Mediatavio
YouTube to MP3PriceBlogPrivacy

Audio extraction

Video to Audio: Tracks, Formats and Sample Rates

Save the soundtrack in a useful format while understanding what extraction does not change.

By Mediatavio Engineering

Extract the soundtrack, not an isolated voice

Video to Audio saves an audio track from a video as a separate file. Speech, music and background sounds already mixed into that track remain mixed. This is not vocal separation, speech recognition or noise removal.

  1. Choose a video containing sound and inspect the preview.
  2. Select MP3, WAV, OGG, FLAC, AAC or M4A for the destination application.
  3. Choose a bitrate for a lossy format. Keep sample rate and channels at Source unless you have a specific delivery requirement.
  4. Sign in, export and download the audio result.
  5. Listen to the saved file and check its duration, channel balance and playback compatibility.

To save only a short interval, trim the video with Video Trimmer first or edit the extracted file with Audio Trimmer. This extraction page does not provide a timeline selection.

Choose a format for the next step

Audio encoders selected by the current implementation
OutputEncodingPractical consideration
MP3MP3 through FFmpegLossy audio for players that expect MP3
WAV16-bit PCMUncompressed samples; potentially much larger files
FLACFLAC through FFmpegLossless compression of the encoded PCM representation
AAC / M4AAACRaw AAC uses ADTS; M4A packages AAC in an MP4 container
OGGOpus natively; Vorbis through FFmpegThe same extension can contain different audio codecs

M4A in this tool means AAC, not Apple Lossless. WAV avoids another lossy codec but writes 16-bit samples; it does not preserve every possible higher-precision source representation. Neither WAV nor FLAC can reconstruct information already discarded by a source AAC or MP3 encoder.

Bitrate affects the lossy output encoders. WAV and FLAC do not use the selected lossy bitrate as a file-size target. Check that your player supports the actual OGG codec rather than relying only on the extension.

What sample rate and channels mean

Sample rate describes how many audio frames represent each second, not playback speed. Converting to another rate resamples the signal; it does not make the recording faster or restore missing high-frequency detail. Choosing more channels cannot recreate a stereo recording from mono.

Source leaves explicit rate and channel overrides unset. A requested value asks the processing engine to convert the audio, subject to codec support. Changing the number of channels can alter the mix, so listen before replacing a delivery file.

For a simple size illustration, stereo 16-bit PCM at 48 kHz needs approximately the following data rate before headers:

48,000 frames/s * 2 channels * 2 bytes = 192,000 bytes/s

This arithmetic is not a measured export result. Compressed file size depends on the encoder and the recording, and an output can still be larger than its source video.

Which track is saved

The FFmpeg path explicitly maps the first audio stream and disables video, subtitle and data output. It re-encodes that stream using the selected audio codec; this page does not offer original-audio stream copying.

-i input.media -map 0:a:0 -vn -sn -dn
-c:a libmp3lame -b:a 192k

This MP3 example shows relevant arguments, not a complete command. Optional sample-rate and channel arguments are added when selected. The native path selects the primary audio track through Mediabunny. A multi-track video is not mixed into one file, and there is no language-track selector; the primary native track and first FFmpeg track need not be the same in every container.

A silent video without an audio track cannot produce a soundtrack. The tool reports the missing track rather than creating invented sound. If a required native audio track cannot be processed, the page tries FFmpeg.

Native processing and the fallback

The native Worker discards video output and requests WAV PCM, AAC, M4A AAC or OGG Opus audio. It needs a secure context and a working source audio decoder; encoded outputs also need their corresponding encoder. MP3 and FLAC use FFmpeg, as do unsupported native combinations or native failures.

Both routes run locally, with account checks handled separately. The native implementation reads the source through a Blob source but buffers the finished output. FFmpeg uses a WASM filesystem, so extracting audio does not eliminate input-memory costs. Long WAV exports can be especially large.

Keep the tab active during export. Cancel terminates the active engine, and the original video remains unchanged. Successful file creation does not guarantee playback support in every browser: test the download in the application that will use it.