This month, iZotope visited Together Boston and presented workshops on drum production, mixing and mastering, and vocal editing! During the mixing and mastering workshop, we deconstructed a drum and bass remix of singer/songwriter Sarah Blacker’s “Shiver,” focusing on controlling dynamics, instrumental separation, and stereo image.
In this post, we’ll be exploring one of the mixing techniques we dove into at Together: Frequency ducking using Trash 2.
When mixing, decisions are constantly being made about how to highlight certain instruments in different sections of the song. If two sounds seem very important to the song, but are fighting for attention, each needs to be given its own space.
There are many ways to deal with this issue:
Equalize (EQ) the sounds differently to give them each different sections of the frequency spectrum to own.
Pan them to give them different sections of the stereo image to own.
Sidechain compress them so that when one sound comes in, the other pulls back.
What if the sounds in the mix are similar? Since their most important frequency content is in the same range, EQing can lead to serious sacrifices without EQ automation to cut frequencies only when the sounds are fighting. We could sidechain compress one off the other, but then information in other sections of the frequency spectrum will also be affected which may create audible “pumping.”
With Trash 2, a far more transparent and time effective option is available—sidechain frequency ducking.
In this example, the female vocal and a distorted mono synth both have a lot of energy around 600Hz. Set up Trash 2 so that when the vocal crosses a set threshold, an EQ cut around 600Hz will trigger on the synth. This way, when the vocal is not the focus of the mix, the synth has all the frequency information important to it, but while the vocal is the focus of the mix, the synth will get out of the way of the voice in a subtle manner that’s not apparent to the listener.
Setting up Frequency Ducking with Trash 2
In the DAW, add Trash 2 as a plug-in on the synth track and select the sidechain as the vocal send. Each host will have a slightly different way to set up sidechaining, but the user manual is always a great place to head for the proper steps. Open the Filter 1 module in Trash two and select a Clean Peak as a filter node.
After setting the node to the frequency to duck, select the Modulation tab and select Envelope as the Mod Type. To the right of the Envelope section are controls for the Target Node. The Target Node is the position of the node when the modulation is engaged, so select the same frequency as before and then select the amplitude to reduce the node by. For reduction of a broader range of frequencies, use a lower Q. For ducking of a narrow range, use a narrow Q.
Now that the amount of amplitude reduction and Q are set up, let’s take a look at the “Envelope” section of the Modulation tab. Select Sidechain so that the threshold will be triggered by the sidechain. In the DAW, start playback in a spot where both sounds can be heard at once. Adjust the threshold to a level a little below the incoming signal, which can be monitored visually with the bar above the threshold slider. From here, it’s all about using your ears to adjust the Attack, Release, Gain and Q to get the desired amount of reduction.
By using this technique, the fighting frequency in the synth pulls back when vocal is active, and leaves the synth sound untouched when the vocal is not. Listen to the audio examples below.
First, listen to the mix without processing. In a song like this with heavy bus compression, the overcrowding of the frequency spectrum becomes even more apparent.
Second, listen to the mix with the frequency ducking dialed into Trash 2.
In this example, the body of the vocal commands more of a presence, with the synth pushed slightly out of the way. If sidechain compression had been used, a lot of the high end sizzle of the synth would be lost, as that would be attenuated as well. By using this technique, a section of the frequency spectrum has been hollowed out a bit more for the vocal to sit in. Because of the threshold control, when the vocalist is not singing, the synth has full reign of the frequency spectrum -- it adapts over time and steps to the side when it needs to!