. Sidechains

  1. What is a sidechain?
  2. What can I use a sidechain for?
  3. Which Logic plugins offer a sidechain input?
  4. Which sound-sources can be used as sidechain inputs?
  5. How do you use Auxes as input to a sidechain?
  6. How do you use Virtual Instruments as input to a sidechain?
  7. What is "ducking", "voice-over", "gating", "de-essing"?
  8. How do you use VST effects that are capable of being controlled by MIDI data?
  9. How do you route audio through the Emagic ES 1 synthesizer?
  10. The Emagic EVOC 20 and sidechains
  11. How do you mute the sidechain input signal?
  12. How do you control the amount of influence the sidechain has?
  13. Can a sidechain be controlled through MIDI instead of audio?
  14. Using sidechains to automate Auto Talk Back
  15. I use an Aux as sidechain, muted its volume, but still hear the Aux signal

.1 What is a sidechain?

Usually you have some audio stream running through an effect, and determine "by hand" the amount of effect that is applied: you tweak the knobs until it sounds just right. Sometimes however you'd like a second audio stream to determine the amount of effect that is applied to the first audio stream. This second audio stream is called a sidechain, and the entire process is often referred to as "sidechaining".

Obviously an effect plugin explicitly needs to be capable of taking in a second audio stream as a "sidechain input". Sidechaining is thus not possible with all effects, but only with ones designed to allow it.

[top] [contents]

 

.2 What can I use a sidechain for?

Sidechaining can be used in various situations, and if you get creative, the sky is the limit. There are however some "standard uses" for sidechains, which will be discussed here. They should at least give you an idea of what sidechains are all about.

 

1. Ducking

Suppose you have a bass and a drum track (both Audio, and not MIDI). However, it seems that the kick drum sound gets "blurred" because of the bass: both have to coexist in approximately the same frequency range, and the presence of the bass tends to overwhelm the kick drum. It would thus be nice if the bass was a bit softer each time the kick drum hits. The solution is known as "ducking" (the bass "ducks" each time the kick hits).

On the bass track, insert a Compressor. However, instead of letting the bass sound itself trigger the Compressor (as would normally be the case), you now pick the kick drum track from the Side Chain popup menu in the compressor. (Note that the names of the Audio Objects will not show up in the popup; you have to pick the right object by looking at the Object's parameter pane or on its channel strip, right below its Output popup). Additionally (important) you have to switch off "Auto Gain" in the Compressor. What you have now is a bass track that only gets compressed when the kick drum is active, thus creating a kind of "pocket" for the kick to sit inside. As for the Auto Gain: when this is on, Logic will compensate the attenuation that takes place in the Compressor, by amplifying the output signal. In this particular case you explicitly do not want that to happen: the entire idea is centred around getting the bass to play softer. Auto Gain would destroy that.

  • The picture shows the setup described above. Notice that the kick is currently playing (level meter) and the Compressor's "Gain Reduction" meter indeed shows the Compressor to be activated.

 

2. Voice-over

You may have noticed that in e.g. documentaries, as soon as the spoken comments begin, the background music gets softer. This is achieved in the same way as described above, under "ducking". A compressor attenuates the music, based on the presence (sidechain) of the voice track.

 

3. Gating

When ducking, an audio track gets less loud the moment a second audio track has output. This was done by inserting a compressor in the first track and let the compressor be "sidechain-triggered" by the second track. Gating does the exact opposite: now sound on the first track is only heard when the second track also has output. This is achieved by inserting a Noise Gate or Silver Gate into the first audio track, and pick the second track as sidechain input to the gate. Sound on the second track will open the gate and thus allow the audio from the first track to pass through. Make sure you set the threshold, attack and release times of the gate properly (depending on the kind of effect you want to achieve).

There are multiple uses for this effect. Suppose you have kick and bass tracks. The kick is nice and tight, but the bass tends to be a bit ahead of the kick. Now apply gating to the bass track, with the kick as sidechain input: this way the bass will only sound when the kick plays, and will thus be as tight as the kick. Another use (which is shown in the picture): you have a kick track and a track with some long chords (pad, strings). By applying gating to the chord track with the kick as sidechain input, the chords will be "chopped up" according to the kick signal -- which you may )or may not) find a nice effect.

