PDA

View Full Version : How To: Ducking: Automatically Lower Music Bed When Speaking


Dailysonic_Adam
Mar 11th, 2005, 10:51 AM
I wanted to contribute something useful to the Podcast Alley, so here's a little audio technique that some of you may not know about, but is something very very useful in putting together your podcasts.

Its called ducking. If you set up your studio gear or software correctly, you can make music bed automatically lower its volume when you (or who ever is hosting the show) begins to speak.

I'll explain how to do this using multitrack software like Pro Tools or Logic, but if you understand the concept, you can set up ducking in almost any studio or software situation - even duck in real time.

1. set an audio track for your voice or whatever you want to trigger to ducking. we'll call this track one.

2. set up up another separate audio track for your music bed or background sound. we'll call this track two.

3. put a compressor on track two. (this an effect or plugin that's come with most multitracking software)

i guess i should quickly explain compression. a compressor reduces the dynamic range (loudness) of a signal. it ofter used to even out the levels of a musician's performance (ie. make the quiet parts sound louder, or the louder parts sound quieter).

as far as we're concerned, there are two important setting on a compressor... threshhold and ratio. threshold defines the volume level that the comp. kicks in. and ratio tells the compressor how much to reduce the volume by. if that still doesn't make sense, here's an article about it (http://www.harmony-central.com/Effects/Articles/Compression/)

4. set up a "side chain" (you usually set this up in the compressor's plugin settings window) so that the output of track ONE triggers the compressor on track two...

normally a compressor listens to *itself*, and kicks in when the volume level reaches its set threshold... but here, you're setting it up to listen to your voice.

5. play with the ratio and threshold levels, and track one and track two's volume, until you've got something you like. when you play with these setting, you're basically fine tuning everything so that:

a. when you're not speaking, the music plays at a good level
b. when you start speaking, you're at a good level, & the music volume lowers itself to a level that you're happy with.

that's it... i hope that makes sense.

BTW - while ducking is great, I don't always use it. I sometimes find that adjusting the music bed's volume manually gives me better results... buts its definately a very useful little trick.

- Adam

TexasMusicForge
Mar 11th, 2005, 12:34 PM
Good stuff from Adam and thanks for contributing to the community here.

Another way to do this if you're not running a "live" show and are using a multitrack audio program like Cakewalk or n-Track to piece your show together is to put all your music on one track and all your narration on a separate track and automate the volume levels for each.

For example, I record my shows as separate clips of a few paragraphs at a time of spoken word and the various bits of music, then lay each into their appropriate track when I assemble the show. I can direct the program to turn the volume up or down at specific points for the music beds and then move the clips for my spoken parts into the correct position on the adjoining track so it all lines up. It's pretty simple and adds a visual element to aid in the editing of your show.

Best regards to all,

Tio Ed
One of millions of guitar players in Austin this week

Inferno
Mar 11th, 2005, 12:57 PM
i tried ntrack and it screwed me!

after i saved the complete file about every 10 seconds it made a noise...
its prob becasue i used the trial version..i have idea.

but thanks i'll check out cake walk so i dont sound so low budget.....or is low budget in these days?

i guess we will find out.

shamigo
Mar 29th, 2005, 02:40 PM
How does one go about doing this in Audacity...come to think of it, it'd be a great idea if someone that is quite apt with Audacity could post some kind of basic's tutorial or something of the sort for us. Either that, or point us to a help page, because (and maybe this is just me) the help for Audacity isn't really comprehensive, and it doesn't really show you how to do detailed stuff.

Thanx to anyone who helps :D

cc_chapman
Mar 29th, 2005, 07:17 PM
I have not seen any way to automatically do this in Audacity. I always just put the music on different tracks and lower the threshold of them so they are not so much over the top.

shamigo
Mar 29th, 2005, 09:55 PM
actually, wot i really wanted to do was kind of like mix tracks. Wot i plan on doing is starting to speak when a song starts/ends, so i wanted to know how thats possible. If worst comes to worst, i'm willing to record my speaking, and then put it over the top of the end or begining of a song to make is sound like i'm cutting in or the song is cutting in...either way, it'd be awesome if someone would help me out. Thx a bunch!

vikingyouth
Mar 30th, 2005, 01:29 AM
some kind soul made a nyquist plugin for audacity to help a bit with this. while it's not automatic it does allow you to control the volume level and time to duck your tracks. here's the code for it, just save it as some name.ny (ie radiofade.ny) and put it into your plug-ins folder. here's the thread it was from. (http://thread.gmane.org/gmane.comp.audio.audacity.general/6775)

;nyquist plug-in

;version 2

;type process

;name "Radio Fade..."

;action "Fading audio to db level..."

;info "by David R. Sky\nReleased under terms of GNU Public License"



;control fo-start "Fade-out start time" real "seconds" 10.0 0.0 600.0

;control fo-dur "Fade-out duration" real "seconds" 1.0 0.0 10.0

;control db-level "Lower db level" real "db" -12.0 -24.0 0.0

;control choice "Keep at constant value, or fade in again" int "0=steady 1=fade back in" 0 0 1

;control db-dur "Low db level duration" real "seconds" 10.0 0.0 120.0

;control fi-dur "Fade-in duration" real "seconds" 1.0 0.0 10.0



; Radio Fade by David R. Sky, March 1, 2005

; Released under terms of the GNU Public License

; http://www.opensource.org/licenses/gpl-license.php




; determine selection duration

(setf dur (/ len *sound-srate*))



; setting pwl break point times

(setf point1 (/ fo-start dur))

(setf point2 (/ (+ fo-start fo-dur) dur))

(setf point3 (/ (+ fo-start fo-dur db-dur) dur))

(setf point4 (/ (+ fo-start fo-dur db-dur fi-dur) dur))



; converting db level to linear

(setf level (db-to-linear db-level))



(setf amp (cond

((= choice 0)

(pwl 0 1

point1 1

point2 level

1 level 1 1)) ; end pwl-0



((= choice 1)

(pwl 0 1

point1 1

point2 level

point3 level

point4 1 1 1

)) ; end pwl-1

)) ; end setq and cond





(mult amp s) ; end mult

Keeme
Mar 30th, 2005, 01:32 AM
I play around in Audacity... what I do is record myself (and the other people in my head). I then add all the tracks I want to play (afterward). If there is a tune I want to play and start my "piece" on I duplicate the tune I cut the first tune around 10 seconds into me talking and apply fade out. I then take the duplicate tune (which is in perfect sync with the first) and adjust its volume at the start. Almost the same thing with closing out... sometimes I just adjust the tracks and catch a good spot.... that or I drink lots of beer and get lucky.

/2cents

KEEME (http://www.keeme.com)

yaz
Mar 30th, 2005, 01:54 AM
when i used to surf, ducking was going under a wave instead of it hitting you in the face...

and keeme, whatever the hell you do, its funny...no need to get so tech...

cc_chapman
Mar 30th, 2005, 05:30 AM
This is my recipe in Audacity

1. Record all of my audio. I just pause to create a silent spot so I know where to put the songs.

2. Drop in the song tracks, station IDs and such each on a seperate track so that I can mess with the levels for each song.

3. Run the Compressor plug in on my vocal track to bump it up appropriately

4. Export as a .wav

5. Encode it as an MP3 with Razor Lame

6. Upload to Libsyn

shamigo
Mar 30th, 2005, 01:59 PM
lol...a little technical for a n00b :oops: , but i'm gonna try and play around. Thanx guys, i'll ask more questions if i need to =)