Documentation

[ GUIDE ]

Guide

Start with the three-step quickstart, then dig into each tool when you need to tune it.

Start here

Quickstart Preparing a sequence

Per-tool reference

Full Pipeline (Wizard & Quick Run)Smart Switch Audio Export Silence Removal Filler Removal Voice Mixing

Your first edit in three steps

Install, prep, and run — the whole flow.

Once Creative Cloud finishes installing Multicam Toolbox, open Premiere Pro and find the panel under Window → Extensions → Multicam Toolbox. The first time you open it, paste your license key on the Activation required screen and click Activate — your key is in your purchase email. Dock the panel wherever you like. Then:

Prep your sequence

One audio track per speaker. Each camera on its own video track. Everything lined up in time.

Tip: use Premiere’s Synchronize to align by audio waveform.

Map speakers to cameras

Open the Wizard (Full pipeline setup) and walk the steps: confirm which tools run, name each speaker, and tick the cameras they appear on. A speaker can appear on more than one.

Tip: rename tracks in Premiere first — the panel auto-fills from them.

Hit Run.

Run executes the passes you enabled. Smart Switch always runs; add silence trimming, filler removal, or a voice mix by toggling those on. It backs up your sequence first, then applies everything in one pass.

Tip: next time, skip the Wizard — Quick Run replays your saved setup in one click.

That’s it for the happy path. If you want each tool on its own, or need to tune the details, jump to the per-tool reference below.

Preparing a sequence

What the panel expects to see on your timeline.

Multicam Toolbox reads your sequence directly — it doesn’t import footage or convert anything. The better your sequence is laid out, the better the results.

Audio tracks

One track per speaker. Lav mics, separate channels from a recorder, or individual mic inputs from a USB interface all work.
Rename each track to the speaker’s name in Premiere — the panel auto-fills from the track names.
A camera’s built-in scratch track is fine as a backup, but leave it unchecked under Speakers so it doesn’t drive switching.

Video tracks

Each camera angle on its own video track (V1, V2, V3…). No nested multicam clip — the panel switches by enabling and disabling clips on separate tracks.
Align cameras in time, usually by syncing to a common audio source or a clap.
Leaving gaps (dead air on a camera) is fine — the panel avoids cutting to a camera that has no footage at that moment.

Habit to build

Duplicate the sequence before running a tool for the first time. Run the pipeline on the copy, compare against the original, and iterate on settings. Once you’ve tuned the knobs for your recording style, you rarely need to touch them again.

Per-tool reference

Full Pipeline (Wizard & Quick Run)

Switch cameras, cut silences, and mix voices in one pass.

The full pipeline runs Smart Switch plus any of Silence Removal, Filler Removal, and Voice Mixing back-to-back on the same sequence. There are two ways in: the Wizard walks you through setup step by step, and Quick Run replays your saved mapping and settings presets in one click. Smart Switch is always on; the other three are opt-in.

When to use it. Start with the Wizard for interview- and podcast-style footage — it is the guided path for your first edit. Once your setup is saved as presets, Quick Run repeats it on the next episode without reopening the wizard.

Before you start

An active Premiere Pro sequence.
At least one video track and at least one audio track with clips.
One audio track per speaker, so the plugin can tell who is talking.

Steps

Open the panel via Window → Extensions → Multicam Toolbox. The status strip shows the active sequence name.
Click Wizard (Full pipeline setup) to open the guided flow. The stepper runs Tools → Speakers → Settings.
Tools: Smart Switch is locked on. Toggle on the extra passes you want — Silence removal, Filler removal (BETA), Voice mix (BETA). The execution-order pills below show the order they will run.
Speakers: name each audio track (e.g. "Host", "Guest 1") and toggle off any track that should not drive switching. Rows tagged AUTO were detected from your layout — override anything that looks wrong.
Speakers → Camera routing: for each camera (V1, V2…), tap the speaker pills for whoever appears on that angle. A speaker can be routed to more than one camera.
Settings: each active tool gets a size card (e.g. Cutting pace Relaxed/Balanced/Punchy). Pick a preset size, or open Advanced for the numeric controls.
If a "Timeline check — fix before running" banner appears, resolve the listed issues first — Run stays disabled until they clear.
Click Run. The plugin backs up your sequence, then a progress overlay runs each stage; the timeline updates when it finishes.
Optional: save your setup for next time. The panel keeps your speaker/camera mapping and your tool settings as two separate presets — a mapping preset and a settings preset — each with its own active slot. Quick Run replays whichever mapping and settings presets are active, together, straight from the home screen.

