Audacity is a free, open-source, cross-platform audio editor. It's a multi-track recorder and waveform editor that handles everything from capturing a single voice memo to cleaning up podcast episodes, digitising old recordings, and assembling layered audio. It works on uncompressed audio for editing and exports to MP3, WAV, FLAC, OGG, and more. This guide covers macOS Tahoe (Apple Silicon), Fedora 43 KDE, and Windows 11.
Installation
macOS (Apple Silicon)
Download the .dmg from audacityteam.org/download . Recent Audacity builds are universal binaries that run natively on Apple Silicon — no Rosetta required. Open the .dmg , drag Audacity to Applications, and launch it.
Via Homebrew:
Note: On first launch, macOS may warn that Audacity is from an unidentified developer. Go to System Settings → Privacy & Security and click "Open Anyway". The first time you record, macOS will also prompt for microphone permission — grant it under System Settings → Privacy & Security → Microphone .
Fedora KDE
Install via Flatpak for the most up-to-date version:
Or via DNF (may be an older release):
Note: If you installed via Flatpak and Audacity can't see your microphone or other input devices, grant it audio access with Flatseal, or confirm that PipeWire/PulseAudio is exposing the device. In the Flatpak sandbox you may also need to select the correct host device from Audacity's device toolbar (see below).
— Placeholder — Audacity on first launch.
Windows 11
Download the Windows installer from audacityteam.org/download . Run the .exe installer — no administrator rights required. Alternatively, install via winget :
Note: Windows Defender SmartScreen may flag Audacity on first run. Click More info → Run anyway . The first time you record, Windows will prompt for microphone permission — grant it in Settings → Privacy & Security → Microphone and ensure the toggle for desktop apps is on. For the best low-latency performance, set the Audio Host in the device toolbar to Windows WASAPI rather than MME or DirectSound — WASAPI gives significantly better timing accuracy and lower latency, which matters for multitrack recording.
Interface Overview
Audacity's window is organised around a few key areas. Along the top are the Transport controls (play, stop, record, pause, skip) and the tool toolbars . Below those sits the device toolbar , where you choose your audio host, recording input (microphone), number of input channels, and playback output. The large central area is the track view , where each track's waveform is displayed on a horizontal timeline. The Timeline / ruler runs across the top of the track area showing time position.
The most important toolbars to know early on are the Tools toolbar (Selection, Envelope, Draw, and Multi-tool), the recording/playback meters (which show input and output levels), and the Selection toolbar at the bottom showing the exact start, end, and length of your current selection in hours/minutes/seconds.
— Placeholder — annotated Audacity interface with transport controls, device toolbar, meters, and track view.
Setting Your Input and Output Devices
Before recording anything, check the device toolbar . Set the Audio Host (on macOS this is Core Audio; on Fedora it's typically ALSA, PulseAudio, or JACK/PipeWire; on Windows 11 use WASAPI for best results), choose your recording device (your microphone or audio interface), set channels to Mono for a single mic or Stereo for a stereo source, and set your playback device (headphones or speakers). Using headphones while recording avoids your microphone picking up the playback and causing echo or feedback.
Recording Audio
First, set your levels. Speak or play at a normal volume and watch the recording meter at the top. Click the meter to enable monitoring without recording. Aim for peaks around -12 dB to -6 dB — loud enough to be clear, but with headroom so the loudest moments don't hit 0 dB and clip (distort). Adjust the input gain slider next to the microphone icon, or the gain on your audio interface, until levels sit in that range.
To record, press the red Record button (or R ). A new track is created and the waveform draws in real time as you record. Press Stop (or Space ) when finished. To add a new recording onto a new track press R again; to append more audio to the end of the current track, use Shift+Record ( Shift R ).
— Placeholder — recording in progress with levels peaking around -12 dB.
Punch-and-Roll and Re-recording
If you fluff a line while recording narration, you don't have to start over. Stop, place the cursor a few seconds before the mistake, and use Punch and Roll Record ( Shift D ) — Audacity plays a short lead-in, then drops into recording at the cursor so you can pick up seamlessly. This is the single biggest time-saver for spoken-word recording.
