TCRM 15 focused on the specific uses of eq to sculpt a mix and the individual sounds within it. It showed how eq can be used to address technical problems (such as noise, rumble, proximity effect and bleed), adjust the tonal characteristics of individual instruments, make frequency-space within the mix, and master the final stereo (or surround) product. But these functions represent only one side of eq usage. This month we’ll explore ways to use eq as a creative effect.
The further away eq moves a sound from a more “natural” tonal quality, the more it can be considered an effect. Though a lot can be done in this regard using traditional board-style filters, often it is the more extreme filter designs that are better suited for creative sonic mangling. Steep slopes, true band-pass abilities, and both resonance and feedback controls all make for better effect-style eq. The ability to automate, modulate, or otherwise trigger control parameters is also very handy.
Band-pass (or extreme gain notching) filters are great for creating emulations of frequency-limited phenomena, or just carving out chunks of an instruments spectrum. The low-fi sound of an AM radio can be emulated by using a bandpass filter to limit the frequency range of a sound to between approximately 200 to 4,500 Hertz. Mini-amp emulations can run about the same range (just sprinkle in some extra distortion). The distinctive characteristics of telephones can also be recreated by limiting range. Both older and newer telephone technologies limit received bandwidth to between 200 to 5000 Hz or so. Cell phones, however, can be harder to emulate well because of the audio gating and data compression schemes also employed.
The distinct sound of listening through a pipe or tube is best recreated using a comb filter. When sound moves through an enclosed space of this type, natural resonances accentuate frequencies that are whole number multiples of a lowest fundamental. Similarly, a comb filter creates a series of individual notches across the frequency spectrum based on multiples of the lowest set frequency. The higher this lowest frequency is set to, the shorter a length of pipe it will emulate. If no comb-filter is available, notch filters can be set individually to do the trick. If the fundamental is low (the pipe is long) then many notch filters will be needed to fill out the audible range. (TCRM 19 and 20 will also show how a delay can be used to achieve this type of comb filter)
Another common (and related) use of bandpass and comb filters is to generate the sound of a bullhorn. This is basically a band-limited comb function, with some distortion added for good measure. The audio needs to be restricted to a range from 200-3500 Hz with a band-pass filter or a combination of HPF and LPF. A comb filter should then be added whose lowest frequency is around 500-600 Hz. Finally, distort as needed (amp simulators work great).
As mentioned in TCRM 14, some filter designs have a distinctive “ring” at the cutoff or center frequency. This generally happens when either the resonance or feedback control is turned up. In some cases, this can actually cause the filter to self-oscillate, generating its own constant tone. Similar effects can also happen when using a very large Q, or when the gain is cranked up very high. These resonant peaks can be used to simulate acoustic formant ideas or, when strong enough, can lend a very electronic/synthesized tone to an effected sound. Below are some more ways to take advantage of this.
By “dynamic effects” I don’t mean compression or limiting—that’s coming next in TCRM 17 and 18--but rather effects that change over time. One of the simplest and most common methods is the filter sweep. This is where the frequency of a filter is slowly adjusted over time as the music plays. It is a mainstay of techno and other dance-club forms where it is generally done rhythmically or across several measures. LPFs with a fair amount of resonance are particularly useful for sweeps, as the accentuation of the cutoff frequency makes the filter change all the more obvious.
Don’t get the wrong idea though, the basic eq sweep is not only for electronica freaks. Ever heard of Jimi Hendrix, R&B, or Funk music…? That’s right, the sweep is also the basic operating principle behind the ubiquitous wah effect so many guitarists have fallen in love with. These are really based on lowpass or resonant bandpass filters with sweepable frequency controls. By moving this peak up and down while playing, the performer can loosely emulate a vocal sound or just quickly modify his/her tone.
However, wah effects become even more powerful when used in conjunction with other eq effects, or even another wah. In fact, when two LPF wahs are placed in series, with the first left at around 2-3 kHz while the second sweeps through the 500-2000 Hz range, a much better emulation of vowel sounds can be made. Similarly, placing a few static notch or band-pass filters before a wah can make a huge difference.
Performing sweeps is generally done in one of three ways. The most straight-forward method (but least precise and most labor intensive) is to simply do it by hand. Good ol’ knob twisting to the beat! Many outboard and software eqs also allow the user to program a specific rate and depth (frequency range) for an automatic sweep. The biggest problem with these is getting the sweep to start exactly on the intended beat and stay in sync with the music over long periods.
A final, and more powerful, solution is to use some sort of automation. Many software plug-ins, as well as a numerous newer hardware eq units, allow automation of the filter frequency. This can be done by MIDI sequencing the data, by auto-write functions, or by the typical graphic-style editor found on most DAWs. In some cases, even the classic analog filters from older synthesizers can be automated through MIDI controller codes using a MIDI-to-CV (control voltage) converter.
No matter how it comes about, eq automation can be a great creative tool, and for a lot more than basic sweeps…. For instance, here’s a fun idea for automating an instrument track:
First, set the filter so that its resonant boost is obvious. Then, set the frequency to the note being played, or to some obvious intervallic relationship (such as an octave, perfect fifth, perfect fourth, or third). Automate the eq to follow the notes of the instrument it’s effecting in such a manner that it plays in unison or creates a harmony to the musical line. I suppose you could even try some 2-voice counterpoint! (has anyone hooked up that electric generator up to J.S. Bach yet?)
