🎓 Journalism Training
Courses › Sound Theory

Lesson 4 of 5 · about 15 minutes

Frequency and Tone: Naming the Bands

You'll be able to

  • Name the common frequency bands and their rough ranges in hertz
  • Find a problem frequency with the spectrogram and the Spectrum tab
  • Explain how harmonics make one note spread across many bands
  • Make room for a voice over a guitar or music bed with a cut instead of a boost

Read first: SoundStudio theory section 4, Waveform and spectrum, section 8, Noise and section 9, EQ. Section 9 has a frequency map with labels like low mids and presence. This lesson puts hertz numbers on those labels, adds the slang you will hear from engineers, and spends time on the most common real problem: two sounds fighting for the same space.

Why names help

When someone says a voice sounds "boxy," they are pointing at a frequency range, whether they know it or not. If you can turn "boxy" into "somewhere around 500 Hz to 1 kHz," you can fix it in thirty seconds instead of sweeping knobs for ten minutes. The names below are craft conventions, not a standard. Engineers draw the borders a little differently, and a deep male voice puts its problems lower than a high female voice. Use the ranges as a place to start looking.

The bands, low to high

NameRough rangeWhat lives thereToo much sounds likeToo little sounds like
Rumblebelow about 80 HzTraffic, air handlers, footsteps, mic handling, windThumps, a vague pressureUsually nothing; cut it on voices
Boom (bass)about 80 to 200 HzChest of a male voice, kick drum, bass guitar, guitar bodyBoomy, bloated, "radio DJ in a closet"Thin, small
Mudabout 200 to 500 HzWarmth of most voices and instruments; room build-upCloudy, thick, talking through a blanketHollow, cold
Boxy / honkabout 500 Hz to 1.5 kHzThe core of the voice and most instrumentsCardboard box, nasal, megaphoneScooped, distant
Presenceabout 2 to 5 kHzConsonants, pick attack, the edge that makes words clearHarsh, shouty, tiringDull, mumbly, behind a curtain
Sibilanceabout 5 to 9 kHzs, sh, t, cymbalsSpitting, hissing esses that sting on earbudsLisping, dull
Airabout 10 kHz and upBreath, shimmer, room sparkle, and hissBrittle, hissyClosed in, like an old MP3

Memorize a pair of anchors. Hum from the power grid is at 60 Hz in North America and 50 Hz in most of the rest of the world, with more lines at its multiples (120, 180 Hz). And the ear is most sensitive between about 2 and 5 kHz, which is why a small boost in presence sounds big and a small excess sounds painful.

Harmonics: why one note lands in many bands

An adult man speaking usually has a fundamental somewhere around 85 to 155 Hz, and an adult woman around 165 to 255 Hz. Yet old telephones, which passed only about 300 to 3,400 Hz, were perfectly understandable. They cut off the fundamental of almost every voice. The words survived because they live in the harmonics and in the consonant noise above.

Every pitched sound works that way. A guitar's open G string vibrates at about 196 Hz. Its harmonics land at 392, 588, 784, 980 Hz and up, each one a whole-number multiple of 196. The pattern of those harmonics, and how quickly each fades, is what makes it sound like a guitar. So one guitar note touches the boom, mud, boxy and presence bands all at once.

presence 502005001k2k5k Hz 196392
One guitar note: the fundamental at 196 Hz and its harmonics stepping upward. A strummed chord stacks several of these, and the higher harmonics plus pick noise spill into the presence range where a voice's consonants live.

Why a voice and a guitar fight

Put a narrator over a strummed acoustic guitar and the words get harder to follow, even when the voice is louder. That is masking. When two sounds share a frequency range, the louder one hides the quieter one there. A voice and an acoustic guitar overlap almost exactly: both have body from 100 to 300 Hz, both have their core in the midrange, and the guitar's pick attack sits right in the 2 to 5 kHz presence band where the voice keeps its consonants.

