Automatic ducking of Voice overs not music??

i am erikd wrote on 1/20/2018, 11:09 AM

I have quite a lot of content similar to above that I am looking for a short-cut in audio editing. You can see two channels of mono audio, each with their own voice speaking from a two person interview. The human eye can easily detect in the waveforms where the person begins speaking and when they are quiet. I want an automatic way to aggressively reduce the audio level envelope effectively to zero when each person is not speaking. This would be faster than me manually editing each portion. I realize that there may be sections that I have to go back and adjust the audio level envelope especially if there is an overlap of audio as they speak on top of each other. I have lots of third party apps such as Vegasaur, VASST Ultimate S, etc. but they all seem to be focused on ducking the music for VO. I want to do aggressive automatic ducking for voice over only. Any ideas on how to achieve this??



Former user wrote on 1/20/2018, 11:18 AM

When you say zero, are you talking 0dB on the meters or are you talking silence (which is minus infinity, not zero). If you are talking 0dB then you can use a compressor with a hard limiter set for the level you want.

i am erikd wrote on 1/20/2018, 11:23 AM

I was hoping to get something very close to silence actually.

Marco. wrote on 1/20/2018, 11:26 AM

I think the silence detection offered by Vegasaur should be close to your need.

i am erikd wrote on 1/20/2018, 11:28 AM

Hi Marco,

I played with the silence detection trying several different level settings but it wouldn't do what I want. It kept selecting the wrong areas for "silence".

ryclark wrote on 1/20/2018, 11:42 AM

Can't you use the Track Noisegate which is intended to turn off the audio below a certain threshold?

i am erikd wrote on 1/20/2018, 12:02 PM

Tried the noisegate which comes close but again by playing with the attack, release and threshold I can get in the neighborhood of what I want but it isn't good enough to rely on. Looks like manual editing might be the only way. I just wanted to see if I could up my game a little bit.

Pashi wrote on 1/20/2018, 2:56 PM

Compressor with sidechain usually used for this feature. You can watch this tutorial.

It's in russian but everything is simple and understandable without voice.

rraud wrote on 1/20/2018, 6:20 PM

The 'db Audioware' sidechain comp is one of the very few that work at all with Vegas. Though it's more work, I prefer 'manual' ducking with envelopes which is more precise, either on the track(s) or sub-master.

FWIW, I've been requesting side-chaining capabilities since the Sonic Foundry days... or at least I used to, since it always fell on deaf ears.

ryclark wrote on 1/21/2018, 7:59 AM

FWIW, I've been requesting side-chaining capabilities since the Sonic Foundry days... or at least I used to, since it always fell on deaf ears.

Magix obviously have Auto Ducking turned on. 😉

Jillian_A wrote on 1/21/2018, 4:54 PM

Here is a quick idea, perhaps you have already tried this.

You now have two audio tracks, A & B. Copy each to a new audio track a & b.

Now use VAAST to duck A based on b. This will duck A everywhere b is speaking. Next, duck B based on a. This will duck B everywhere a is speaking. Finally, delete a & b and touch up the ducked areas on A & B.

I have not tried this, but it is much easier to touch up already ducked tracks than it is to go through and create hundreds of ducking areas.

Hope this helps.

Former user wrote on 5/19/2018, 3:32 AM

So current vegas 15 build still doesn't have auto ducking built in. I have to pay for a plugin>?

jetdv wrote on 5/23/2018, 7:07 AM

If you just delete the "silent" sections, you can use the Voice Over command in Excalibur to do that.