In corporate situations, speech intelligibility is the number one priority for sound reinforcement. The philosophy here is different than when you’re working with full-fidelity music: Your goal is not to make voices sound more aesthetically pleasing, but “clearer.” It’s all about ease of understanding, versus enjoyment.
Like any sound reinforcement situation, all sorts of factors are at play, from room acoustics to microphone choices to speaker technique to processors in the signal chain. But for simplicity’s sake, we’re going to focus on the acoustic properties of speech, how those properties affect intelligibility, and share some very basic tips for EQing out problems in corporate situations where time is of the essence.
The Speech Spectrum
First, a nano-refresher: Speech sounds are created when air vibrates through vocal cords; this frequency spectrum is then shaped by “formants,” resonances in the vocal tract. (This is why two people speaking the same word at the same pitch will sound completely different.)
The human voice sits in a very wide frequency range, with the speaking voice spectrum falling between about 50 Hz to 5 kHz. The majority of energy, however, is concentrated between 300 Hz and 3 kHz (aka the telephone band).
For purposes of intelligibility, we’ll look at this spectrum in the context of language. Words are generally defined by changing vowels and consonants, with consonants being the most important. You can loosely divide the human speaking voice range in three general bands:
The fundamental range falls around 80 to 180 Hz for male voices and about 160 to 250 Hz for female voices.
The “vowel” area ranges between about 350 Hz and 2 kHz, with energy dropping off steeply above 1 kHz at normal intensity.
The “consonant” range falls between about 2 kHz and 4 kHz, with “sibilance” generally falling in the 5 to 8 kHz range.
So, you could consider the frequency band around 1,000 to 4,000 Hz range to be the most important when it comes to perceived intelligibility. These are just general guidelines; there’s a great deal of overlap, and each person’s voice is different, so use these as a starting point.
Dynamic Attack
Very generally speaking, a normal conversational voice lands in the 50 to 70 dB range; the dynamic range of spoken voice can vary up to 20 dB at close range. Most speakers tend to emphasize points by raising their voices. Acoustically, enhancing that vocal effort changes both the volume level and the frequency spectrum of the voice. That center band of frequency emphasis, the “vowel” area, will rise as a voice rises—as much as two octaves. Yet the increase in energy in a voice’s upper frequency range, the “consonant” range, is very small in comparison. So in other words, yelling does not increase intelligibility.
Quick, Corrective EQ
Understanding these fundamentals will help you root out sonic anomalies, which can often be remedied with a few quick-and-dirty EQ tricks if you’re crunched for time.
For general clarity, drop in a high-pass filter around the low end of the speaker’s fundamental. Try an 80 Hz setting and move up from there.
Sometimes, a wide-Q boost of a few dB in that crucial 1 to 4 kHz range will add the intelligibility you need.
Need to give a “thin” voice a little more authority? Beef up the 100 to 250 Hz range.
Is your speaker a poor articulator? Try enhancing the 2 kHz to 5 kHz range (but watch for sibilance issues).
Remember, these are very general tips, and every speaker and situation is unique. We’ll take a look at more complex processing solutions in upcoming columns, so watch this space!
Sarah Jones is a writer, editor, and content producer with more than 20 years' experience in pro audio, including as editor-in-chief of three leading audio magazines: Mix, EQ, and Electronic Musician. She is a lifelong musician and committed to arts advocacy and learning, including acting as education chair of the San Francisco chapter of the Recording Academy, where she helps develop event programming that cultivates the careers of Bay Area music makers.