VEED Clean Audio
Remove background noise from any speech recording while keeping quiet words intact.
VEED Clean Audio: Remove Background Noise from Any Recording
Upload a noisy recording and get the voice back clean. VEED Clean Audio strips out traffic, fans, hum, and room noise while keeping quiet and distant words intact, the words most noise removers swallow.
It’s the same model behind the Clean Audio button in the VEED editor, where it handles thousands of recordings a day. On Mitte it takes any audio or video file up to 30 minutes long.
What can you do with it?
- Clean up a podcast episode. Recorded at a kitchen table with the fridge humming? The hum goes, the conversation stays.
- Rescue interview audio. Cafe chatter, street noise, and air conditioning come out, including under the soft-spoken guest.
- Fix the sound on a video. Upload the video itself and its audio track gets cleaned.
- Tidy voiceovers and course lessons. Home recordings come back sounding like they were made in a treated room.
- Prep speech for subtitles or dubbing. Less noise behind the voice means fewer mistakes in the steps that follow.
- Match levels across clips. Every output lands at the same loudness, so recordings from different days cut together without volume jumps.
How to use it
- Upload your recording. Audio (wav, mp3, m4a, aac, ogg) or video (mp4, mov, webm). For video, the audio track is used.
- Run it. A 10-minute file comes back in about 2 to 5 minutes.
- Get your clean track. Download it, or send it straight into your next step on Mitte.
What you get
One audio file, the same length as your recording, so it drops back onto your timeline in sync.
The output is 48 kHz mono in lossless FLAC (or WAV if your editor prefers it), leveled to -19 LUFS with peaks capped at -1.1 dBTP. That’s a standard loudness for spoken word, ready to publish.
Settings
- Strength: how hard the background is pushed down under the voice. The default is the tuned setting. Lower it to keep a little room tone behind the speaker. Silence between words is always fully cleaned.
- Loudness: the target level, from -40 to -8 LUFS. Leave it at -19 for most speech; go to -16 if you want a hotter podcast master.
- Output format: FLAC (lossless, about half the size) or WAV.
Tips
- Try the default strength first. It’s the setting VEED validated across real recordings, and it rarely needs touching.
- Too dry? Ease the strength down. A voice with zero room behind it can sound unnatural on camera. A bit of tone under the speech fixes that.
- It’s built for speech. For sung vocals and music, use the Vocal Splitter.
- Stereo comes back mono. Multi-channel recordings are mixed down to a single channel.
- Split sessions longer than 30 minutes. Files up to 512 MB are accepted.