SoundForgePro Official 15-day trial

Audio Restoration

How to Use the Noise Gate in Sound Forge Pro

In this guideSections
    How to Use the Noise Gate in Sound Forge Pro

    A noise gate in Sound Forge Pro silences audio that falls below a threshold you set. When the signal is loud enough — a voice, an instrument, a transient — the gate opens and lets it through. When the signal drops below the threshold — silence between sentences, the noise floor between notes — the gate closes and attenuates or mutes it. The result is a recording where the quiet sections are actually quiet instead of quietly noisy.

    Part of the Sound Forge restoration hub. New to the editor? Start with the complete Sound Forge 2026 guide.

    Quick answer: Open Effects → Noise Gate for the legacy processor, or load coreFX Gate from the included coreFX Suite when it is available in your plug-in list. Measure a noise-only pause, place the threshold above that noise but below the quietest word or note you must preserve, use a fast attack, and lengthen release until endings stop sounding chopped. Preview the weakest passages before you render.

    Two Gate Plugins in Sound Forge Pro

    Sound Forge can expose a built-in Noise Gate and the coreFX Gate plug-in. They share the same basic job, but they are not parameter-for-parameter copies. The current Sound Forge pricing page lists the coreFX Suite as included; the Sound Forge Pro 12 manual documents the built-in Effects → Noise Gate command and the Statistics workflow used to measure noisy silence.

    Noise Gate (Effects → Noise Gate) — the built-in processor documented in the Sound Forge manuals. It removes signal below a threshold and provides timing controls for how the gate opens and closes. Treat the menu label and available presets in your installed build as the authority; legacy installs and upgraded systems can expose a different plug-in inventory.

    coreFX Gate — part of the coreFX Suite listed with the current Sound Forge plan. Its documented gate controls are Level (the threshold) and Attack/Release. Use it in a previewable effect chain when you want to compare settings before committing the result. Don’t assume that a control shown in a third-party gate tutorial — Hold, Hysteresis, Range or Floor — exists in these Sound Forge gates.

    Sound Forge 2026 noise gate workflow measuring room noise setting threshold and checking attack and release
    Sound Forge 2026 gate workflow: measure a noise-only pause, keep the threshold below quiet words, then preview attack and release before applying.

    Choose the processor whose controls you can verify in your installed build, then judge it on the actual file. The useful comparison is not which interface looks newer; it is whether the quietest wanted syllables and decays pass cleanly while the unwanted pauses close without an obvious jump in room tone.

    Threshold, Attack and Release

    Threshold — the level in dB below which the gate closes. Any signal quieter than the threshold gets attenuated. Any signal louder than the threshold passes through. Setting this correctly is the whole job — everything else is refinement.

    Set the threshold in the gap between the highest unwanted noise and the lowest wanted signal. Use Tools → Statistics on a noise-only selection to measure the noise peak, then preview the quietest wanted syllable or note. If noise peaks near -42 dBFS and the softest wanted speech reaches -28 dBFS, a threshold around -38 to -35 dBFS is a sensible test range. If wanted speech falls to -30 dBFS, the threshold must stay below that level or the gate will cut it; raising the threshold above quiet program material is not a cleanup technique.

    Attack — how quickly the gate opens once the signal crosses the threshold, in milliseconds. Too slow an attack clips the beginning of transients — the first few milliseconds of a consonant, a drum hit, a note onset — because the gate hasn't fully opened yet. Too fast an attack can cause the gate to snap open abruptly on signals that approach the threshold gradually.

    For dialogue, start with a short attack and audition words that begin quietly, especially sibilants and soft consonants. A 1–5 ms test range is useful, not a preset to copy blindly. Percussive material may need the fastest opening the processor allows; a slower attack deliberately softens an onset. Increase it only when a very fast opening clicks or sounds unnaturally abrupt.

    Release — how quickly the gate closes after the signal drops below the threshold, in milliseconds. Too fast a release creates an abrupt cutoff that sounds like the audio was edited with scissors. Too slow a release lets noise through after the program material ends, defeating the purpose of the gate.

    For dialogue, 100–200 ms is a practical first preview range, not a universal target. Listen to sentence endings, breaths and room reverb. If tails are chopped, lengthen release or lower the threshold. If noise remains audible for too long after every phrase, shorten release carefully. Always check several quiet and loud passages because one setting can behave differently across the file.

