Compression ratio
The compressor's strength: at 4:1, every 4 dB above the threshold comes out as only 1 dB.
See it
What it is
Compression ratio says how strongly a compressor reduces movement above the threshold. At 4:1, every 4 dB the input tries to rise above the line becomes 1 dB at the output. With a -18 dBFS threshold, an input at -10 dBFS sits 8 dB over the line and settles around -16 dBFS before makeup gain, assuming a hard knee and enough time to react.
Lower ratios such as 2:1 are gentle; 3:1 or 4:1 is a practical starting range for steadying speech; very high ratios approach limiting. Hardware often fixes the choices for you: the UREI 1176 gives you 4, 8, 12, and 20 to 1 as buttons, and pushing all four in at once is the famous all-buttons-in mode engineers use to wreck a drum room on purpose. Pick the ratio together with the threshold, watching the gain-reduction meter and listening for whether the performance still breathes.
Gotcha: ratio is not a percentage and it does not predict the result by itself. A 10:1 compressor whose threshold is never crossed does nothing, while 2:1 with a low threshold can work constantly. Knee, attack, release, and any automatic makeup gain can make the same ratio sound completely different.
Ask AI for it
Use FabFilter Pro-C 2 on this voiceover with a 3:1 Ratio, 6 dB Knee, 10 ms Attack, and 100 ms Release. Set the Threshold for 3 to 6 dB of gain reduction on the loudest phrases, leave Auto Gain disabled, and level-match the processed signal to bypass with manual makeup gain. Keep consonants crisp and reject any setting that pumps between words.