In professional audio post-production, achieving transparent noise gate settings for dialogue tracks is a critical skill that separates clean, natural-sounding audio from abrasive, distraction-filled sound. Whether you're mixing for film, television, podcasts, or voiceovers, the goal is to suppress unwanted background noise — hums, air conditioning, traffic, or microphone handling — without introducing audible artifacts or making the dialogue sound choppy or robotic. A properly configured noise gate can remain completely invisible to the listener, preserving the natural dynamics and breath of the performance while cleaning up the noise floor. This article provides an in-depth guide to understanding, setting, and fine-tuning noise gates for dialogue, delivering a professional result every time.

The Fundamentals of Noise Gating

A noise gate is a dynamics processor that attenuates audio signals below a defined threshold. When the signal level drops below that threshold, the gate "closes," reducing the gain and thus silencing low-level noise. When the signal exceeds the threshold, the gate "opens," allowing the audio to pass at full gain. While the concept sounds simple, the challenge lies in making the gating action imperceptible. For dialogue, any abrupt or mismatched timing can sound unnatural, clip transient sounds, or cause a pumping effect.

Unlike compressors, which reduce gain above a threshold, gates operate below the threshold. An expander is a variant that reduces gain gradually below the threshold, offering a smoother, more musical alternative to a hard gate. Many modern plugins combine gate and expander modes, as well as additional features like lookahead and sidechain filtering.

Core Parameters and Their Impact on Transparency

To achieve transparent noise gating, you must understand each parameter and how it interacts with the human voice. Below are the primary controls found on most noise gates.

Threshold

The threshold sets the level (in dB) at which the gate engages. It must be set just above the noise floor but below the quietest part of the dialogue. Setting it too high causes the gate to close on soft speech; setting it too low lets noise through. For dialogue, typical thresholds range from -40 dBFS to -20 dBFS depending on the recording level and noise floor. Use a spectrogram or a silent section of the track to identify the noise floor, then set the threshold about 2–3 dB above that level. Avoid setting it too close to the noise floor to prevent the gate from "chattering" open and closed on background fluctuations.

Attack Time

The attack determines how quickly the gate opens once the signal exceeds the threshold. For dialogue, a fast attack (1–10 ms) is essential to capture the initial transient of a word — especially plosive consonants like "p" and "b." If the attack is too slow, the start of words may sound truncated or muffled. However, an attack that is too fast can introduce a click or pop. Most engineers set attack between 3 ms and 8 ms for transparent results, adjusting based on the speaker's articulation.

Hold Time

Hold time keeps the gate open for a minimum duration after the signal falls below the threshold. This prevents the gate from slamming shut between syllables or during short pauses in speech. A hold time of 20–100 ms is typical for dialogue. Too short of a hold causes chattering on rapid speech; too long can let noise bleed through during brief silences. Listen to plosive and fricative transitions to fine-tune hold time.

Release Time

Release controls how quickly the gate closes after the signal drops below threshold and the hold time expires. This is arguably the most critical parameter for transparency. A release that is too fast produces an unnatural, abrupt cutoff of the room's natural reverb tail or breath sounds. A release that is too slow keeps the gate open, letting noise pass between phrases. For dialogue, start with a release of 50–200 ms. Faster releases (e.g., 50 ms) work well for dry close-mic'd vocals, while slower releases (150–200 ms) suit roomy dialogue or tracks with natural reverb. Always adjust release to match the decay of the background noise.

Range

Range (also called floor or depth) determines the maximum gain reduction when the gate is closed. A range of -60 dB is common for dialogue, effectively silencing the noise. However, using too much reduction can emphasize any gain-staging issues or bring up electrical hum during silent passages. Some engineers prefer a range of -20 dB to -40 dB to avoid an unnatural "dead air" feel, especially if the background noise is consistent (like HVAC hum). For a more natural sound, consider using an expander with a ratio of 2:1 to 4:1 instead of a hard gate.

Knee and Hysteresis

Many advanced gates offer a "knee" control that smooths the transition from open to closed, mimicking the behavior of an expander. A soft knee (6–12 dB) produces a gradual reduction, reducing the audible onset of gating. Hysteresis uses separate threshold levels for opening and closing: a higher threshold to open and a lower threshold to close. This creates a "dead zone" that prevents the gate from oscillating near the threshold. Both features are valuable for dialogue transparency.

Sidechain Filtering: Frequency-Conscious Noise Gating

One of the most effective ways to make a noise gate transparent is to use a sidechain filter. This allows the gate to only open when specific frequencies are present — typically the fundamental frequencies of human speech (e.g., 100 Hz–300 Hz for male, 200 Hz–500 Hz for female). By filtering out low-frequency rumbles or high-frequency hiss from the sidechain input, the gate will ignore noise that contains different frequency content, reducing false triggers. For example, if the background noise is dominated by a 60 Hz hum, you can high-pass filter the sidechain above 100 Hz so the hum doesn't cause the gate to open. This technique keeps the dialogue clean without the gate reacting to noise that is spectrally distinct from the voice.

