of those picture morphing programs, where you simply insert a marker in one spot on one picture, and then a matching marker on the other picture, over and over until you're done, and then the program does all the tweening. i.e. you put marker A on the tip of a nose in pic A, and then marker A in the tip of the nose in picture B, and so on.

In this case, I could put marker at a spot in the mp3, and then a matching marker on the frame I want it to sync to in the video, and then when all done making markers, the program automatically stretches/shrinks what is needed to make it all synch up. Is there anything like this on the market yet?

[manono]

Assuming you have AC3 audio, you can use, AC3scale (http://thunder.eeap.cwru.edu/~zhuo/audioscale.html) to stretch or compress it.This cuts or duplicates audio frames to make it fit the target length. Very useful for small adjustments.To cut or add silence, use Delaycut.

Gee, I never heard of that one, Alan. Looks handy for some. A GUI would help for the DOS challenged, like me. Of course DelayCut is the standard when you have a constant asynch. Either that or adjusting the delay. But in this case I think he's talking about progressive asynch. Otherwise why would he even be talking about stretching the audio or video?

And when I was talking about adjusting the video framerate, I was thinking of VDub for AVI or manipulating MPEG-2 video's length through the use of DGPulldown. Not only do you not have to use Vegas for these operations, but it's not even possible to perform these operations using Vegas, I don't think.

I've shrunk audio a few seconds on a 1 1/2 hour project before and didn't notice any artifacts or irregularities.

Surely you're not suggesting that reencoding doesn't result in quality loss?

Other software is not required when you have something that can intelligently accomplish what you need to accomplish

It is when the software you're using isn't designed for, nor can accomplish, the goal in the best manner possible, the goal here being synching up the audio and video. But since we still don't know the kind of audio or video (or the video framerate), this discussion is pretty much just academic.

[NukeyDoo]

Anyone? (regarding my question 2 posts above)

[DB83]

Anyone? (regarding my question 2 posts above)

What you want to do is incredibly difficult if not impossible, if I read you correct.

You have a mp3 of a recording and want to match it with a live performance. If the recording is from that live performance then it could be achieved by varying the length of the video to match the audio (one a small amount tho) - that is what is being discussed here.

But a studio recording will have totally different timings to the live performance. The song could even have minor variations in music played. Synching will be impossible IMHO

[NukeyDoo]

Hi

Yep, that is pretty much what I want to do except it's a live recording (mp3) of the original band playing a song, that I want to match up with a live video concert of a different band playing the same song.

The video is muted so you only hear the real band playing it, so I just want to sync the audio up with the video. The band that covered the song, didn't do anything to the arrangement. It was done exactly as the real band does it, so the timings aren't really off much, but I just don't understand how to stretch and sync these things up. I'm new to all of this.

p.s. I don't mind if things don't match 100% perfectly. the video was shot kind of far away and badly, and this is just practice for me anyway.

Thanks for any information

[Jensen]

If i could make a suggestion, I have been working with something similar. If you take the live footage and cut it in a couple of places and replaces the cuts with lets say footage from the audience shots, you could then theoretically sync the rest of the footage to where it fits. The best cut would be where the live footage gets off place or out of sync, then go to where its back on sync or move it. A duplicate shot of the audience might not be so obvious if done right. Or maybe a different audience shot from a different video perhaps with the same quality, Just a suggestion

Jensen

[AlanHK]

Yep, that is pretty much what I want to do except it's a live recording (mp3) of the original band playing a song, that I want to match up with a live video concert of a different band playing the same song.

The video is muted so you only hear the real band playing it, so I just want to sync the audio up with the video.

You can do this in the sound editor, Audacity.
See http://audacityteam.org for help docs.

First demux your original soundtrack. Load it in Audacity. Then import the new soundtrack.

You can play both tracks together, or mute one.
Find a distinctive moment of sound near the beginning on the original track. Mark it, then find the corresponding point in the new track. Move the new track to align.

If the timing is exactly the same, you're done. But go near the end and see if they're still in sync. If not, stretch or compress the new track to fit.

When the timing is right, delete the original track and export your new one as a new stereo WAVE file and convert it to whatever format you need for the video.

Or you might keep the introductory section of the original and delete the music part, then mix them.
It's pretty easy in Audacity once you work out what the tools do.

There is a tutorial on multi-track mixing at
http://quicktoots.linuxaudio.org/toots/audacity/
or http://www.transom.org/tools/editing_mixing/200404.audacity.html
that covers the details.

[guns1inger]

Just note that audacity doesn't pitch shift in the same pass. If you find that stretching or compressing the audio changes the tone, you may have to run a pitch shift afterwards to get it sounding correct again.

[AlanHK]

Just note that audacity doesn't pitch shift in the same pass.

Actually, you can use the "change tempo" effect to change duration without changing pitch.

But I assumed that any stretching or compression would be by a few percent at most, if so there should be no perceptible pitch change.