Understanding Noise Gates in Detail

A noise gate is a dynamics processor that attenuates or completely mutes audio signals that fall below a set threshold. The core idea is simple: when your desired dialogue or vocal is present, the gate remains open, allowing the audio to pass through naturally. When the speaker pauses or stops talking, the gate closes, reducing the background noise that would otherwise be audible in the silence. Noise gates are widely used in post‑production for film, television, podcasting, and live sound to clean up dialogue tracks without affecting the natural cadence of speech.

Unlike a noise reduction plugin that continuously processes the entire signal, a gate operates in a binary manner—it either lets audio through or does not. This makes it particularly effective for removing low‑level, constant noises like air conditioning hum, computer fan noise, or distant traffic. However, using a gate effectively requires understanding its parameters and their interaction with the source material. A poorly set gate can sound robotic, chop off the ends of words, or create unsettling pumping effects.

Key Parameters of a Noise Gate

Threshold

The threshold determines the level at which the gate opens and closes. Set it just above the level of the background noise, so the gate closes during quiet sections but opens when dialogue begins. Use a meter that shows the RMS or peak level of the noise floor to set a precise threshold. If the threshold is too high, parts of quiet dialogue—such as the end of a whispered sentence—will be cut off. Too low, and the gate never closes, allowing noise to pass throughout.

Ratio

On many noise gates, the ratio controls how much the signal is reduced when the gate is closed. A ratio of 20:1 or higher effectively mutes the signal. Some gates allow lower ratios (like 4:1 or 2:1) to turn the processor into an expander, which attenuates noise more gently—useful when you want to preserve some ambience but still reduce noise. For most dialogue cleanup, a high ratio (full gating) works best.

Attack

Attack controls how quickly the gate opens once the signal exceeds the threshold. A fast attack (1‑5 ms) ensures that the beginning of a word is not clipped. However, if attack is too fast, it can cause a click or “explosive” sound on plosives. For dialogue, an attack time around 1–10 ms is typical, depending on the speaker’s articulation and the nature of the noise.

Hold

Hold time keeps the gate open for a set duration even after the signal drops below the threshold. This prevents the gate from chattering rapidly on short pauses or when the speaker takes a quick breath. A hold of 50–200 ms helps maintain natural continuity. Overly long hold times can leave noise audible during longer pauses, while too short a hold can cause the gate to close prematurely, chopping off the tail of a word.

Release

Release controls how quickly the gate closes after the signal falls below the threshold (and after the hold period ends). A smooth release (50–300 ms) allows the sound to fade naturally rather than slamming shut. If the release is too fast, you may hear a click or abrupt change in ambience. If too slow, the gate will stay open through parts of the pause, letting noise back in. Adjust the release so that the noise floor rises and falls imperceptibly.

Range

Range sets the maximum amount of attenuation when the gate is closed. A range of −60 dB to −80 dB will effectively mute the background noise. Some gates allow a range of only −20 dB, which is useful when you want to reduce noise without removing it completely—creating a more natural sound floor for certain styles of dialogue.

Advanced Techniques

Sidechain Filtering

Many noise gates offer a sidechain input that lets you key the gate from an external audio source. In dialogue editing, you can feed a copy of the dialogue track that has been filtered to emphasize the voice’s frequency range (e.g., 200 Hz–4 kHz). This makes the gate open only when the speaker’s voice is present, even if the original track contains loud background noises that might otherwise trigger the gate. To do this, insert a parametric EQ on the sidechain to boost the voice frequencies and cut the low rumble and high hiss.

Frequency‑Keyed Gating

Some advanced plugins allow you to set the gate to respond only to specific frequency bands. For example, you can make the gate open only when there is energy in the vocal range, while ignoring low‑end thumps or sibilance. This is extremely helpful when the background noise contains intermittent sounds (like a dog bark) that fall outside the speaker’s range. The gate remains closed during those noises, eliminating the need for manual editing.

Expanders vs. Gates

An expander is a dynamics processor that reduces the level of signals that fall below the threshold, but in a gradual way. The ratio is typically lower (e.g., 2:1). Expanders are excellent for cleaning up dialogue when you want to retain some ambient information—for instance, in a documentary where natural room tone is part of the storytelling. Use an expander when the background noise is not too loud and you want to avoid the unnatural silence that a full gate can create.

Practical Step‑by‑Step Workflow

Follow this procedure to apply a noise gate to a dialogue track in your digital audio workstation (DAW):

  1. Analyze the noise floor. Solo the track and find a section with no dialogue. Note the peak or RMS level of the background noise using a level meter. Set the initial threshold about 3–6 dB above that level.
  2. Set attack and release. Start with attack at 5 ms, hold at 100 ms, and release at 150 ms. Listen to a full sentence and adjust: if the beginning of words are cut, shorten attack; if the ambience bumps back in too quickly, lengthen release.
  3. Adjust the range. Set the range to at least −60 dB to fully mute the noise. If the background noise is not extreme, try a range of −30 dB to −40 dB for a more natural decay.
  4. Dial in the hold. Increase the hold time until the gate stays open through natural breaths and short pauses (typically 100–200 ms). Avoid excessive hold, or the gate will leave noise audible during longer pauses.
  5. Use sidechain filtering if needed. Insert a high‑pass filter (cutting below 100 Hz) and a low‑pass filter (cutting above 8 kHz) on the sidechain path. This prevents low‑frequency rumbles or high‑frequency hiss from falsely opening the gate.
  6. Listen in context. Play the dialogue against the background music and sound effects. Adjust the gate so that it doesn’t create distracting level changes or reveal unintended artifacts.

