Ducking and Sidekick 3 (VST Plugin)

[In 1.5]

On the Adobe forum, Emmett posted a link to a tutorial about Ducking, which uses a very nifty free* VST plugin called Sidekick 3. It was at that stage untried, and there were no references to Audition on the site, but since this has been a bit of a thorn in the past, I thought I'd have a go at it.

What I shouldn't have done was read the tutorial - it's basically a bum steer*. Okay, doing it the suggested way works, but it uses busses quite unnecessarily. It was a particularly bum steer when it comes to Audition 1.5, because it causes C++ crashes... but after some experimentation with both 1.5 and 2.0, I have a methodology that works with both. It's not particularly easy to use with 1.5, because VST effects don't display the way their makers intended, which is a pain. But hey, the fact that it works at all is quite something - this is miles better than attempting to use the envelope generator, especially in 2.0 where you get a compression meter. This is a more-or-less duplicate of what's in the Adobe thread, incidentally, for those who don't go there.

In 1.5:

The voice track has to be fed directly to the output, but with a sidekick send on it, and on this copy of sidekick, all the control sliders are set to the left. You can't bypass it, but the Sidekick fader in the Audition mixer should be set to zero - it's merely providing a virtual channel send.

The music bed also has sidekick as a VST, but this time you have to select it as an input, and mute the direct track feed in the output. This is important, or you get no ducking! For some strange reason it seems to take the Gamma input, rather than the alpha one, (YMMV) and I set that to 1.000 - it's only an enabler after all. Leave all the 'disabled' sliders set to disabled, and don't touch the bypass, whatever you do - this will make the output latch up. Duck/key needs to be set to 1.00 - right over to the right, and by juggling the attack, release and threshold sliders you can achieve ducking on the music bed.

It's a bit fierce to set up, but it does work after a fashion, and produces a better result than the envelope generator does. You have to persevere with the settings though.

In 2.0:

Okay, as I suspected, you don't need to use busses to get this to work in AA2.0 - it works fine within channels in exactly the same way it does in 1.5 (which is what I suspected, and partly why I wanted to test it there first).

It's a whole lot easier to set up in 2.0 because you get a compression meter display that you don't get in 1.5, and that really does make a difference to the ease of use.

All you need to do is to get an instance of Sidekick into both tracks and set the dry-wet mix according to what you want to duck or hear. So that's wet for the music bed and dry for the voice channel. You need an input (which appears to be the beta channel this time  ) for the music bed, and again, this is where you have to set Sidekick up - it will pick up the control signal automatically from the voice track, which gets sent normally to the output.

In 3.0:

Works just the same as it does in 2.0. I should perhaps mention that which control channel you need to select depends on which way around you have your tracks arranged. If I put the voice on track one, and the music on track two, then it's the alpha channel you have to select on track two, and nothing on track one (because it's providing the control signal). Then set up for tasteful ducking, and mix down the result.

I have no idea why the tutorial thinks that you need busses for this in 2.0, because you simply don't - it works fine in the channel as it is, and does exactly what's required - sends a control signal from one to the other. Yes, it's pretty good, and when you've set it up correctly, does exactly what it says on the box.

It seems to work quite well with a really fast attack time, and a release time set to about 370ms - set the threshold to taste (about -34dB with the voice track I used)

The guy that made this VST deserves a coconut - this is pretty good.

* Have subsequently heard from the writer of the tutorial, and I think he's going to modify it somewhat, in line with the implementation  I've suggested.

*Sidekick 4 isn't free any more, but it's not expensive.

Dumb question here, but didn't I read somewhere that Audition 2.0 does this already without the pluggin? 

While I cannot find it now, I swear that I've seen a "tutorial" on the net somewhere that explained how to do this by putting a compressor on the background track, and sidechaining the input of the compressor to the vocal track.  It also mentioned something about is sending the dry side of a delay to activate the compressor while sending the wet side to the mix, so that the background track "ducks" before the vocal starts.

I probably have no idea what I am talking about here, but was just curious.

I probably have no idea what I am talking about here...

