EQ and Compression on a Voice

EQ changes the tone of a voice. Compression changes how steady its level is. Used lightly, together, they turn a decent recording into something that sounds like radio. Used heavily, they turn it into something that sounds like a fast-food drive-through speaker.

Do this after noise cleanup and editing, and in this order: EQ first, then compression. Compressing a voice full of rumble means the compressor reacts to the rumble.

How SoundStudio's EQ works

The EQ tab has ten bands plus Low Cut and High Cut. Drag a numbered dot on the curve: left and right moves its frequency, up and down sets its gain, the mouse wheel over a dot makes it narrower or wider, and a double-click flattens it. Click a band to open the band editor and type exact values for Type (Bell, Low shelf, High shelf, Notch), Frequency, Gain and Width.

The EQ is live: tick it on and you hear it while you play. It only goes into the audio when you press Apply EQ to Selection / All, or when you tick Apply Effects at export. Nothing happens on its own. Keep that in mind for every tool in this lesson.

The theory page has a map of what lives where at EQ: shaping tone, and the frequency basics at #t-spectrum.

The voice EQ moves

1. High-pass at 80 Hz

Set the Low Cut around 80 Hz. Most speaking voices carry little useful sound below that, and the lows are where rumble, desk bumps and HVAC live. A deep male voice may want 70 Hz; a light voice can go to 100 Hz. Listen while you raise it: stop just before the voice starts to sound thin.

2. Cut mud at 200 to 400 Hz

Home recordings pile up energy in the low mids. It sounds thick, cloudy, "talking into a cereal box." Find the exact spot by sweeping:

1. Take a Bell band, set it narrow and boost it a lot, around +10 dB.
2. Play the voice and drag the dot slowly from about 150 Hz to 500 Hz.
3. Somewhere it will suddenly sound awful and honky. That is the problem frequency.
4. Turn that boost into a cut: 2 to 4 dB, medium width.

SoundStudio's Spectrum tab can help: click a peak on the curve and it offers Cut 4 dB (narrow), Notch it out or Boost 3 dB (wide), then moves a free EQ band there.

3. Presence at 3 to 5 kHz

Consonants live here, and so does the sense that someone is talking to you rather than near you. A wide boost of 2 to 3 dB around 3 to 5 kHz makes words clearer, especially under music. More than about 4 dB usually turns harsh and tiring. If the voice already cuts through, skip this step.

4. Tame the S sounds

S, SH and the T in "tea" are sharp bursts somewhere up around 5 to 9 kHz, depending on the voice. Boosting presence can make them stab. An EQ cut there would dull every word all the time. A de-esser only turns that band down when an S happens.

SoundStudio's De-Esser is in the Voice tab. It tames harsh S and SH sounds around the Frequency you choose. To find that frequency, play an S-heavy line ("Six city sessions") and watch the live spectrum, or let Analyze point it out: it reports harsh "s" sounds with the frequency measured from the S sounds themselves. Set it so the S sounds sit back without turning into a lisp. If you hear "th" where an "s" should be, back off.

Compression, with numbers

A compressor turns the sound down automatically when it gets louder than a level you set. Then you turn the whole thing back up. The result: loud words come down, quiet words come up relatively, and the voice sits at a steady level a listener doesn't have to ride with the volume knob.

The controls, with numbers for a spoken voice. The theory page covers the same ground at Dynamics.

• Threshold: the level where compression starts. Set it so only the louder words cross it. A starting point of -20 dB is common; adjust from there.
• Ratio: how hard it pushes back. At 3:1, a word 6 dB over the threshold comes out 2 dB over. At 4:1, a word at -12 dB with a -20 dB threshold is 8 dB over and comes out 2 dB over, at -18 dB. For voice, 2:1 to 4:1.
• Attack: how fast it reacts once a sound crosses the threshold. 5 to 20 ms for voice. Too fast flattens the snap of consonants. Too slow lets the front of loud words through.
• Release: how fast it lets go after the level drops. 100 to 250 ms for voice. Too short and you hear it flutter; too long and quiet words after a loud one stay squashed.
• Knee: how gently it starts working near the threshold. A soft knee sounds more natural on voice.
• Makeup Gain: the gain you add afterward to bring the level back up. In SoundStudio this is the only gain the Compressor adds.

The number that matters is gain reduction: how many dB the loudest words are turned down. For a natural spoken voice, aim for 3 to 6 dB on the loud words. If you can clearly hear the compressor working (breathing, pumping, the room swelling up in pauses), back it off.

Presets and Fit to This Audio

The Dynamics tab's Suggested Presets set the Compressor, Limiter and Gate values together but never switch anything on. The Voice preset uses 3:1, with the threshold a little under the voice's average level. Press Fit to This Audio after choosing it and SoundStudio measures the file and sets Threshold, Makeup and Gate to suit. The line under the presets tells you what it measured and set. Treat it as a starting point and listen.

Then tick Enable to hear it live, and when it sounds right press Apply Compressor to Selection / All. If you leave it only enabled, the export will not include it unless you tick Apply Effects. The Export box warns in red when live effects will be left out, which is the usual reason an export sounds quieter than what you heard in the app.