Common Mistakes and How to Avoid Them

Setting the Threshold Too High

One of the most frequent errors is setting the threshold too high, causing the gate to close during quiet speech—especially at the ends of sentences or on softer consonants. This results in choppy, unnatural dialogue. Always check the gate’s behavior on the softest spoken words in the track. If necessary, lower the threshold or increase the hold time to prevent premature closing.

Using a Release That Is Too Fast

A release time below 30 ms often produces a click or a “zipper” sound as the noise floor is abruptly reintroduced. This is especially noticeable during pauses. Increase the release time to at least 100 ms and listen for a smooth fade out of noise. If the release is too long, the gate may not fully close before the next phrase begins, letting noise bleed through.

Ignoring the Attack Phase

Dialogue that starts with a plosive (like “p” or “b”) can cause a thump if the attack is too fast. Conversely, a very slow attack will make the beginning of a word sound muted or “sucked in.” Find a middle ground—usually between 3 and 10 ms—by listening to the first syllable of each phrase.

Over‑Gating by Using Too High a Ratio

While a ratio of 20:1 or higher is common for dialogue, some situations call for a gentler approach. If the background noise is low and relatively consistent, a lower ratio (like 4:1) acting as an expander can provide a more natural result. Experiment with the ratio to see whether full gating or expansion sounds better for your specific track.

Not Using a Look‑Ahead Feature

Many modern noise gate plugins offer a look‑ahead setting (usually 1–10 ms). This allows the gate to open slightly before the signal crosses the threshold, preventing the attack of a word from being clipped. Enable look‑ahead if your DAW supports it, but be aware that it introduces a small amount of latency. For real‑time monitoring, you may need to compensate with delay.

Combining Noise Gates with Other Tools

A noise gate is most effective when used as part of a broader cleanup chain. Here are common ways to combine it with other processors:

  • EQ before the gate: Apply a high‑pass filter to remove low‑frequency rumble (below 80 Hz) and a low‑pass filter to reduce high‑frequency hiss (above 12 kHz). This reduces the energy that the gate “sees,” making it easier to set a clean threshold.
  • Compression after the gate: Once the background noise is removed, a compressor can even out the dynamic range of the dialogue. Insert a compressor after the gate with a low ratio (2:1 or 3:1) and a threshold that catches only the loudest peaks. This prevents the dialogue from sounding inconsistent.
  • Noise reduction plugins: For persistent noise that the gate cannot remove (like a constant 60‑Hz hum), use a standalone noise reduction plugin (e.g., iZotope RX, Waves WNS) before the gate. The gate will then handle only the residual noise.
  • De‑essing: Sibilance (harsh “s” sounds) can open the gate unexpectedly. Place a de‑esser before the gate to tame those frequencies, or use a frequency‑keyed gate that ignores the sibilant range (5–8 kHz).

Real‑World Application Scenarios

On‑Location Dialogue with HVAC Noise

Suppose you have a dialogue track recorded in a room with an air conditioning unit that adds a constant low‑frequency hum. Set the threshold just above the hum level, use a fast attack (5 ms) and a medium release (150 ms). Apply a high‑pass filter in the sidechain to prevent the hum from opening the gate. The result is clean dialogue with no hum during pauses, while the words remain natural.

Podcast with Traffic Background

For a podcast recorded near a street, the background is inconsistent—cars pass by at random intervals. A standard gate will open whenever the traffic gets loud, letting in unwanted noise. Use a frequency‑keyed gate that responds only to frequencies in the vocal range (200 Hz–3 kHz). Set the threshold low enough to capture quiet speech but high enough to ignore distant traffic. The gate will open only when the host speaks, ignoring the passing cars.

Documentary Narration with Room Tone

When a narrator records in a studio with some natural room tone, a full gate can create an unnatural dead silence between sentences. Instead, use an expander with a 2:1 ratio and a low threshold. The level of the room tone will be gently reduced by about 6 dB, preserving the sense of space while still cleaning up the track. This technique is also useful for maintaining the ambient realism in binaural or immersive audio formats.

Choosing the Right Noise Gate Plugin

Most DAWs include a stock noise gate that is perfectly capable for dialogue editing. However, third‑party plugins often offer more advanced features:

  • Waves NS1: Combines noise reduction and gating with a simple interface. Good for quick cleanup of dialogue with constant noise.
  • FabFilter Pro‑G: Offers a wide range of parameters, including an intelligent look‑ahead, smooth envelopes, and a flexible sidechain EQ. Well‑suited for precise editing.
  • iZotope RX Voice De‑noise: While not a traditional gate, it provides frequency‑dependent gating that is extremely effective for dialogue. Use it in conjunction with a simple gate for best results.
  • Soundtheory Gulldörfer: A specialized gate plugin that uses a neural network to detect speech. Ideal for noisy, chaotic backgrounds where a traditional gate would struggle.

External Resources for Further Learning

To deepen your understanding of noise gates and dialogue editing, explore these authoritative resources:

Final Thoughts

A noise gate, when applied with careful attention to its parameters, can transform a dialogue track from noisy and distracting to clear and professional. The key is to treat the gate as a surgical tool, not a blunt instrument. Spend time understanding the noise characteristics of your source material, set the threshold accurately, and always listen to the processed track in the context of the full mix. By combining a gate with EQ, compression, and other dynamics processing, you can achieve dialogue that sounds natural, intelligible, and free from unwanted background noise. With practice, you’ll develop an instinct for adjusting attack and release times to match different speaking styles and acoustic environments.