by Bruce Bartlett
Hey Bruce, how have you been? I have a quick question: When separating instruments in the mix, does it sound better to have them slightly overlap or make sure that the frequency ranges are separated completely? Thanks –Nick Papps
Each instrument produces a wide range of fundamental frequencies and harmonics. All the frequencies that an instrument makes, and their relative levels, are called the spectrum of the instrument.
It’s impossible to separate the spectra of the instruments completely because each instrument covers a wide range of frequencies. So there always has to be some overlap. For example, a bass guitar produces frequencies from about 41 Hz to 10 kHz including its harmonics. That doesn’t mean you should filter out everything below 10 kHz in the guitars so they don’t overlap the bass guitar’s spectrum. The idea is to reduce the bass a little in the guitars to leave more space for the bass guitar. You might reduce the upper mids or highs in an acoustic guitar slightly so the vocal can be heard clearly, filter out highs above 4 kHz in electric guitars so they don’t cover up the cymbals and tom attacks, and so on.
As another example, if you took all the mids and highs out of the bass guitar so that they don’t mask other instruments, the bass guitar would lose its clarity and definition. The EQ needed to reduce masking is only about 3 to 6 dB in most cases, but you have to go by ear.
It’s not a good idea to solo each instrument and EQ it to sound good by itself. You need to equalize it in context with the other instruments, because each instrument masks others and affects how each instrument sounds. That also affects the EQ you need to apply.
When you solo an instrument, a good use for EQ is a highpass (low-cut) filter. Filter out frequencies below the lowest fundamental frequency that the instrument produces. Here’s one way to do that: Start with a highpass filter set to a Q of 1.7 and a frequency of 40 Hz. While playing the track, gradually raise the filter frequency until the sound starts to thin out, then back off a little. Filtering out the deep lows on each instrument and vocal reduces breath pops, low-frequency leakage, and rumble from traffic and air conditioning.
Good luck, Nick!
Hey Bruce, in your book you mention how to thicken up lead vocals a bit. If I understand correctly, you say record a second track and make the volume slightly lower than the first. Am I supposed to record a second take, or should I just duplicate the first pass? Also, are there any other good tricks to do to lead vocals? Thanks — Nick Papps
To thicken the vocals, usually you record a second take. This is called “doubling the vocal”. It sounds less mechanical than duplicating the first pass and sliding the second pass about 25 milliseconds earlier or later than the first pass. Pan both tracks to center. In Heavy Metal productions, it’s common to double the rhythm guitar, and pan the two parts all the way left and right.
John Lennon asked his recording engineers if they could double his vocals electronically. They did it with a tape recorder and called it “ADT” for “Automatic Double Tracking.” You hear it on John’s vocal in “I am the Walrus”. You can create the same effect with a DAW using the time-sliding method.
A similar effect is called “Chorus”. A chorus stomp-box or plug-in delays the incoming signal about 15-30 milliseconds, adds it to the original signal, and varies the delay. The chorus effect varies its own delay; you don’t need another plug-in to do that. It sounds less mechanical than combining a vocal with its delayed replica using a constant delay time.
The most popular vocal effect is reverberation, which simulates room acoustics. Set the reverb time to about 0.5 second to simulate a small room, or about 1.6 seconds to simulate a concert hall. Or just leave the vocal dry (without reverb), as in “One Hand In My Pocket.” Contemporary recordings tend to be drier than 1980’s recordings. Fast tempo songs usually sound best with shorter reverb times.
A popular vocal effect for rockabilly or rock ‘n’ roll genres is slap echo or slap-back echo. It’s a single echo with the delay set to about 130-150 milliseconds. It sounds better if you turn down the high frequencies in the delayed signal. You can hear slap echo in Lennon’s vocal in “A Day in the Life” and “Instant Karma”, and in most songs by Creedence Clearwater Revival.
A slow repeating echo (delay time about 500 msec and some feedback to make the echo repeat), mixed very quietly under the vocal, can be effective for slow ballads. Set the delay time so that the echo repetitions match the tempo of the song.
Compressing the vocals is very common to keep their volume more constant. For starters, try a 3:1 ratio, 20 msec attack time and 250 msec release time. Gradually turn down the threshold until the gain reduction is about 6 dB. (Of course, these suggestions do not apply to all songs). One vintage compressor plug-in that sounds great is the dbx 160. Vocal compression is used almost always with rock. For folk music, use gain riding (volume envelope automation) instead because it sounds more natural.
If the singer’s “s” sounds (sibilant sounds) are excessive, use a multiband compressor set to compress only the band from 5 kHz to 20 kHz, where the sibilant frequencies are. Try a 1 msec attack time, 200 msec release time, and 5:1 compression ratio. Gradually turn down the threshold until the amount of sibilance sounds good to you.
Bruce Bartlett is a recording engineer and microphone engineer (www.bartlettaudio.com). He is the author of Practical Recording Techniques 6th Edition and Recording Music On Location. Please send Bruce your recording questions via L2PNet.com.