Video to MP3 Converter & Audio Extractor

The built-in sample. On a plain visit the tool makes a 6-second stereo test recording at 44,100 Hz: 6 × 44,100 = 264,600 samples per channel. As 16-bit WAV that is 44 header bytes + 264,600 × 2 channels × 2 bytes = 1,058,444 bytes. As MP3 at 192 kbps it is about 192,000 × 6 ÷ 8 = 144,000 bytes, roughly 14% of the WAV.

Trimming. To keep 1:30 to 2:00 of a soundtrack recorded at 48,000 Hz, the tool starts at sample 90 × 48,000 = 4,320,000 and stops at 120 × 48,000 = 5,760,000, keeping 1,440,000 samples, exactly 30 seconds. At 192 kbps that is about 720,000 bytes.

The common mistake: a higher bitrate than the source. Most phone and web videos carry AAC audio at about 128 kbps. Converting a 3-minute soundtrack to MP3 at 320 kbps gives 7.2 MB instead of 2.9 MB, 2.5 times bigger, and it cannot sound better than the 128 kbps it came from; it is a second lossy encode, so it is very slightly worse. Choose Original to copy the AAC stream untouched, or 128–192 kbps MP3 if you need MP3.

MP3 was developed largely at the Fraunhofer Institute in Germany, with Karlheinz Brandenburg as one of its main authors, and standardised as MPEG-1 Audio Layer III (ISO/IEC 11172-3) in 1993; Fraunhofer picked the .mp3 file extension in July 1995. The LAME encoder started in 1998 and became the reference-quality open-source MP3 encoder; this page uses lamejs, a JavaScript port of it. Fraunhofer ended its MP3 licensing programme in April 2017, and the format is now patent-free.

WAV comes from Microsoft and IBM’s Resource Interchange File Format (RIFF) of 1991: a short header followed by raw PCM samples, which is why its size can be worked out exactly. Video files rarely carry MP3 or WAV today. MP4 and MOV almost always hold AAC, standardised in 1997 as MPEG-2 Part 7; WebM holds Opus, standardised as RFC 6716 in 2012, or the older Vorbis. Copying those streams as they are is the only lossless way to get a video’s sound out.

This tool extracts the audio from a video, or re-saves an audio file, as MP3, lossless WAV or the original stream. It replaces the separate Audio Extractor, whose decode-to-WAV with a preview player is now the WAV option. It reads the file’s structure first, so you can see each stream before choosing, and it keeps the source sample rate instead of resampling to your sound card’s.

Everything runs in your browser: common formats use the browser’s built-in decoders and a small MP3 encoder; MKV, AVI, unusual codecs and the stream-copy option use a WebAssembly build of ffmpeg. Your video is never uploaded. It does not download from YouTube or any other site; use it on recordings you own or have permission to use.

It is for podcasters, students, video editors and anyone who wants just the sound of a clip.

Related tools: GIF from Video, Audio Muter, and Video Speed.