  • The picture shows the setup described above. Notice that the kick is currently playing (level meter) and the pad is as well -- so apparently the kick has opened the "pad gate".

 

4. De-essing

De-essing is mainly used on vocals. Sometimes sibilant sounds (like "s"-sounds) on a vocal track sound too loud or harsh. Dedicated de-essers suppress the appropriate frequency-range to get rid of this unwanted sound. If you don't have a dedicated de-esser, you can use sidechains to achieve the same effect. Aside: Logic currently has a dedicated de-esser, which isn't very good though, so this trick might still be useful. Apart from that, this same technique can be applied in entirely different contexts as well, so read on.

The basic setup is the same as used when ducking (see above). This time however you don't use an entirely different audio track as the sidechain input. Instead you use a copy of the original vocal track, but strongly enhance the offending frequency, so that this frequency drives the compressor and softens the harsh ess-sounds. In the ducking example, the kick not only drove the bass-compressor, but was also sent to the main output. In the case of de-essing we don't want that: the vocal-copy-with-enhanced-offending-frequency is only used to drive the compressor and should not be part of the regular output -- of course.

Now all that's left to do is create the "enhanced vocal copy". For this you don't need to copy the actual audio file. Instead, send the original vocal track to e.g. bus 1, and locate the Bus-1 object in your Audio Mixer (or take some Aux object, and assign bus 1 to its input -- however, read the section on mixing busses and auxes for possible caveats). In the Bus object insert a Fat EQ, and tune the EQ to enhance the frequency you want to get rid off. Set Bus 1's output to "bus 2". If you have a Bus 2 object in your mixer, set its output to "No Output" or drag its Volume slider all the way down, as you don't want the enhanced signal to turn up in your mix. In the (original) vocal track, insert a Compressor, and pick Bus 2 as its sidechain input. Adjust threshold and ratio on the Compressor such that the Compressor gets mainly triggered by the offending frequency (i.e. Threshold should be low enough so that the Compressor is triggered at all, but no so low that it's being triggered all the time).

This same technique can be used to suppress any frequency range you like. If you e.g. have a drum track where one tom is way too loud, run the drum track through a compressor whose sidechain input takes a copy of the drums. Make sure that in the copy the tom's frequency is exaggerated, using a Fat EQ. Not using a sidechain would compress the entire drum track too heavily, and all the time -- which is not what you'd want.

[top] [contents]

 

.3 Which Logic plugins offer a sidechain input?

In the effects department: Compressor, Silver Compressor, Expander, Noise Gate, Silver Gate and AutoFilter are the Logic plugins offering a sidechain.

In the Virtual Instrument department: ES1, EVOC20 offer a sidechain.

[top] [contents]

 

.4 Which sound-sources can be used as sidechain inputs?

In Logic Platinum each sidechain capable plug-in (i.e. each plug-in with a Side Chain popup menu in its interface) can use Tracks, Inputs and Busses as sidechain inputs. In other versions of Logic sidechain input is limited to Tracks only. These are the only possible sidechain inputs.

[top] [contents]

 

.5 How do you use Auxes as input to a sidechain?

If you want to use the output of an Aux Object as sidechain input, just set the Aux's output popup to one of the busses, and use that bus as the input to the sidechain. Note that this is using the "bus as virtual node", and thus means that you don't need a Bus Object in your mixer at all (see the section on Busses and Auxes of this FAQ for an explanation of the difference). If you need the Aux output as a sidechain input, but also want it to turn up in your mix, do either of the following (supposing some Track object sends to Bus 1, and Bus 1 is set as the input for the Aux):

