Auto-ducking of voice tracks

[Tranquil_Cosmos157F]

Problem

More and more users are editing podcasts, Zoom/Teams events, video interviews done with tools like Riverside etc. in Premiere Pro. Especially after we got Text-based editing, I’ve seen an increase in clients doing this kind of editing in Premiere Pro.

One thing they all have in common is that they’d like to auto-mix voices that are on different audio tracks. Say you have two or three audio tracks, and you want to duck the levels on the tracks of those who are not speaking at the moment to avoid spill/bleed/leakage/phase issues, noise, and generally low audio quality. For now, this is mostly a manual task.

Examples

This is a screenshot of a client’s timeline of an interview with two microphones (track 1 and 2), and as you can see, they’ve done a lot of manual cutting back and forth between the tracks.

Manual cutting of interview with two microphones

I have tried to trick the Essential Sound panel into auto-ducking one voice against another, calling one of them a music track, but it’s impossible to find a setting that makes it work 100% of the time.

Tricking Auto-Ducking into ducking one microphone against the other

FR:

So my feature request is: I want to select all the voice clips, and make the AI duck them an amount of dB, very much like the auto-ducking we have for music.

We need to be able to choose one voice as the winning voice if/when multiple voices compete for our attention.

[Colin G Smith]

Great idea and you're right that editing multiple dialog tracks is becoming more frequent. 

[R Neil Haugen]

Oh YES! YES! ... and yes, I'm shouting yes!

This would be so freakingly incredibly useful, and guaranteed, there's a LOT of us that would be using it daily!

And ... two of my fav experts/teachers in one thread! Jarle I've met and worked with, Colin ... seen so many times on screen. Awesomeness abounding!

[Christian.Z]

Please submit a Bug report / Feature request at the Adobe User Voice: https://adobe-video.uservoice.com/

[Tranquil_Cosmos157F]

Uservoice? Really? Which year is it? 

[Richard van den Boogaard]

I would love to see this implemented, but I would also like to see it expanded with an option for frequency ducking, where you choose to only attack the frequencies that fight with voice in the music track and leave the lower frequencies in tact.

This creates a much fuller sound in the edit and is akin to an "outside the club" effect or a trick that is often used in deep and trance house music, just before 'the drop'.

Side chaining is not an option in PPro, so enabling this - with one or multiple voice tracks - would be greatly appreciated.

[Tanuki M]

Yes. indeed, that integration would be also crucial when dealing with translation voiceover recordings, when clip that is in "original" language and is desired to be ducked against translation one. For now only two ways - doing it by hand with volume keyframes, or assign "Dialogue". match lodness, remember gain step, then reasign as "Ambience" and then do the ducking. - quite time consuming to be told. And second option can really surpass time spend, against doing it manually sometimes.

[markschlicher]

Yes, critical for interviews

[Trevor_Asq]

WT Automixer is a Duggan style auto mixer plugin that can achieve this a bit (although dipping down to silence is probably going to be too breathy for this or auto ducking).

I think Beatrig has a side chain compressor working in PP. Which would work for e.g.  compressing native voice when translation is occuring.

Avid Media Composer has 'strip silence' which can be useful in this kind of scenario - but again it still results in removing too much (i.e. little laughs / reactions that are lower level) or too little.

What would make editing this kind of dialogue way quicker would be 'range select' or 'region select' like FCP7 had, FCPx has (I think) and Audition, Reaper etc DAWs have.

Here's how Resolve and Audition can be edited whilst playing back.