Basic Editing
Make sure the Selection tool (the I-beam, F1 ) is active. Click in the waveform to place the cursor, or click and drag to select a region. The Selection toolbar at the bottom shows the exact start, end, and length of your selection.
Cut, Copy, Paste, and Delete
Select a region, then: Cut ( Ctrl X / Cmd X ) removes it and closes the gap; Copy ( Ctrl C ) copies it; Paste ( Ctrl V ) inserts the clipboard at the cursor; Delete ( Delete ) removes the selection and closes the gap, exactly like Cut but without touching the clipboard.
Trimming
To keep only a portion of a recording and discard everything else, select the part you want to keep , then choose Edit → Remove Special → Trim Audio ( Ctrl T ). Everything outside the selection is removed. This is the inverse of Delete and is the fastest way to top-and-tail a clip — selecting the good take and trimming away the silence and false starts at either end.
Silencing a Selection
Sometimes you don't want to remove audio (which shifts everything after it) but simply want to mute a section in place — a cough, a click, a breath. Select the region and choose Edit → Remove Special → Silence Audio ( Ctrl L ). The waveform in that region is flattened to silence while keeping the overall timing intact.
— Placeholder — a noisy region selected and silenced in place.
Fades
To smooth the start or end of a clip, select a short region and apply Effect → Fading → Fade In or Fade Out . A one-second fade at the very start and end of a recording prevents the abrupt "click" of audio starting or stopping at full volume.
Undo
Audacity has unlimited undo within a session ( Ctrl Z ), and redo with Ctrl Y (macOS: Cmd Shift Z ). Edit freely — you can always step back.
Noise Reduction
Noise reduction removes consistent background noise — hum, hiss, fan or air-conditioner drone. It works in two passes: first you show Audacity a sample of just the noise, then you apply the reduction to the whole recording.
Step 1 — Get a noise profile. Find a stretch of your recording where no one is speaking but the background noise is present — even half a second works. This is why it helps to record a few seconds of "room tone" (silence) at the start of every session. Select that quiet region, then go to Effect → Noise Removal and Repair → Noise Reduction and click Get Noise Profile . The dialog closes; Audacity now knows what the noise sounds like.
Step 2 — Apply the reduction. Select the whole track ( Ctrl A ), open Noise Reduction again, and this time adjust the three settings and click OK :
Noise reduction (dB) — how much to cut. Start around 12 dB. Sensitivity — how aggressively Audacity decides what counts as noise; start around 6. Frequency smoothing (bands) — start around 3. Use the Preview button before committing. Push too hard and you'll get a watery, "underwater" artefact on the voice — back off the reduction amount until it sounds natural. It's better to leave a little noise than to mangle the voice.
— Placeholder — the Noise Reduction dialog after capturing a noise profile.
Splitting Tracks
"Splitting" can mean a few different things in Audacity, depending on what you're trying to do.
Splitting a Clip at a Point
To cut one clip into two independent clips on the same track (so you can move them separately), place the cursor where you want the split and choose Edit → Clip Boundaries → Split ( Ctrl I ). The clip is divided at that point with no audio removed. To split a selection out into its own clip with silence on either side, use Split New to move it to a brand-new track.
Splitting One Recording into Multiple Files
If you've recorded a long session — say, several songs from a vinyl record, or chapters of a reading — and want to export each segment as its own file, use labels . Place the cursor at the start of each segment and press Ctrl B to add a label; name it (this becomes the filename). When you've labelled every split point, choose File → Export → Export Multiple , set "Split files based on: Labels", and Audacity writes one file per label region in a single pass.
Splitting a Stereo Track into Two Mono Tracks
Click the track's name to open the track dropdown menu and choose Split Stereo to Mono . This gives you independent left and right channels — useful when each channel holds a different microphone, or when one channel is cleaner than the other.
— Placeholder — a session marked with labels, ready for Export Multiple.
Exporting Audio Files
Audacity's own project format is .aup3 — that's what File → Save Project writes, and it preserves all your tracks, labels, and edit history. But .aup3 won't play in a normal media player; to get a shareable audio file you need to export .
