Audio workflow
Use this workflow to get clearer speech while minimizing processing artifacts.
- Back to: Guides
1) Identify the recording type
- Speech-only (podcast, interview, narration)
- Speech + background ambience (events, outdoor)
- Multi-speaker / overlapping dialogue
2) Start with the cleanest source
- Prefer WAV or original recorder output when available
- If extracting from video, export audio with minimal additional compression
3) Choose the right primary tool
- Denoise: best for steady noise (hiss, hum, fan, HVAC)
- Voice isolation: best for complex noise (crowd, traffic, mixed sources)
4) Refine conservatively
- Run a conservative pass.
- Listen for artifacts (warble, metallic tone, missing consonants).
- Increase intensity only if the voice remains natural.
5) Check intelligibility
- Listen on headphones and phone speakers
- Confirm key consonants (t/k/s/f) remain clear
6) Leveling (basic)
- Ensure the voice stays consistently audible across the whole clip.
7) Export appropriately
- Use a speech-appropriate bitrate/format for your platform.
Related pages
- Use cases: Audio use cases, Privacy workflows
- Examples: Remove background noise, Isolate a voice
- Guides: Troubleshooting, Export settings