  1. Set the Aux's output to e.g. Bus 2, and use that as input for the sidechain. Use the Bus 2 Object, with its output popup set to a regular output (like "Output 1-2") to add the Aux's output to the mix.
  2. Set the Aux's output to some regular output (like "Output 1-2"). At the same time use the first Send slot on the Aux to send the Aux signal to Bus 2. Use Bus 2 as input to the sidechain. If you have a Bus 2 object, set it's volume all the way down, or set its output popup to "No Output", in order to prevent the Bus 2 signal from turning up in the mix (which would effectively double the Aux's output).
  3. Use 2 Auxes, both with their input popup set to Bus 1 (i.e. receiving input from the Send of the same Track object). Set the first Aux's output to Bus 2, and use that as the sidechain input. Set the second Aux's output to a regular output, like "Output 1-2".
  • The picture shows option 2 above. The kick on Track 1 is sent to Bus 1, which is used as input for Aux 1. The Aux has an inserted delay, and we want this delay to be part of the mix and have it "chop up" the pad on Track 2 (i.e. this is gating, as described in section 2 above -- this time however the delayed signal should do the chopping instead of the original kick).
    The output of the Aux is set to Output 1-2, so the delay ends up in the mix. At the same time the Aux sends to Bus 2, which is used as input for the sidechain of the Gate that is inserted in Track 2.
    Note that the Volume fader on the Bus 2 object is dragged all the way down, to prevent the Bus 2 signal (which is the delayed signal) from turning up in the mix once more (remember, the Aux already added this delayed signal to the mix). The Bus 1 object's Volume fader is not set to zero, since otherwise the Aux would not get any input. To prevent the Bus 1 signal (which is just the plain kick) from turning up in the mix (which Track 1 already takes care of), we set its output popup to "No Output" (hidden by the Compressor's window). We could have set Bus 2 to "No Output" as well, instead of dragging its fader down.

With Logic's extremely flexible routing, there are probably more ways to achieve the same goal. If you're somewhat unsure of how audio is routed in Logic, trying to get all 3 possibilities mentioned above to work is a good exercise...

[top] [contents]

 

.6 How do you use Virtual Instruments as input to a sidechain?

Virtual instruments do not show up as a possible inputs in a plugin's Side Chain popup. In order to sidechain the output of a VSTi into some effect, just send the VSTi's output to a bus, and use that bus as input to the sidechain.

  • The picture shows a Track with an inserted Compressor which should be triggered by the output of AudioInst 1 ("ducking", see section 2 above). The Audio Instrument's first Send slot is set to Bus 1, and Bus 1 is taken as the sidechain input in the Compressor. Notice how the Bus 1 Object's volume fader is dragged all the way down, to prevent the Bus 1 signal (which is just a copy of the Audio Instrument's output) from turning up in the mix. Also note the Compressor Auto Gain being set to off. See section 2 above, under "ducking", for an explanation.

[top] [contents]

 

.7 What is "ducking", "voice-over", "gating", "de-essing"?

See section 2 above.

[top] [contents]

 

.8 How do you use VST effects that are capable of being controlled by MIDI data?

Some VST effects allow you to influence their behaviour by sending MIDI data to them. As an example: Antares' AutoTune lets you send MIDI note events to the effect, and have the effect retune your audio track to match the pitches as indicated by the MIDI notes currently played. However, if you insert a VST effect in an Audio Tracks' regular Insert slot, there's no way you can send it MIDI data. Since Logic originally wasn't equipped to deal with such effects, the Emagic programmers have created a not entirely intuitive way to deal with this (but, hey, it's working, so lets not complain).

The idea is simple: insert the VST effect in the Instrument slot of an Audio Instrument channel strip. This is important: you don't insert the effect as you normally would, in a regular insert slot, but you insert it as if it is a Virtual Instrument (which it obviously isn't). If you then open the effect's interface you'll see it now suddenly has a Side Chain popup! Here you pick the audio track which you want to run though the effect (i.e. the track that you'd normally inserted the effect in). Assign an Arrange track to the Audio Instrument, and put the controlling MIDI data on that track. Finally turn the volume down on the Audio Track's channel strip, as you don't want the un-effected signal to turn up in your mix.

  • The picture shows a Track with audio data that you want to have processed by dfx Skidder (a very nice free plugin from DestroyFX, that lets you do MIDI controlled gating of audio). The track itself is muted (see volume slider), but used as sidechain input to Skidder, which is inserted as instrument in an Audio Instrument channel strip. The MIDI on the "AudioInst 1" track controls the effect.

[top] [contents]

 

.9 How do you route audio through the Emagic ES 1 synthesizer?

Insert the ES 1 on an Audio Instrument track and then select an Audio Track in the ES 1's sidechain popup. Now set the LFO of the ES 1 to "EXT". The ES 1 will then use the audio signal as a modulation source (instead of the regular LFO). This means that you won't be able to modify the audio signal itself (i.e. you can't route the audio through the filter),but you can use the audio to modulate anything you like. So if the filter cutoff is set to be modulated by the LFO, then now the cutoff will follow the audio signal. You could use this e.g. to have the filter synchronize to a drumloop.