    Closed-gate behavior — the built-in Noise Gate and the documented coreFX Gate do not expose the adjustable Floor/Range control described in many generic gate tutorials. A gate suppresses the low-level signal when closed. If that produces unnatural dead silence, use a downward expander or manual volume envelope for partial attenuation, or preserve controlled room tone between edits.

    On spoken-word and location recordings, complete silence can be more distracting than the original room tone. When that happens, do not look for a nonexistent Floor knob: reduce the amount of gating, use an expander, or edit consistent room tone under the pauses. The goal is a stable background, not the lowest possible meter reading.

    How to Set the Threshold Correctly

    Start by running Tools → Statistics on a section that contains only the background noise. Note its maximum sample level; the manual explicitly identifies this as a way to estimate noise amplitude for the Noise Gate. Then find the lowest wanted word or note. Place the threshold between those two readings and preview. Two to four decibels above a stable noise peak can be a useful first test, but only if it remains safely below the wanted signal.

    If you don't have a clean noise-only section, preview the file with the gate active and watch the level meters. The gate should be visibly closed during pauses — the meters should drop noticeably — and open cleanly when the program material starts. If the meters don't drop during pauses, the threshold is set too low. If speech gets cut off, it's set too high.

    A threshold that is too high commonly removes soft consonants, breaths and phrase endings. Lower it while looping the weakest passage, then recheck a noise-only pause. If there is no clean level gap between wanted signal and noise, a gate cannot separate them reliably; use broadband or spectral noise reduction instead.

    A common mistake is setting the threshold based on how much noise sounds audible rather than on the actual level difference between noise and program material. If the noise floor is at -42 dBFS and the quietest speech is at -28 dBFS, there's 14 dB of separation to work with. Setting the threshold at -35 dBFS gives 7 dB of margin on each side — enough for the gate to open reliably on speech without triggering on noise.

    Chattering: What It Is and How to Stop It

    Gate chattering is the rapid opening and closing of the gate when the signal hovers near the threshold. It sounds like a stuttering or pumping effect on a quiet passage — the gate triggers open and shut repeatedly as the signal fluctuates just above and below the threshold level. It's the most common artifact from a noise gate, and it almost always means one of two things: the threshold is set too close to the level of the program material, or the release time is too short.

    Moving the threshold down is usually the first thing to try. If the signal regularly dips below the threshold during breaths or brief silences between words, the gate triggers closed on those dips and opens again immediately. Setting the threshold lower so those dips don't cross it stops the repeated triggering.

    The two Sound Forge gates documented here do not expose a Hold control. If lowering the threshold is not enough, lengthen Release so brief dips do not become abrupt closures. If the source still chatters, the signal-to-noise gap may be too small for a hard gate; use an expander or repair the underlying noise instead.

    The ProSoundWeb noise-gate guide explains the same timing trade-off in more detail: release must be long enough to avoid an audible closure during brief dips, but short enough that unwanted room tone does not linger between phrases.

    Don’t raise the threshold when quiet endings are already falling below it; that makes the cutoff worse. Lower the threshold, lengthen Release, or abandon the gate if noise and speech overlap. Community reports about spoken-word gating repeatedly point to the same audible failure: the noise vanishes in pauses and rushes back under the voice.

    Noise Gate vs Noise Reduction: Different Tools for Different Problems

    A noise gate and the NR-2.0 noise reduction tool do not do the same thing, and reaching for the wrong one wastes time.

    A gate is useful when unwanted low-level sound is mainly objectionable in pauses and there is a clear level gap between that sound and the wanted material. It does not identify "noise" by frequency or character. When the gate opens for speech or music, every sound above and below it in the recording passes through together.

    Noise Reduction handles the other problem: constant noise that runs underneath the program material as well as in the gaps. HVAC hum, tape hiss, computer fan noise, mic preamp hiss. These problems don't have clean pauses the gate can use as a reference — the noise is always present, whether the speaker is talking or not. The full NR workflow is in the noise reduction guide.

    A quick diagnosis is to listen while the wanted signal is active. If the fan, hiss or hum remains distracting under speech, gating the pauses will only increase the contrast. Capture a representative noise print and preview conservative broadband reduction instead; use DeHum or narrow EQ for a tonal hum when appropriate.

    A recording-community discussion of vocal gating describes this failure mode plainly: a hard gate can make dialogue sound less natural because the background disappears in pauses and returns under speech. Treat that as a listening warning, not a universal ban. A light gate can work when the recording already has a healthy signal-to-noise ratio; severe background noise needs a different tool.

