Support Tutorials

Tech Note #5: Duck! The levels are falling!!!!

By Marc Schonbrun

Revision 1.0 - June 28, 2007




In my tutorial video “Firewire Returns + SpectraFoo”, I employed a cool +DSP effect to automatically lower the music while I was talking. This is also known as “ducking”.

To implement this effect, I set up a graph in the +DSP window (⌘4) as follows:

  1. I added the +DSP Plug-in MIOComp S/Sfrom the “Insert” drop-down list:
    image001

    MIO Comp is a stereo compressor with a mono sidechain input (the sidechain is critical for making the ducker).

  2. Now, we have to cable the MIOComp plugin to the graph I/O:

    1. Click and drag from DAW 1 and DAW 2 to the left and right input of the MIOComp.
      image003
    2. Click and drag from Analog 1 (the microphone input) to the SC input of the MIOComp (This is the side-chain input):
      image005
    3. Click and drag from the left and right outputs on the MIOComp to Process 1 and Process 2.
      image007
  3. At this point, everything is wired up. Now, double click on the MIOComp primitive to bring up its user interface:
    image009

    At the top of the UI, you will see a “Detector Source” with a text box that you can toggle. Click and hold down so that it reads “SC-IN” which stands for “Side Chain Input”.

What we’ve done is tell the compressor to listen to the sidechain (which happens to be my voice) to engage the compressor and not the level of the incoming audio from iTunes.

The threshold should be set very low for this instance, but the exact amount of threshold is dependant on how loud you are talking. -37.7 dB did the trick for me.

Go ahead and depress the “Comp Enable” switch to turn on the compressor. I’d recommend turning off “Auto GN” or automatic gain, as you won’t need it here. The idea is to turn it down low while you talk.

Now, in order to hear this, we have to switch to Mixer Output/Routing so that we can listen to the output of the compressor (which is fed from the process returns 1 + 2). Press (⌘3) to take you to the output Patchbay and change FW Return 1 and 2 to Process 1 and 2. This is now feeding the process returns to SpectraFoo from the compressor. To be able to hear the effect live, go ahead and switch your monitor outputs to Process 1 + 2 as well.

image011

The next part is all ears…

Now that we’re all patched in and we can hear the effect – go ahead and start a track in iTunes. Switch back to the +DSP (⌘4) to look at the compression settings.

You’ll want to set the ratio very high as we don’t want this a little lower, we want this a lot lower. I chose a 17.4:1 ratio. Normally, ratios this high would sound very “slammed”, but it’s only acting as a leveler and you won’t hear the music when I’m speaking very well, so it’s not that important.

Once the ratio is set: the two most important knobs to set are the attack and release of the compressor.

Attack sets how long after you start speaking until the compressor lowers the volume of DAW 1 + 2. I experimented with this until it felt natural; you don’t want it too slow, and you don’t want it too fast. 82.4 ms (milliseconds) was the right choice for the job; as I said, it’s all ears. The release sets how long the compressor “holds on” to the signal (in this case, how long it turns iTunes down) after you’ve stopped speaking. If it’s set too fast, it will sound odd, if it’s set too slow, it will take too long to bring the music back.

The hardest part about the release was allowing enough time for me to breathe and naturally pause between sentences without the level creeping up while allowing me to stop speaking and bringing the music back up at a natural pace. Again, I listened and moved the control until it sounded right and 1000 milliseconds was right.

That’s it!

Looking back at the SpectraFoo/FireWire Return Tutorial movie that I used this in, you can see the action of the ducker in SpectraFoo. I have copied a frame from the movie below. The red arrow points to the time that I started speaking over the music. You can see that above the arrow (before I started talking), the music is loud, and below the arrow, the level of the music has been reduced to make room for my narration:

DuckerAction

If you want to watch the movie, and listen to the ducker in action, here is the link: Firewire Returns+SpectraFoo Tutorial Movie.

Considered yourself ducked.

Revision History:

  • Revision 1.0 - June 28, 2007 - Initial Release