Please also read the answers to similar questions in the chapter on Emagic plug-ins, and the chapter on Effects.

[top] [contents]

 

.10 The Emagic EVOC 20 and sidechains

Please read the section "Setting up and starting to use the EVOC 20" in the chapter on Emagic plug-ins of this FAQ, in which sidechaining with both the Evoc Polysynth and the Evoc Track Oscillator is explained.

[top] [contents]

 

.11 How do you mute the sidechain input signal?

Suppose you use an Audio Track or a Bus as the input to a sidechain of a certain effect or VSTi, but you don't want the sidechain signal itself to end up in your mix. Easy: just turn down the volume on the Track or Bus object. This will not affect the sidechain itself -- apparently the sidechain signal is routed to the effect before the channel strip's volume fader. Instead of turning the volume down, you can also set the channel strip's output popup to "No Output".

In the previous sections you'll find multiple practical examples where the sidechain signal needed to be muted.

[top] [contents]

 

.12 How do you control the amount of influence the sidechain has?

Suppose you use Audio Track 1 as sidechain input to a Compressor that's inserted on Audio Track 2. This means that the compressor will kick in when Track 1 has output (ducking, see section 2 above). No problem getting this to work. However, it might be that the effect Track 1 has on the compressor's functioning is too strong. Ideally you'd like to have a gain-knob right at the sidechain's input, to determine how strong the incoming sidechain signal is. There is no such knob however. Fortunately there is a solution...

Send Track 1 to e.g. Bus 1 (using one of the Sends). In the Compressor on Track 2, pick Bus 1 as the sidechain input. Now the Send dial on Track 1 (i.e. the one that determines how much you send) is effectively your sidechain gain-control. The trick thus is to not use Track 1 as sidechain input directly, but use a bus instead. Note that you don't need any Bus Object for this: you simply use the "bus as a virtual node" (see the section on Busses and Auxes for details). Imagine the possibility of, for example, automating the level of this send, when you need to precisely "dose" the amount of effect through your sidechain ...

[top] [contents]

 

.13 Can a sidechain be controlled through MIDI instead of audio?

Simple answer: no. Sidechains are an audio-thing, period.

Less simple answer: even though the above is true, there might be a workaround which might or might not work, depending on your needs. Insert a VSTi in an Audio Instrument channel strip. In the instrument pick a patch that produces a very constant sound (volume-wise). Now route the instrument's output to a bus, and use that bus as the input to the sidechain. MIDI notes on the instrument track will thus generate a constant-amplitude sound, which in turn will trigger the sidechain.

[top] [contents]

 

.14 Using sidechains to automate Auto Talk Back

First: what is Auto Talk Back? Suppose you have a singer in the booth, and you want to be able to talk to him as soon as recording the current take finishes. Normally you would have a microphone sitting next to you, with an on/off switch. As soon as the take is finished, you hit the 'on' switch, and anything you say is routed through your console and can be heard in the booth. Obviously you want to switch the microphone off again when starting a new take, so that noise from the control room doesn't disturb the recording.

Using Logic's sidechain capabilities, it's possible to fully automate this process. I.e. you let Logic "press the on/off" button for you automatically. The way to set this up is as follows:

On some unused Audio Track put a signal with a constant volume. This can be a dedicated short piece of audio which you set to loop for example. Set the output for this track to "No Output". On a Bus insert a compressor with low threshold and high ratio settings, and Auto Gain set to OFF, and use the Audio Track as sidechain input. Finally route the microphone signal to the bus. That's it (note that you've just created a ducker, as described in section 2 above).

Now when Logic plays, the constant tone signal will trigger the compressor, thus attenuating the microphone signal. As soon as Logic stops, the compressor will no longer be triggered, and you'll be able to communicate with the singer.

Note: if the microphone is attached to one of your soundcard's inputs directly, you could use an Input object instead of a Bus.

[top] [contents]

 

.15 I use an Aux as sidechain, muted its volume, but still hear the Aux signal

You probably have a Bus Object somewhere, assigned to the same (virtual) bus node as your Aux's input. Please read the section on "Using Busses and Auxes simultaneously: potential problems and confusion".

[top] [contents]