Go to File → Export and choose a format:
Export as WAV — uncompressed, lossless, large files. Best for archiving or handing audio to another editor. Export as MP3 — compressed, small, universally playable. Best for podcasts, sharing, and web. Set the quality to around 192–256 kbps for speech, higher for music. Export as FLAC — lossless but compressed (roughly half the size of WAV). Good for music archives. Export as OGG — open, compressed alternative to MP3.
In the export dialog you can fill in metadata (title, artist, track number) which players and podcast apps will display. On modern Audacity, MP3 export is built in — you no longer need to install the separate LAME encoder.
Remember the distinction: Save Project keeps your editable work in .aup3 ; Export produces the finished listenable file. Do both — save the project so you can come back and re-edit, and export the file you actually share.
— Placeholder — exporting to MP3 with metadata fields filled in.
Practical Workflows and Tips
A Clean Voice-Recording Workflow
A reliable order of operations for spoken-word recording: (1) set your input device and levels, aiming for -12 dB peaks; (2) record two or three seconds of silent room tone, then your content; (3) use Punch-and-Roll to fix flubs as you go; (4) when done, capture a noise profile from the room tone and apply Noise Reduction to the whole track; (5) trim the dead air off the start and end; (6) silence any remaining clicks, coughs, or long pauses; (7) apply a short fade in and fade out; (8) run Effect → Volume and Compression → Loudness Normalization to bring the whole thing to a consistent perceived loudness (around -16 LUFS for podcasts); (9) Save the project, then Export to MP3.
Normalise vs. Amplify vs. Loudness Normalization
Amplify raises or lowers the whole selection by a fixed amount. Normalize scales the audio so its peak hits a target level (e.g. -1 dB). Loudness Normalization targets perceived loudness (LUFS), which is what streaming and podcast platforms actually measure — use this one for final delivery so your audio sits at a comfortable, consistent volume.
Useful Habits
Always record room tone. A few seconds of silence at the top of every session gives you a noise profile and material to patch gaps. Save the project early and often ( Ctrl S ) — .aup3 projects keep all your undo history. Work non-destructively where you can — duplicate a track ( Ctrl D ) before applying heavy effects so you have the original to fall back on. Use Solo and Mute on track headers to audition individual tracks in a multi-track project. Zoom in ( Ctrl 1 ) to make sample-accurate edits and out ( Ctrl 3 ) to see the whole project.
Key Keyboard Shortcuts
macOS users: substitute Cmd for Ctrl throughout.
Tips and Gotchas
Save Project and Export are not the same thing. The most common beginner surprise is saving a .aup3 project and then being unable to play it in a music app. Always Export to get a standard audio file; Save Project to keep editing later.
Don't over-apply Noise Reduction. Aggressive reduction produces a tell-tale watery, robotic artefact on voices. Use the lowest amount that gets the job done, and always Preview first.
Record with headphones. Monitoring through speakers lets the microphone re-capture the playback, causing echo and, at worst, feedback squeal.
On macOS , if Audacity records silence, check System Settings → Privacy & Security → Microphone and confirm Audacity is allowed. You may need to quit and relaunch after granting permission.
On Fedora KDE , the Flatpak build runs in a sandbox. If your microphone or interface doesn't appear in the device toolbar, use Flatseal to confirm audio permissions, make sure PipeWire is running, and try selecting the PulseAudio host in Audacity's device toolbar. For low-latency multitrack work you can route Audacity through JACK/PipeWire.
On Windows 11 , if Audacity records silence even though microphone access is granted, check Sound settings → More sound settings → Recording tab — right-click your microphone and confirm it's set as Default Device. Also ensure the microphone isn't exclusively claimed by another app. If you're using an external USB audio interface, WASAPI Exclusive mode gives the lowest latency but prevents other apps from using the same device simultaneously — switch to WASAPI Shared if that's a problem.
Set the project sample rate before recording. The Project Rate (bottom-left) defaults to 44100 Hz, which is fine for most uses; use 48000 Hz if your audio will accompany video. Mixing sample rates between tracks forces resampling and can subtly degrade quality.