How do I dub?

[Mike Kujbida]

Welcome back Norman 🙂
You may get other suggestions but here’s two ways I would do it.

One way, and probably the easiest, is to re-record the entire sentence and drop it on top of the problem sentence.
Hopefully this was done in a controlled environment that you have access to.

The other way is to use the word you say you already have.
As long as the word stands by itself, this can be done fairly easily.
By that, I mean that it’s pronounced the same way and said with the same emphasis.
Standing by itself means that the previous word doesn’t slur (run into) into it nor does it slur into the next word.
Slur probably isn’t the correct term but hopefully you know what I mean by that.
Ungroup the audio from the video on the desired clip (the “elsewhere” part).
Now split the audio so the word is by itself.
Copy it and paste it over top of the mispronounced word.
Go back to where you did the split and group all the events back together again so that they move as one when you need them to.

[Norman Willis]

Hi Mike.

>>Ungroup the audio from the video on the desired clip (the “elsewhere” part).

Would I make a duplicate of the audio track, and then delete all but the word I need?

And then how do I ‘paste’ it over the top of the offending word?

[Mike Kujbida]

Norman, there’s no need to make a duplicate of the audio track unless you feel more comfortable doing that.

Once you get the replacement audio clip, simply drop it on top of the offending portion.
Doing it this way replaces what is underneath.

[D. Eric franks]

Mike’s gotcha covered (as usual!) and I’d only add that I use the waveform to visually line up the syllables as closely as possible, using two tracks (and zoooooooomed way in). Rarely, you might need to do a little time stretching/compression to get it to match, but if the waveforms look the same, it’s almost guaranteed that the lipsync will be perfect.

If it’s just a word, it’ll be really easy. If it’s one word and a common one, you should have no problem at all snip-snip paste. Sometimes it’ll be painfully easy, sometimes you might have to hunt for the right pronunciation/inflection and sometimes you’ll need to re-record.

[Dave Haynie]

I have done stuff like this pretty often.

First thing.. find your offending word, and cut the the audio portion out of the main A/V clip at that point. Slide it down to a separate track just to keep it around. Do the cut as close to the edges of the word as possible. If the background is quiet, this is easy to visualize… if there’s too much background noise, you may have to listen carefully to find the cut points.

Next find the replacement word. For this, try to find one as isolated from other speech as possible, and make the cut correspondingly as wide as possible… ok, you don’t have to cut, but it may be easier. Copy that isolated section to your problem area.

Next, try to line up the word as much as possible, using the original word as a guide. No big deal if it’s a little off, just match the leading bits, and most viewers will never notice.

Finally, blend in the lead and tail sections to the original audio… this is why you want the largest dead-air around that replacement word (assuming you have several to choose from) as possible. Blending from crowd noise to crowd noise is idea… no one will hear the edit. If you do a direct cut, there’s probably a glitch there, even if it’s a small one. Move the edges of the replacement clip that overlap (and crossfade) with the original audio to the cleanest sounding replacement.

-Dave

[Bob Peterson]

You probably want to replace the word rather than mix two words together. That would be an overwrite in SF assuming, of course, that both words have the same length.

[Norman Willis]

Hi guys. Thank you for all the excellent suggestions. I’m kind of slow, and I am still getting comfortable with Vegas, and I while I have Sound Forge, I have never really used it, so is it possible to do this whole thing inside Vegas?

Before I start editing, I normally use Sony Noise Reduction 2.0, dropping about 12db (“Default for fast computers”) out of the whole track. Then I bump the noise gate up until the static drops out. I apply that to the whole timeline before I do any cutting at all.

If I have the right replacement word already spoken in the sound track, then how would it be if I duplicate the sound track, find the right replacement word on the upper track, and cut away everything but the good word and the background sound on either side of it?

Then I could match the good word up to the lower track, and then zoom way in, insert a volume envelope on each track, and create kind of a ‘cross-fade’ in the space both before and after the words in question, so that the good word gets dubbed in, and the bad word gets muted out?

Would that work? Or would the Sound Forge thing be way better?

Thanks.

Norman Willis
http://www.nazareneisrael.org

[Bob Peterson]

Yes, Vegas can handle this quite easily by doing what you describe. SF might be needed if you wanted to do something more sophisticated. One big advantage of using Vegas, is that the editing applied to your soundtrack is not destructive. It can always be reversed to get the original sound. Being able to reverse a change is, IMHO, a good thing.