Stupid DAW Tricks Part I : Frequency-based Dynamic Control
By Tom Volpicelli on July 14 2008 03:00 PM | Permalink | Author Info
We’re all familiar with functions like de-essing and multi-band compression. It’s easy, slap on a plug-in, fiddle around a bit and you’re done. There are times however when some issues are not easily controlled through a single setting and the “remedy” starts getting in the way. Wouldn’t it be nice if we could add a little bit of “intelligence” to make up for these difficulties?As an example, suppose you have a track where the vocalist starts singing several sibilant syllables. You determine that the 7-8k range needs a bit of taming and you set the de-esser to reduce these at a given threshold. This works great for most of the song, but in some sections the vocal is much lower in level and it doesn’t reach the threshold of the de-esser (so no de-essing is performed), or so loud that the de-esser becomes obtrusive. In this first installment of Stupid DAW Tricks, we'll examine an easy method for setting up our DAW software to handle this kind of situation better and more efficiently.
Before we get into the implementation details, a little background on how a typical de-esser works. A de-esser usually consists of a compressor that is controlled by a side-chain that is fed a duplicate of the original signal equalized so that the offending frequency is most prominent. This equalized signal does not become part of the output of the compressor, but is only used to control how and when compression takes place by using this as the input to the circuit or software logic used to control the amount of compression (Figure 1.)
To duplicate this functionality in a DAW our first task is to be able to send the original signal to at least two separate locations, the input of the compressor for the source signal to be processed, and another to a channel with an EQ that can be used for the side-chain to the compressor. We also need to decide if we should insert the compressor in the original track or use two dedicated channels for processing. If using an analog console, latency delay due to digital processing isn’t a concern, so inserting the compressor in the source track should be fine. However when we are dealing with digital workstations latency is a concern. We don’t want the signal to the side-chain to be delayed compared to the source signal or the compressor’s response will be late and will miss the sibilant part of the source track.
Many DAWs have latency compensation, but since it’s DAW-dependent and may not work as intended when feeding back a signal I’ll take the more conservative approach and use two dedicated channels for processing. This can be done as follows:
- Configure the source channel (channel 1 in Figure 2, below) so that its output is sent (using your aux or bus sends) to two other channels. Turn off the output to the main bus for channel 1 so that the source signal only goes to these two other channels.
- In channel 2 insert an EQ/filter. Using a combination of high pass and low pass filters, band-pass filter the signal to the frequency that you want the compressor to react to. For example if we wanted to remove the 7-8k range, set the high pass filter somewhere around 5k with a steep slope and the low pass around 10k. You can also use just the high pass filter if you want to de-ess all frequencies above a given range, or use only a low pass filter if you want to remove plosives or in situations like balancing uneven frequencies of a bass guitar. Output this channel to another bus or aux send again making sure that it is not part of the mix but only used for controlling the compressor.
- In channel 3 insert a compressor that has side-chain capability and output this channel to the main bus. The side-chain of the compressor should be setup to accept the bus or aux send from channel 2 above. Settings on the compressor are dependent on the problem that you are trying to fix, but as in most de-essing applications you want a fast attack and release so that it does its job quickly and gets out of the way. Set the threshold so that it reacts only to the problem you are trying to fix, and the ratio to the amount of reduction you would like.
We now have a basic working de-esser. One immediate advantage is that we can now choose whatever compressor we like instead of being limited to onesupplied in a de-esser plug-in (assuming that the compressor that you want to use has side-chain capability). Want a warmer overall sound? Choose a vintage or tube emulation style compressor. Want transparency? Choose a comp with less “color”.
Another advantage is that we have complete control over how the side-chain behaves. Let’s go back to our original problem of de-essing a vocalist with a wide dynamic range. One way to “add intelligence” is to automate the level of the filtered track (channel 2). During softer sections raise the level of channel 2 so that it hits the fixed threshold of the compressor by the same amount as the louder sections.
In addition to automating channel 2 you could try adding a second compressor with make-up gain after the EQ. This will even-out the level of the side-chain so that the de-esser reacts more consistently.
In cases where you do not want the de-esser to engage you can automate a mute, or add an expander or gate to channel 2. This technique often comes in handy in mastering. For example when you might want to de-ess a vocal in some sections, but do not want to de-ess a hi hat when there is no vocal.
With total control over the side chain there are also some interesting functions and effects that we can create.
One example would be to use another track other than the source track for the side-chain. Let’s say that you have some really huge guitars with a lot of bottom end. During parts of a song where these are playing by themselves it sounds great, however once the bass and the rest of the music comes in there too much competition and the low end starts to become a bit too “large” or muddy. We can use an EQed copy of the bass guitar as a side-chain to lower the guitars in those parts in order to act as a sort of “ducker”. This can also work between kick and bass.
Another interesting idea might be to use a click track for the side-chain to create an interesting tremolo effect. Beyond using the side-chain channel to control a compressor it may also be used to control a gate or expander and the same ideas above apply except that your are raising the source track (with an upward expander) or gating a track based on the frequency of the source or other tracks.
As they say, the possibilities are endless and you are only limited by your imagination.