    The two tools can be used together. Run NR-2.0 first to reduce constant noise across the whole file, then apply the gate to clean up any remaining noise in the quiet sections. This combination often produces cleaner results than either tool alone, especially on location recordings where there's both a constant noise floor and quieter gaps between program material. That sequence fits into the broader signal chain covered in the mastering guide.

    Applying the Gate: Selection vs Whole File

    Sound Forge Pro applies the gate destructively when run from the Effects menu directly. Before running it, decide whether you want to gate the entire file or just specific sections.

    Ctrl+A to select all and apply the gate to the whole file. Or drag a selection over specific sections — particularly useful when only part of a recording has the noise problem, or when different sections need different threshold settings.

    When the noise floor changes across a file, split the work into sections and preview each one. Don’t use a more aggressive threshold for wind simply because it is louder: wind often overlaps speech across a wide frequency range, so a hard gate may expose the problem rather than solve it.

    For files where the noise floor varies — a recording that gets quieter in some sections and louder in others — applying different gate settings to different selections produces better results than a single pass. Select the noisier sections, apply a more aggressive threshold. Select the quieter sections, apply a more conservative threshold. Two passes, two selections, one file.

    Filter before the gate only when unwanted energy sits outside the useful range of the source — for example, subsonic handling noise below speech. A filter that removes the same frequencies from both noise and voice may not improve their level separation. Use the Paragraphic EQ to remove a verified band, then remeasure and reset the gate threshold. Don’t carry the old threshold forward after EQ changes the level.

    Frequently Asked Questions

    Where is the Noise Gate in Sound Forge Pro?

    Use Effects → Noise Gate for the built-in processor. The current Sound Forge plan also lists the coreFX Suite; load coreFX Gate from the plug-in list if it is present in your installation. Upgraded and legacy installs can expose different plug-in inventories, so verify the exact menu in your build.

    What threshold should I set for the noise gate in Sound Forge Pro?

    Measure a noise-only selection with Tools → Statistics, then measure or preview the quietest wanted signal. Put the threshold between them. Two to four decibels above a stable noise peak is a useful starting test only when it stays below every syllable or note you need to preserve.

    What attack and release times should I use for a noise gate in Sound Forge Pro?

    For dialogue, preview a short attack around 1–5 ms and release around 100–200 ms, then tune by ear. Shorten attack if the start of words is lost. Lengthen release if phrase endings or room decay are chopped. These are test ranges, not universal settings.

    Why does my noise gate sound choppy or stuttering in Sound Forge Pro?

    The signal is hovering around the threshold. Lower the threshold and lengthen Release while looping the weakest wanted passage. The documented Sound Forge gates do not expose Hold. If speech and noise occupy the same level range, use an expander or noise-reduction process instead of forcing a hard gate.

    What's the difference between the Noise Gate and Noise Reduction in Sound Forge Pro?

    The Noise Gate attenuates audio during quiet sections where signal falls below a threshold. It only helps when the noise is isolated to quiet gaps in the recording — between sentences, between notes. The Noise Reduction tool (Tools → Noise Reduction) uses a noise print to reduce constant background noise that runs underneath program material. Use the gate for intermittent noise in pauses; use NR for constant noise like hiss, hum, or fan noise. The full workflow for NR is in the noise reduction guide.

    Does Sound Forge's gate have a Floor or Range control?

    The built-in Noise Gate and the documented coreFX Gate do not provide the adjustable Floor or Range control found in some third-party gates. For partial attenuation instead of hard silence, use a downward expander, a volume envelope, or controlled room tone.

    Can I use the Noise Gate and Noise Reduction together in Sound Forge Pro?

    Yes. Apply conservative broadband noise reduction first when noise remains under the wanted signal, then remeasure the residual noise and set the gate for pauses. Preview both stages together and bypass each one separately; stacking aggressive processing can create pumping or watery artifacts.

    If noise remains audible under speech, move to the broadband noise-reduction workflow instead of forcing a higher gate threshold. Use Spectrum Analysis or the hiss-and-hum diagnosis guide when the unwanted sound is tonal or frequency-specific.

    Last fact-checked: August 2, 2026 against the current Sound Forge 2026 plan, Sound Forge Noise Gate and Statistics documentation, coreFX Gate controls, and cited gate-engineering guidance.