Well, I agree with that bit! Audition has been notorious for not doing this just about since it came out - an amazing omission considering how much it's used for radio production. There is an envelope follower, but it really doesn't work properly as a ducker at all, and it's a right pain to set up. If you read anything at all, it certainly wasn't about any built-in effect within Audition, because nothing in there except the envelope follower and vocoder accepts inputs from more than one track.

If you can find a plugin compressor that takes a sidechain input, and Audition can provide it, then yes, that would work too, I'd think. Do you know of one?

I am very new to Audition...  I still don't know my way around it, I just remember seeing a post somewhere on the internet that explained how to do it in 2.0.  Unfortunately, I just cant find it again. (...and can't guarantee that the author of that post knew what they were talking about either.)

Can someone please explain clearly in non-technical jargon what this thing does please.

In a nutshell, "ducking" will lower the volume of one track whenever something is present on another.  A common example would be a music bed on track 1, and a (spoken) vocal on track 2.  This pluggin would lower the volume of the music during the spoken vocals (so that you can better hear the voice) and then bring it back to normal volume when there is no vocal. 

This is done quite often in commercial production.  (I think it is also done in music production to bring out certain elements like a bass drum)

Can someone please explain clearly in non-technical jargon what this thing does please.

Ducking is a non-technical term. There's even a Wikipedia reference to it - here.

don't spend too much time in the Radio, V/O, BBC forum, but w/o knowing for sure the 'sidekick' was a plugin I grabbed the thread precisely because ducking has been a hole in CE/AA's repetoire

I've had work arounds and other plugs & software but several areas where I prefer to use AA ths wil help

AND w/o knowing that the thread would be useful it provided exactly the type of info I was expecting

(probably don't need a vote of everybody who logged in . . . that would be a bit petty after all, but precisely because I read thread because of the title I thought I'd chime in . . .  it's Sat I've a benefit brunch to attend . .. don't want to engage pithier thought)

[the 'sidekick' was a plugin]

don't spend too much time in the Radio, V/O, BBC forum, but w/o knowing for sure the 'sidekick' was a plugin ...

Fair point - I've added that to the thread title, even though it's in the first line.

Oh, I had no problem with original title . . . it's what caused me to drop into the radio forum and resulting info (thanks by the way appreciated it and will find it useful . . . when turning around a live performance into a marketable product before the show ends it helps to stay with as few applications as possible and for obvious reasons I'd like to keep the pst tracking editing in AA, & for this type of thing side chain control is quite useful . .. I typically found I'd either set up parallel sessions in two editing programs or try to use a hardware compressor, anyway from initial experiments this looks like it will be pretty helpful) matched title to a 't'

headline writing is among my many non skills so I tend to not quibble excessively . . . 'course when I had responsibility for editing such things I did encourage use of a verb (which was not all that common in MSM when i started, and I would expect Brit implementation to be different anyway)

http://reaper.fm/reaplugs/ includes their compressor with sidechaining, but I'm not clear how that would work in the context of Audition and/or Sidekick.  I guess the answer is I should try.  It would be handy if Audition had channel functions built in.  That enables you to do stuff like having a delay on the voice channel relative to the output to the music bed channel on which a ducking compressor is working, so that you can set the music bed level to drop a tiny bit before you hear the voice.

Thanks for that info Steve I've often thought about trying that Sidekick but was dubious about its effectiveness

Actually if you add a delay after Sidekick on the voice channel, it does delay the voice but doesn't delay the send to the music bed.  So the music bed level can be set to drop just before the voice is heard.

I haven't tried the Reaper sidechaining compressor in Audition as without a sidechain input I don't see how it would work - I don't think there's a way of sending it the necessary input from the other track.  Of course it can be used as a normal compressor in any event.  That's the cunning thing about Sidekick - it combines the two functions of compressor and sidechaining in one effect.  But that does mean the the compression is no better than Sidechain provides.

Actually if you add a delay after Sidekick on the voice channel, it does delay the voice but doesn't delay the send to the music bed.  So the music bed level can be set to drop just before the voice is heard.

It's quite interesting to do it like that - you have to extend the attack time of the ducking quite a bit to get an acceptable fade on the music, and the delay time you choose is quite critical too (I used about 100ms) - but it works fine.