Understanding Dynamic Range Compression

Dynamic Range Compression (DRC) is a fundamental audio processing technique that reduces the difference between the quietest and loudest portions of an audio signal. In the context of dialogue, this means bringing up whispered phrases while taming shouted outbursts, creating a consistent listening experience. Without compression, audiences often struggle to hear soft speech or are startled by sudden peaks. The goal of DRC is not to squash life out of a performance but to maintain intelligibility across varying delivery styles and recording conditions.

The core principle involves automatically lowering the gain of signals that exceed a chosen threshold. The amount of gain reduction is determined by the ratio, and the timing of this reduction is governed by attack and release settings. Modern compressors also offer features like knee for smoother onset of compression and lookahead to anticipate peaks. Understanding these parameters is essential for applying DRC tastefully to dialogue.

Key Parameters of a Compressor

Threshold

The threshold sets the level above which compression begins. For dialogue, a good starting point is to place the threshold just above the average speaking level, often around –20 dBFS for a well-recorded voice. This ensures that only peaks are attenuated while maintaining natural dynamics in softer passages.

Ratio

The ratio defines how much gain reduction is applied once the signal exceeds the threshold. A ratio of 3:1 means that for every 3 dB over the threshold, only 1 dB passes through. For dialogue, ratios between 2:1 and 4:1 are common. Higher ratios can work for voiceovers or narration that require extreme consistency, but they risk sounding unnatural if overused.

Attack and Release

Attack time controls how quickly the compressor responds after the signal crosses the threshold. Fast attack times (1–5 ms) clamp down on sharp transients like plosives or lip smacks, while slower attacks (10–30 ms) preserve the natural punch of consonants. Release time determines how long it takes for the compressor to stop reducing gain after the signal falls below the threshold. For dialogue, release times of 50–150 ms often work well, though faster releases can cause audible pumping and longer releases may create a dull, “stuck” feel.

Knee

The knee parameter smooths the transition from uncompressed to compressed states. A hard knee (0 dB) applies compression abruptly at the threshold, which can be useful for aggressive limiting. A soft knee (6–12 dB) begins compressing gradually before the threshold, yielding a more transparent effect that’s often preferred for natural-sounding dialogue.

Makeup Gain

Compression reduces the overall signal level, so makeup gain is used to bring the output back to an appropriate level. Automatic makeup gain features can be convenient, but manual adjustment allows finer control. The goal is to match the perceived loudness of the original signal while enjoying the improved consistency.

Choosing the Right Compressor for Dialogue

Different compressor designs impart distinct sonic characteristics. For dialogue, transparency is usually desired, but some situations benefit from character.

  • VCA Compressors (e.g., API 2500, SSL Bus Compressor) are clean, precise, and great for controlling peaks without coloration. Many DAW stock compressors emulate VCA behaviour.
  • FET Compressors (e.g., UREI 1176) offer fast attack and a punchy, sometimes aggressive sound. They can add presence to thin dialogue but require careful setting to avoid distortion.
  • Optical Compressors (e.g., LA‑2A, Teletronix LA‑3A) provide smooth, musical compression with a slower attack, making them excellent for preserving the natural flow of speech. They’re a favourite for voice‑over and podcasting.
  • Digital/Plugin Compressors like FabFilter Pro‑C 2 and Waves R‑Compressor offer advanced features such as sidechain filtering, lookahead, and variable knee. They are versatile tools for any dialogue‑focused workflow.

When selecting a compressor, consider its ability to handle fast transients without introducing artefacts. A good dialogue compressor should allow you to hear the vocal character intact after gain reduction.

A Step‑by‑Step Dialogue Compression Workflow

Step 1: Prepare the Track

Start with a clean recording. Apply a high‑pass filter around 80‑100 Hz to remove low‑frequency rumble and microphone handling noise. This prevents the compressor from reacting to unwanted subsonic energy. Use a de‑esser if sibilance is prominent, as compression can exacerbate harsh “s” and “sh” sounds.

Step 2: Set the Threshold

Play the dialogue at its loudest section and lower the threshold until you see 3–6 dB of gain reduction on the louder phrases. Avoid exceeding 8 dB of reduction in one stage, as this can cause audible artefacts. If more compression is needed, consider serial compression (two compressors in sequence with low ratios).

