Audio Production on Linux

What You'll Learn

Linux is a serious platform for audio. Professional DAWs such as Ardour, Reaper, and Bitwig Studio run natively, and modern Linux audio (PipeWire) gives low latency without the fiddly setup of the past. This page covers three kinds of work:

• Voice-over — narration for YouTube videos, podcasts, and courses
• Recording instruments — guitar, bass, vocals, and keys through an audio interface
• Music production — MIDI, virtual instruments, mixing, and mastering

How Linux Audio Works (in One Minute)

Your apps (DAW, browser, OBS, Discord)
          │
          ▼
     PipeWire  ← the modern sound server on Fedora, Ubuntu 22.10+, Mint 22+, Arch…
          │      speaks all three "languages" apps use:
          │      PulseAudio (desktop apps), JACK (pro audio), ALSA (low level)
          ▼
   ALSA drivers ──► your sound card / USB audio interface

Older guides tell you to install and start a separate JACK server. With PipeWire you don't need to: pro audio apps that want JACK are automatically connected through PipeWire, and your browser and system sounds keep working at the same time.

# Check you're running PipeWire (should mention "PipeWire")
pactl info | grep "Server Name"

Step 1: Set Up for Low Latency

Latency is the delay between playing a note and hearing it back. Under ~10 ms feels instant; above ~20 ms is hard to play along to.

1a. Real-time permissions

# Ubuntu/Debian: JACK support through PipeWire (often already installed)
sudo apt install pipewire-jack

# Real-time limits for the "audio" group: installing jackd2 asks
# "Enable realtime process priority?" — answer YES (it creates the settings file)
sudo apt install jackd2
sudo usermod -aG audio $USER         # then log out and back in

# Fedora: PipeWire's JACK support is normally installed already;
# real-time priority is granted automatically through RTKit
sudo dnf install pipewire-jack-audio-connection-kit

1b. Set a smaller buffer while recording

# Buffer ("quantum") of 128 samples at 48 kHz ≈ 2.7 ms per buffer
pw-metadata -n settings 0 clock.force-rate 48000
pw-metadata -n settings 0 clock.force-quantum 128

# Put it back to automatic when you're done
pw-metadata -n settings 0 clock.force-quantum 0

If you hear clicks or crackles, raise the quantum to 256. Tools like qpwgraph (org.rncbc.qpwgraph) show and change audio connections visually.

Step 2: Audio Interfaces and Microphones

• USB audio interfaces: any class-compliant interface works with no drivers: Focusrite Scarlett, PreSonus, Behringer UMC, MOTU M-series, Audient iD, and most others. Search "<model> class compliant" before buying.
• Focusrite Scarlett (3rd/4th gen): the alsa-scarlett-gui app replaces Focusrite Control for routing and the internal mixer.
• USB microphones (Blue Yeti, Rode NT-USB, Shure MV7): plug-and-play.
• Avoid interfaces that need a Windows-only control app for basic operation. Thunderbolt-only interfaces are hit and miss.

Set the interface as your input/output in Settings → Sound, or pick it inside the DAW. Choose the Pro Audio profile for an interface to see all its individual inputs and outputs.

Step 3: Choose Your DAW

Our picks: Audacity for voice-over; Ardour (free) or REAPER (paid, cheap) for recording and mixing; Bitwig if you make electronic music and come from Ableton.

Step 4: Plugins (Effects and Instruments)

Linux plugins come as LV2, CLAP, and VST3. Many commercial developers (u-he, Toneboosters, Airwindows, Vital) ship Linux versions — check before buying.

Windows-only VST plugins

yabridge runs Windows VST2/VST3/CLAP plugins inside Linux DAWs using Wine. Many work well; heavily copy-protected ones (iLok) often don't. Install the plugin's Windows installer with Wine, then run yabridgectl add "$HOME/.wine/drive_c/Program Files/Common Files/VST3" and yabridgectl sync.

Workflow 1: Voice-Over for YouTube and Podcasts (Audacity)

1. Room first: record in a small room with soft furnishings; get close to the mic (a fist's width to a hand's width away).
2. Set levels: in Audacity, click the recording meter to start monitoring. Speak normally — peaks should hit around −12 dB, never 0.
3. Record 5 seconds of silence first (room tone), then your script. Leave a pause and re-read any line you fumble; cut the mistakes later.
4. Clean up, in this order:
   • Noise Reduction: select the silence → Effect → Noise Removal and Repair → Noise Reduction → Get Noise Profile. Then select all → Noise Reduction again → around 6–12 dB → OK.
   • Filter Curve EQ or Bass and Treble: a low cut (high-pass) at ~80 Hz removes rumble.
   • Compressor: threshold −18 dB, ratio 3:1 evens out loud and quiet words.
   • Loudness Normalization: −14 LUFS for YouTube, −16 LUFS for podcasts.
5. Export: File → Export → WAV (for the video editor) or MP3 (for podcasts).

Workflow 2: Recording Guitar, Bass, and Vocals (Ardour)

1. Plug the guitar into the interface's Hi-Z / Instrument input, the mic into a mic input (turn on 48 V phantom power for condenser mics).
2. Open Ardour → New Session → Audio/MIDI system: JACK/PipeWire, sample rate 48 kHz, buffer 128 or 256.
3. Track → Add Track: one mono audio track per input. Connect each track's input to the right interface input (click the input button at the top of the track).
4. Guitar tone: on the guitar track, add a Guitarix or Neural Amp Modeler plugin. You record the clean signal but hear the amp — you can change the tone later.
5. Click track: set the tempo in the ruler and enable the metronome (click icon in the transport bar).
6. Arm the tracks (red button), press Shift+R to arm the transport, then Space to record.
7. Mix: balance faders first, then EQ (cut before you boost), then compression, then reverb on a bus shared by all tracks.
8. Export: Session → Export → Export to Audio File(s). WAV 24-bit for mastering; use the loudness analysis to target −14 LUFS for streaming.

Workflow 3: Producing Music With MIDI

1. Plug in a USB MIDI keyboard — it's detected automatically.
2. In your DAW add a MIDI track with an instrument plugin: Surge XT or Vital for synths, sfizz with a piano SFZ library for keys, Hydrogen or DrumGizmo for drums.
3. Record or draw notes in the piano roll. Quantize to tighten timing.
4. Build the arrangement: intro, verse, chorus. Duplicate regions and vary them.
5. Mix with the same order as above, then master: gentle EQ, glue compression, a limiter, and a loudness check.

Troubleshooting

• Crackles / dropouts — raise the buffer (quantum) to 256 or 512, close browsers during recording, and make sure your user is in the audio/realtime group.
• DAW doesn't see the interface's inputs — select the Pro Audio profile for the interface (Settings → Sound, or pavucontrol → Configuration).
• Hearing yourself twice / echo — turn off either the interface's direct monitoring or the DAW's input monitoring.
• Flatpak DAW can't find plugins — Flatpak DAWs use Flatpak plugin packages (search Flathub for "LV2" / "VST3"), or install the DAW and plugins from the same source.

What's Next?

• Put your voice-over on video: video editing on Linux →
• Stream your finished tracks from your own music server →