Extract a video’s audio as MP3 or WAV

Choose the file for its next job

MP3 is useful for a listening copy or a compact spoken-word delivery. WAV is useful when another editor needs decoded audio without another lossy compression step. WAV will generally be larger because it stores the samples rather than using the same kind of perceptual compression.

If the original soundtrack already lost detail during recording or encoding, a WAV export preserves the decoded version of that soundtrack. It does not restore the lost information. Calling that export “higher quality than the source” would be misleading.

For spoken material, check whether the source is mono, stereo or a more complex mix. Reducing it to mono may be reasonable for a single voice, but it can change music or spatial material. Know what the recipient asked for before changing channels.

Extract from the best available source

If you need only a passage, trim it deliberately. Do not use silence at the beginning of a listening preview as evidence that the file is empty. The source may contain a silent lead-in before the conversation begins.

1. Choose the original video rather than a copy downloaded from a messaging app. Additional compression in that copy may already have damaged the sound.
2. Open the MP3 extraction tool for a listening copy, or the audio extractor when you need another supported format. Choose the output settings the recipient needs.
3. Run the tool and read the result’s duration and audio information. A source with no audio should produce an explanatory error, not an empty file that looks successful.
4. Download the audio and play the beginning, a middle passage and the end. Then listen through the full deliverable before approving it.
5. If you will edit the sound, keep this extracted file alongside the video and identify which source it came from. Do not lose the link between the audio and its picture.

Check the extraction did not change the job

• Confirm the expected speaker and sound track were extracted, especially when the original has alternate languages or commentary.
• Listen for clipping, channel imbalance, dropouts and unexpectedly missing music. Compare a troublesome passage directly with the source.
• Check that the duration covers the intended content. Container timing and a player’s rounded display may differ slightly, so investigate a meaningful missing passage rather than a display rounding difference.
• Check seeking near the end. A download that begins correctly may still be incomplete.
• Open the file in the system where it will be used. An editor importing sound successfully is a different check from a browser showing a download link.

Keep extraction separate from cleanup

Extract first and assess what the recording needs. Noise reduction may help a steady background but may also damage speech. Loudness normalization changes the level over the programme; it does not make a clipped consonant undistorted.

Keep an untreated extraction for comparison. When cleanup makes the voice dull or watery, reduce the treatment or return to that copy. Do not keep processing merely because each new file has a more promising name.

For an audio-only publication, check the host’s delivery requirements. If it asks for a compact format, create that final copy from the best available intermediate rather than repeatedly converting between compressed formats.

Questions about this workflow