Settings worth knowing

Cutting pace (Smart Switch): Relaxed holds longer with fewer cuts; Balanced is the default for most shows; Punchy cuts more often. Open Advanced for the exact minimum/maximum camera times and the "ignore gaps shorter than" filter.
Tightness (Silence removal): Gentle removes only long dead air; Balanced tightens pauses without rushing; Aggressive cuts hard. Advanced exposes the duration, noise floor, and breathing-room margins.
Presets (Mapping & Settings): Two independent preset types: a mapping preset (which speaker sits on which camera) and a settings preset (your tool choices and their values). Each has its own active slot, and Quick Run replays both active presets together.
Voice mix: Opt-in denoise, compression, and volume leveling per speaker before the mix. Leave it off if you prefer to process audio elsewhere.

Tips

Run on a duplicate of your sequence the first time — it is easier to compare results and dial in settings.
If the panel warns the sequence "was already processed," duplicate the original before running again so you are not re-editing an edited timeline.
If silence cuts feel too tight, raise the "breathing room" values in Advanced before touching anything else.

Smart Switch

Multicam switching driven by who is talking.

Smart Switch watches each speaker-labeled audio track and switches to the camera assigned to whoever is currently speaking. Use it when you already have clean audio and just want the multicam cuts done for you.

When to use it. Pick this when you want switching only — no silence trimming, no audio processing. It uses the same Speakers and Settings screens as the wizard, minus the tool-selection step.

Before you start

An active sequence with at least one video and one audio track.
One audio track per speaker.
Clear mapping in your head of which speaker appears on which cameras.

Steps

Open the panel and click Smart Switch under "Or run one tool."
Speakers: name each audio track and toggle off any track that should not drive switching. AUTO-tagged rows were detected from your layout.
Camera routing: for each camera (V1, V2…), tap the speaker pills for whoever appears on that angle. A speaker can be routed to multiple cameras if the framing covers more than one person.
Settings: pick a Cutting pace (Relaxed / Balanced / Punchy), or open Advanced for the exact camera-time and gap controls.
Click Run. Your sequence is backed up first, then the switches are applied.

Settings worth knowing

Cutting pace: The one-tap control for how the edit feels — Relaxed holds longer, Punchy cuts more often. Balanced is the default.
Minimum / maximum camera time (Advanced): The shortest hold before another switch is allowed, and the longest a camera stays even if the same person keeps talking. Raise the minimum to calm a jittery cut; lower the maximum for a snappier feel.
Ignore gaps shorter than (Advanced): Brief speaker overlaps are ignored so the edit does not whip around when two people step on each other for half a second.

Tips

Mapping the same speaker to multiple cameras is the easy way to get multi-angle coverage — the plugin will pick between them.
If you also want silence trimming or a voice mix, use the Wizard instead — it runs Smart Switch plus the extra passes in one go.

Audio Export

Per-track WAV stems for editing outside Premiere.

Audio Export writes each selected audio track in the active sequence to its own WAV file. Use it when you want to hand off clean stems to a sound designer, mix externally, or archive raw speaker tracks.

When to use it. Use this whenever you need individual WAVs rather than a single mix. If you want a single mixed output, use Voice Mixing instead.

Before you start

An active sequence with at least one audio track that contains clips.
A writable folder on disk (the plugin defaults to your project folder).

Steps

Open the panel and click Audio Export.
Confirm the Export Path — it defaults to your project folder. Edit it if you want output somewhere else.
Set the Base Filename. Files are written as basename_1.wav, basename_2.wav, and so on.
Under Source Tracks, deselect any tracks you do not want exported. Empty tracks are skipped automatically.
Click ▶ Export Tracks. Status messages stream in the panel as each track is written.

Settings worth knowing

Export Path: Absolute path to the output folder. Must be writable. Pre-filled from the active project when possible.
Base Filename: Everything before the _N.wav suffix. Defaults to the sequence name.
Source Tracks: Checkboxes for each audio track. Output order matches the track index — top audio track becomes basename_1.wav.

Tips

Rename your sequence before running if you want the default base filename to be meaningful — it is faster than editing the field.
Empty tracks are auto-deselected but still show in the list — that is intentional, so you can see the full layout at a glance.

Silence Removal

Detect dead air and cut it out of selected audio tracks.

Silence Removal scans the audio tracks you pick, finds gaps of silence using FFmpeg, and ripple-deletes them automatically — the surrounding clips slide together so your timeline tightens up in one pass. It backs up your sequence before making any cuts, so you can always compare against the original.

When to use it. Use this when you want precise control over what "silence" means — threshold, minimum gap length, and margin — without running the full pipeline.

Before you start

An active sequence with audio to analyze.
FFmpeg (bundled with the plugin — no separate install needed).
Enough disk headroom for temporary analysis files in the project folder.

Steps

Open the panel and click Silence Removal under "Or run one tool."
Speakers: pick which audio tracks to scan — toggle off anything you want left untouched.
Settings: choose a Tightness (Gentle / Balanced / Aggressive), or open Advanced for the exact thresholds — each one has a short explainer right in the panel.
Click Run. Your sequence is backed up, then the plugin runs FFmpeg on each track and trims the dead air it finds.
Review the result on your timeline.

