Interviews are recorded wherever the person happens to be: a busy café, a street corner, an office with the air conditioning on, a video call with a guest whose laptop sits next to a window. You rarely get to choose the room, so the cleanup afterwards matters more than for almost any other kind of recording.
This guide covers sit-down interviews indoors, street and vox-pop interviews outdoors, and remote interviews, with the steps that make the biggest difference to each.
Start by listening for three things
Before you change anything, play the recording on headphones and note:
- The background. Is it steady (air conditioning, a fridge, distant traffic) or changing (chatter, cutlery, passing cars, music)?
- The balance. Is the guest much quieter than the interviewer, or the other way round? Single-microphone interviews often favour whoever was closest.
- The damage. Any moments where the voice distorts, a gust of wind hits the microphone, or someone talks over the answer?
Background noise is what a noise remover is for. Balance is a levelling problem. Damage can only be partly repaired, so knowing where it is helps you plan the edit.
Sit-down interviews indoors
Indoor interviews usually suffer from steady room noise and echo.
- Clean before you edit. Remove the background from the raw file first, so every cut, fade and caption is built on clean audio.
- Clean each microphone separately if you recorded on two tracks, then balance them. Two clean tracks mixed together sound far better than one mixed track cleaned afterwards.
- Expect echo to stay. A large, hard room makes voices sound distant. That is the room, not noise on top of the voice, and noise removal doesn't take it away. Next time, sit closer to the microphones or move to a softer room.
Street interviews and vox pops
Outdoor interviews have the hardest noise: traffic, crowds, wind, sirens, music from shops. They also tend to be recorded on a single handheld or camera-mounted microphone.
- Use the original camera or phone file, not a copy shared through a messaging app.
- Remove the background in one pass. A speech-focused cleaner handles changing noise much better than classic noise reduction, which needs a steady noise profile and struggles with a passing bus.
- Watch the edges of each answer. If a car horn lands in the middle of a key sentence, the cleaned version will be much quieter, but a trace may remain; choose a different take for the edit if you have one.
- Accept some atmosphere. A little street sound under a vox pop tells viewers where they are. After cleaning, if the result feels too isolated, you can mix in a little clean ambience from a pause.
- Wind: see how to remove wind noise from video and audio, and use a furry windscreen next time.
Remote interviews
Remote interviews recorded over a call combine two problems: each person's room, and what the call software did to the audio (compression, its own noise suppression, dropouts).
- Ask the guest to record their side locally if you can, and send you the file. It will be far better than the call recording.
- If you only have the call recording, clean it as one file and focus on intelligibility rather than perfection. Why your Zoom recording sounds bad explains what can and can't be fixed.
Step-by-step cleanup
- Pick the best source for each voice: separate tracks, local recordings or the camera's original file.
- Remove the noise. Drop the file into the noise remover for audio, or for a filmed interview the video noise remover, which leaves the picture as it is and replaces only the audio. Our speech AI keeps the voices and removes the sound around them.
- Turn on Voice clarity if voices sound boxy or muddy; it trims rumble and boxiness and adds presence.
- Preview the original against the cleaned version at the same moment, especially in the noisiest section and on the quieter speaker.
- Set loudness for the destination: Podcast for an audio show, YouTube for online video.
- Download, then edit: balance the voices, cut the pauses, add music if you use it.
Balancing two voices
If one speaker is still much quieter after cleaning, raise their track or the sections where they speak. Do this after noise removal, so you are raising a clean voice rather than a voice plus its background. Gentle compression helps even out a speaker who varies a lot, but don't overdo it; a natural conversation has some dynamics.
Things that can't be fixed
Be honest with yourself and your editor about these:
- Crosstalk. When two people talk at once, both voices stay. A speech tool keeps speech, including the speech you didn't want.
- Music under the conversation. Background music in a café is reduced, but a speech tool is not a music separator; it may leave traces, and it won't keep the music cleanly if you wanted it.
- Heavy distortion. Our tool softens light clipping before it cleans, but a voice that broke up badly stays damaged.
- A whisper under a loud crowd. If you can't make out the words on headphones, no tool can recover them reliably.
A field checklist for next time
- Microphone as close as the situation allows; a clip-on for sit-downs, a handheld held near the chin for the street.
- Windscreen on outdoors, always.
- Record a few seconds of the location before the first question.
- Monitor on headphones while recording.
- For remote guests, ask for a local recording.
- Keep the originals untouched.
If you publish interviews as a show, podcast audio cleanup has presets for podcast loudness. Longer or frequent interviews may need more hours than the free allowance; the pricing page has the plans.