Practical Workflow for Setting a Transparent Noise Gate

Follow this step-by-step approach when setting up a noise gate on a dialogue track:

  1. Analyze the noise floor: Locate a silent section of the recording (between phrases or after the performance ends). Note the RMS and peak levels of the noise. Listen to its character — is it broadband, hum, hiss, or room tone?
  2. Set the threshold: Start with the threshold around 3–6 dB above the noise floor. Play the dialogue and watch the gate's metering. If the gate closes on soft words, reduce threshold; if it stays open during silence, raise it.
  3. Adjust attack and hold: Set attack to 3–5 ms initially. Use a fast attack for plosive clarity but not so fast that it clips transients. Add enough hold (30–80 ms) to prevent chattering on quick syllables.
  4. Tune release time: Listen to the end of phrases. Adjust release so the gate closes just after the natural reverb or breath fades. If you hear a "chop" on the tail, increase release. If you hear noise between words, decrease release.
  5. Refine with knee or expander mode: If the gate sounds abrupt, soften the knee or switch to expander mode with a ratio of 3:1. This creates a more gradual noise reduction.
  6. Use sidechain filtering: Insert an equalizer in the sidechain path to filter out frequencies below 100 Hz or above 8 kHz. This prevents low-frequency rumble or high-frequency hiss from opening the gate.
  7. Listen in context: Always audition the gate against the full mix, not just in solo. Background noise may be masked by music or sound effects, allowing you to use a less aggressive gate.
  8. A/B with the dry signal: Toggle the gate on and off to ensure the differences are minimal. If you can hear the gate working, it's not transparent.

Advanced Techniques for Transparent Noise Gating

Automated Thresholds

For dialogue with widely varying levels (e.g., whispering and shouting), a static threshold cannot work well. Use automation to lower the threshold during quiet passages and raise it during loud sections. Some DAWs allow clip gain adjustment before the gate, effectively normalizing the input level so a single threshold works across the track.

Multi-Band Gating

Multi-band gates split the signal into frequency bands and apply independent gating to each. This is useful when noise is concentrated in certain frequencies. For instance, you could gate low-frequency rumble aggressively while leaving mid-range noise more open, preserving the air of the voice. Multi-band gating can be very transparent but requires careful balancing of band crossover frequencies.

Use a De-Noiser Before Gating

If the noise floor varies or contains complex spectral content (like wind or refrigerator hum), consider using a spectral de-noiser (e.g., iZotope RX, Waves NS1) to reduce noise broadly, then a gate to catch residual noise. This tandem approach yields cleaner results than gating alone. The de-noiser handles continuous broadband noise, while the gate deals with impulse noise or sudden increases in background level.

Ducking Sidechain Alternatives

In some scenarios, you may want to use a sidechain to trigger a gate from another source — for example, using a separate voice track to gate a room microphone. This is more about noise gating for dialogue recording than mixing, but it can produce extremely transparent results when the gate is keyed from the close mic.

Common Mistakes and How to Avoid Them

  • Pumping: Caused by release time that is too fast or too slow. Pumping is a rhythmic noise floor change that sounds like a "wow" effect. Fix by adjusting release and using a softer knee.
  • Chattering: Rapid opening and closing on noise fluctuations. Increase hold time, adjust hysteresis, or raise threshold slightly.
  • Truncated consonants: Attack too slow means plosives are cut off. Reduce attack time to under 5 ms.
  • "Dead air" between words: Range too high. Reduce range to -20 dB or -30 dB to leave some ambient noise, or use an expander instead of a gate.
  • Not listening in context: A gate that sounds transparent in solo may be obvious in the mix. Always check with music and effects playing.
  • Over-gating: Attempting to remove every bit of noise can destroy the natural feel. Allow some room tone or ambient noise to remain for a realistic soundstage.

Real-World Examples: Film vs. Podcast vs. Voiceover

The ideal noise gate settings differ by application. In film dialogue, the goal is often to leave natural room tone to maintain spatial continuity — a gate with a range of -20 dB and a slow release can work well. In podcasts, listeners expect highly intimate, clean audio, so a more aggressive gate with a range of -40 dB to -60 dB is common, often combined with a de-esser and compressor. Voiceover for commercials or narrations requires pristine clarity with no artifacts; these tracks benefit from expander mode with ratio 2:1 and a threshold just above the noise floor. Always tailor your approach to the medium.

Many excellent noise gate plugins exist. For a straightforward, easy-to-use gate, the stock plugins in your DAW (like Pro Tools' DigiRack Gate or Logic's Compressor in gate mode) provide solid results. For more control, consider Waves Vocal Rider (which automates level before gating), FabFilter Pro-G (excellent visual feedback and sidechain filtering), or iZotope RX Voice De-noise for spectral cleaning. For further reading, consult Sound on Sound's article on creative gate use and The Pro Audio Files' noise gate tutorial.

Conclusion

Transparent noise gating for dialogue tracks is not about eliminating noise at all costs — it's about controlling it in a way that feels natural and unobtrusive. By carefully adjusting threshold, attack, hold, release, range, and sidechain filtering, you can clean up recordings without sacrificing the human quality of the voice. Remember to always listen in context, avoid common pitfalls like chattering or pumping, and when in doubt, err on the side of a gentler gate or an expander. With practice, you'll be able to set a noise gate that is completely transparent, leaving only the intended performance to shine through. For further deep dives into dynamics processing, explore the links provided and continue refining your technique.