Step 3: Adjust the Ratio

Begin with a ratio of 3:1. If the dialogue still has excessive variation, increase gradually to 4:1. For whispered sections that are still too quiet, a lower ratio (2:1) may suffice. Trust your ears over visual meters.

Step 4: Fine‑Tune Attack and Release

Set the attack to 5–10 ms to catch peaks without dulling the initial consonants. Adjust release between 50 and 150 ms; a good test is to listen to a phrase ending with “p” or “t” – if the release is too fast, you’ll hear a “pumping” or “breathing” sound after the word. If it’s too slow, the compressor may still be reducing gain when the next word begins, making the dialogue feel squashed.

Step 5: Use a Soft Knee

Enable soft knee mode (6–10 dB) for a more natural transition. This is especially important for dialogue that has dynamic variations common in emotional performances.

Step 6: Add Makeup Gain

Bring the output level up until the perceived loudness matches the original signal or feels appropriate for the mix. A/B the compressed and uncompressed signal to ensure you haven’t introduced an unnatural change in tone.

Step 7: Listen in Context

Dialogue compression should be evaluated within the full mix – music, sound effects, and ambience. What sounds good in solo may not work when other elements compete for the same frequency range. Make small adjustments while the entire track plays.

Advanced Techniques for Superior Dialogue Control

Multiband Compression

Multiband compressors split the audio into frequency bands (e.g., low, mid, high) and apply independent compression to each. This is useful for treating dialogue problems that occur in specific ranges. For instance, a narrow band around 200–400 Hz can tame boxiness, while a high band above 5 kHz can control sibilance without affecting the rest of the vocal. iZotope’s guide to multiband compression explains practical applications for voice.

Sidechain Compression for De‑essing

Sidechain compression uses an external signal to trigger gain reduction. For de‑essing, you route the dialogue to a compressor’s sidechain input and insert an EQ that boosts the sibilant range (around 6–8 kHz). When sibilance exceeds the threshold, the compressor dips only those frequencies, leaving the rest of the speech untouched. This method is more precise than standard de‑esser plugins and can be used creatively to duck dialogue behind music (e.g., radio voice‑overs).

Serial Compression

Instead of one compressor doing heavy lifting, use two or three with low ratios (2:1, 1.5:1) for a total of 4–6 dB reduction. This technique distributes gain reduction across multiple stages, reducing artefacts and preserving natural dynamics. Serial compression is standard in broadcast and film dialogue mixing because it yields smooth loudness consistency without a “compressed” sound.

Common Mistakes and How to Avoid Them

  • Over‑compression: Applying too much gain reduction causes dialogue to sound lifeless and fatiguing. Aim for 3–6 dB of reduction per stage and use your ears to judge whether the vocal retains its energy.
  • Pumping and Breathing: This occurs when the release time is too fast, causing the compressor to audibly “open up” between words. Increase the release time until the effect disappears.
  • Ignoring the Loudness Standard: Broadcast and streaming platforms often adhere to loudness standards like ITU‑R BS.1770 (e.g., –23 LUFS for TV, –16 LUFS for podcasts). Compression alone does not guarantee compliance; use a loudness meter alongside DRC. EBU Tech 3341 provides guidelines for loudness metering.
  • Neglecting the Microphone: A poor recording cannot be fixed with compression. Ensure good mic technique, a quiet environment, and proper gain staging before reaching for a compressor.
  • Using Only One Type of Compressor: Different compressors excel at different tasks. A transparent VCA might be best for peak control, while an optical compressor can add warmth. Layer them wisely.

Conclusion

Dynamic Range Compression remains one of the most effective tools for managing dialogue levels. By understanding its parameters, selecting the right compressor, and applying a methodical workflow, you can achieve clear, consistent speech that remains natural and engaging. Advanced techniques like multiband compression, sidechain de‑essing, and serial compression offer further refinement for professional results. As with any audio processing, critical listening and practice are irreplaceable. Continue exploring resources from trusted engineers and manufacturers to deepen your expertise. For further reading, consult Sound on Sound’s compression basics or Production Expert’s guide to dialogue compression.