Dynamic and Modulating eq
There are also many filter designs which directly incorporate time-based elements. Dynamic equalizers, for example, measure an incoming signal’s amplitude and map that to the filter’s frequency control. These can be set to increase their frequency when the input signal gets louder. Depending on the settings, the effect can range from subtle, to auto-wah, to practically re-synthesized sounds. Some of the more sophisticated designs include a secondary trigger input so that the dynamic contour of a control input (audio, MIDI, or CV) will be mapped to the frequency of the filtering done to the primary signal input.
Modulating filters also change their frequency over time, but at a much faster rate. This is because they do so by acting instantaneously, based upon a control waveform (such as a sine, triangle, saw, or square). As the modulating waveform moves up into the positive portion of its cycle, the frequency of the filter rises. When the waveform descends, so does the filter frequency. Often, these control signals are set at a very low frequency so the filter changes are perceived discreetly, but they don’t need to be. Control signals (modulators) whose frequency approaches the audible range will may start to be heard as their own tone, and even create sidebands. The creation of these sidebands and the perception of a new timbre makes these higher frequency modulations a form of audio synthesis all their own.
There are many vocoder designs and variations on this classic effect. The basic method, however, is really quite simple and worth explaining here.
The frequency range of an input signal called the modulator (usually a vocal, but not always) is split into numerous bands using band-pass filters. The amount of energy present in each band is measured and used to control the output level of corresponding frequency bands of a second audio signal, called the carrier (often a keyboard or guitar). By doing this, the frequency/time contour of the vocal signal is mapped onto the second carrier signal lending it a vocal-like quality as well. This second audio source can come from live instruments, recordings, noise generators, or oscillators. The effectiveness (amount of wackiness and/or intelligibility) of the effect depends on the exact nature of both signals as well as the number, frequency, and nature of the filters. As a rule, all other things being equal, the more filter bands available, the more intelligible the results will be.
On top of this basic operating principle, many specialized functions have been added by different manufacturers. Rather than rely on a vocal input, some vocoder models can use electronic vocal emulators. Some include detector circuitry so that the unit can respond differently to vowels versus consonants. Still others have the ability to shift formant frequencies by shifting or remapping the frequency relationships between vocal (modulator) and carrier signals. This can be used to either “gender-bend” or just further modify the vocoding effect. Further effects can be made on units that allow modulation of the frequency and/or amplitude of the carrier, or the center freq of the filter bands.
Of course, the control input to a vocoder doesn’t have to be vocal…. Synthesizers, samples, and acoustic instrumental sources can also create some very cool effects. Also, by switching the normal input configuration, vocals themselves could be the carrier and guitars the modulator. One great trick is to vocode long sustained guitar sounds with the sound of a drum track to produce choppy, rhythmic guitar sound that has a drumlike sonic character!
Vocs v. Talks
Often confused with vocoders, talkboxes are a physical/acoustic way of imprinting vocal qualities on non-vocal sounds. These designs use enclosed boxes, with a speaker inside and a tube running out. When the tube is placed just inside a performer’s mouth, the sounds coming from inside the box are naturally filtered by the human mouth, throat, vocal folds, lungs, and nasal cavity. A microphone is then used to capture the newly filtered sounds. Some of the more famous uses of this effect are on "Do You Feel Like We Do" by Peter Frampton and "Livin' On a Prayer" by Bon Jovi.
Well, that does it for our three-chapter look at eq. Next time we’ll discuss compression and other forms of dynamics processing.
John Shirley is a recording engineer, composer, programmer and producer. He’s also a Professor in the Sound Recording Technology program at the University of Massachusetts Lowell and chairman of their music department. He considers creative use of eq and modulation techniques paramount to human evolution… and in using good ol’ Johann Sebastian to solve our energy needs. Check out his wacky electronic music CD, Sonic Ninjutsu, at http://cycling74.com/products/c74music/
Supplemental Media Examples
The following audio examples of some of the techniques described in TCRM 16. For continuity, each use the same raw tracks as TCRM 15, taken from a demo recording of “Liars” by The Bay State.
A bandpass filter, or a combination of highpass and lowpass filters, can be used to create a low fidelity effect that emulates things like small radios, mini-amps, or telephones. Here, the vocals are band-limited and lightly distorted. The end of the line (“…to say yes.”) is left dry: TCRM16_1.wav
To demonstrate the effect of comb filters even better, a guitar is used. First, the dry guitar without the comb filtering: TCRM16_3.wav
Now with a comb filter (does it sound like it’s going through a tube or pipe?): TCRM16_4.wav
Next, the frequency of the comb is reduced, adding more bands and changing the timbre. If we continue to use our pipe model… the pipe just got bigger! TCRM16_5.wav
By sweeping the cutoff frequency of a lowpass filter with resonance and a higher Q, a wah effect is created: TCRM16_6.wav
A sine-modulated bandpass filter with resonance can create anything from subtle to wacky or transformative effects. You decide which this is…: TCRM16_7.wav
The original guitar recording through a vocoder: TCRM16_8.wav
The original guitar recording through a talk box: TCRM16_9.wav
Thanks again to The Bay State for the use of the raw tracks to this demo recording of “Liars.” All material used by permission; all copyrights reserved. Demo recorded by Michael Testa.
The Bay State is Tom Tash, Drew Hooke, Susanne Gerry and Evan James.
Check out their music on iTunes (including the commercial release of “Liars” or visit them on facebook at: http://www.facebook.com/thebaystate