Settings worth knowing

Tightness: The one-tap control — Gentle removes only long pauses, Balanced is the safe default, Aggressive chases every gap. All three set the Advanced sliders below.
Shortest silence to cut (Advanced): Gaps shorter than this are left alone. Raise it to keep natural pauses, lower it to trim breaths and ums.
Shortest sound to keep as speech (Advanced): Sounds shorter than this are treated as noise and folded into silence. Helps avoid false positives from coughs, page turns, or mouse clicks.
Breathing room before / after speech (Advanced): Extra silence to keep on either side of the speech. Start with generous values; tighten them once you trust the detection.

Tips

Run on a duplicate sequence until the thresholds feel right — undo gets awkward if you ripple before you review.
If it is cutting off the start of words, increase "breathing room before speech". If sentences feel clipped at the end, increase "breathing room after speech".

Filler Removal

Cut um, uh, er, and ah with an on-device speech model.

Filler Removal (BETA) uses an on-device speech model to find filler words — um, uh, er, ah — and trims them out of the selected audio tracks. The model runs entirely on your machine: your audio, media, and timeline never leave your computer. (If you opt in to anonymous usage sharing in Settings, the panel sends anonymized detection metrics — counts and timing only, never your audio.) The first time you use it, the panel downloads the model, and Run stays disabled until it is ready.

When to use it. Use it to tighten conversational speech without hand-scrubbing for verbal tics. It runs standalone, or as one of the passes in the Wizard.

Before you start

An active sequence with speech audio to clean up.
A one-time model download on first use — a ~75 MB speech model plus a ~2 MB recognizer, fetched automatically with a progress bar in the panel.
One audio track per speaker gives the cleanest results.

Steps

Open the panel and click Filler Removal (BETA) under "Or run one tool." If the model is still downloading, wait for it to finish — the tile shows progress.
Speakers: toggle off any track you do not want scanned.
Settings: choose a Sensitivity (Light / Medium / Aggressive), or open Advanced for merge and padding controls.
Click Run. Your sequence is backed up first, then detected fillers are trimmed out.
Review the result. Prefer to check before cutting? Turn on "Mark instead of cut" in Advanced to drop markers on fillers instead of removing them.

Settings worth knowing

Sensitivity: Light catches only obvious, isolated ums (safest); Medium gets most fillers with very few false hits; Aggressive removes everything it can find — spot-check the result.
Merge fillers within / Keep padding around cut (Advanced): How close two fillers must be to be removed as one, and how much audio to leave on either side of each cut so words are not clipped.
Mark instead of cut (Advanced): Places a marker on each detected filler instead of removing it — useful for reviewing what the model found before committing.
Trim welded fillers (Advanced): Shortens a filler fused onto a real word instead of skipping it — helps with "and-uh" style run-ons.

Tips

The model download happens once and is cached — later runs start immediately.
Run on a duplicate sequence the first time so you can compare, and start on Light before moving to Aggressive.

Voice Mixing

Denoise, compress, normalize, and mix to a single WAV.

Voice Mixing analyzes each included track, applies denoise and compression, normalizes loudness, then mixes the result into a single WAV and drops it onto a new audio track at the start of the sequence.

When to use it. Use this when you want a finished voice mix laid back into the sequence — one track to rule them all — without hand-mixing each speaker yourself.

Before you start

An active sequence with at least one audio track.
FFmpeg (bundled).
Write access to the project folder for temporary and output files.

Steps

Open the panel and click Voice Mixing (BETA) under "Or run one tool."
Speakers: toggle off any track you do not want in the final mix.
Settings: pick a Polish level (Light / Standard / Heavy), or open Advanced for the Noise reduction and Compression controls.
Click Run. The plugin denoises, compresses, levels, and mixes the included tracks, then drops the finished mix on a new audio track at the start of the sequence.
Review the mix on your timeline. Re-run with a different Polish level if it needs more or less cleanup.

Settings worth knowing

Polish: The overall cleanup strength across denoise and leveling. Light keeps the room’s character; Standard is broadcast-style; Heavy is maximum cleanup but can sound processed.
Noise reduction (Advanced): Off, Light, Medium, or Heavy. Start with Light; go higher only if the room or mic is obviously noisy. Heavy can smear sibilance.
Compression (Advanced): Light preserves dynamics; Standard is a safe default; Heavy is for when speech needs to cut through music or background noise.

Tips

If you want individual WAVs instead of a mix, use Audio Export.
Voice Mixing is one of the passes the Wizard can run — enable Voice mix there to fold it into a full edit instead of running it on its own.

Get help

Stuck, found a bug, or want a feature? Our Discord is the fastest way to reach us — ask in #help, report problems in #bug-reports, and suggest ideas in #feature-requests. Prefer email? Reach us at support@multicamtoolbox.com.

Join the Discord

Get the plugin Have a question?