The beginner fix is to turn the voice up. It works a little, then the voice sounds pasted on top. The better fix is to carve a small space. Cut the guitar or music bed by 2 to 4 dB with a wide bell somewhere around 2 to 4 kHz, right where the voice's clarity lives. The guitar barely sounds different on its own. With the voice on top, the words come forward. Do the same in the low end: a high-pass on the voice around 80 Hz, and maybe a little less guitar around 200 Hz, stops the two from piling into mud.

Cut before you boost, and cut the thing in the way rather than boosting the thing you want. Moves of 2 to 3 dB are big. If you need more than 6 dB, the real problem is probably the microphone or the room.

The tutorial's radio spot workflow puts a music bed 15 to 20 dB under the voice. That handles level. The carve handles frequency. Do both and the bed can sit a little higher without burying anyone.

Example: the library board promo

A volunteer cuts a 30 second promo for the Ashford Library Board's book sale. Her voice is recorded close on a USB mic, and the bed is a fingerpicked guitar track from a local duo, The Corner Lot. On playback, the words "Saturday at nine" smear into the guitar.

Before: voice peaks at -3 dBFS, bed 12 dB below, no EQ. After: high-pass on the voice at 80 Hz, a 3 dB cut at 300 Hz on the voice for mud, a wide 3 dB cut at 3 kHz on the guitar, bed now 16 dB under. Same words, and a listener in a car can make out the time without turning it up.

Try it: find the bands, then carve a space

Open SoundStudio. Load a voice recording and a piece of music with guitar or piano in it. Your own recordings are best.

  1. Load the voice. Press V to switch to Wave + Spectrogram. Point at the bottom of a word and read the frequency of the lowest bright band: that is near your fundamental. Point at an "s" sound: you will see a bright smear high up, usually above 5 kHz. Write both numbers down.
  2. Find a pause. Point at the faint glow during the silence. A steady line at 60 Hz (or 50 Hz) is hum. A haze across the top is hiss. Thumps at the bottom are rumble.
  3. Select a few seconds of speech, open the Spectrum tab and use Analyze Selection. Hover across the curve and match the bumps to the band names in the table above.
  4. Open the EQ tab and set Low Cut to 80 Hz, as the tutorial suggests. Listen for rumble vanishing while the voice stays full.
  5. Load the music file. Find a peak in its Spectrum curve between 2 and 4 kHz. Clicking the curve offers a narrow cut, a notch or a wide boost and sends that frequency to the EQ; set the band to a wide, gentle cut of about 3 dB instead of the narrow one.
  6. Press Apply EQ to Selection / All on the music. Then use Mix Files Together to play the voice over the music with the music about 15 dB lower. Close the mixer, press Ctrl+Z on the music to take the cut back out, and play the mix again. Listen for the words getting clearer with the cut in place.

Write down which band each problem lived in and what you cut. Name the problem first, then reach for the knob.

Checklist

  • Rumble below 80 Hz: high-pass voices.
  • Mud around 200 to 500 Hz: a gentle cut clears more than a treble boost.
  • Presence 2 to 5 kHz: where words live; carve space here in the music.
  • Sibilance 5 to 9 kHz: tame it gently, or the speaker lisps.
  • Judge EQ at equal loudness; louder always fools you for a moment.

Check yourself

1. A voice recording sounds thick and cloudy, like talking through a blanket. Which band is most likely overloaded?

2. Harsh "s" and "t" sounds that hurt on earbuds live mostly in which range?

3. A guitar plays a note with a fundamental of 196 Hz. Where is its third harmonic?

4. A narrator is hard to understand over a strummed acoustic guitar. What is usually the best first move?

5. Which filter removes rumble and handling noise from a voice?

Terms in this lesson

Fundamental
The lowest frequency in a note, which sets its pitch.
Harmonic
A frequency at a whole-number multiple of the fundamental: 2x, 3x, 4x and so on.
Masking
When a louder sound hides a quieter one in the same frequency range.
Presence
The range of roughly 2 to 5 kHz where consonants and clarity live.
Sibilance
Hissy energy from s, sh and t sounds, roughly 5 to 9 kHz.
High-pass filter
A filter that passes frequencies above a set point and cuts those below. Also called a low cut.

Full glossary

Sources