Pulling the audio out of a video is one of the most common conversions there is: a recorded lecture you want to listen to on a walk, a song from a video you filmed at a concert, the audio of an interview for transcription, or a voiceover you need to edit separately. This page opens with MP3 as the output; drop MP4, M4V or MOV files and convert, and you get the soundtrack only, with the video discarded. The audio in most MP4 files is AAC, which is already compressed, so the MP3 is a second encoding. Phone videos usually carry 128 to 256 kbps audio, and 192 kbps is a good default; for speech 128 kbps is enough and keeps files small. The video itself does not need to be re-encoded, but the converter still has to read the whole file, so a long video takes a while and the tab must stay open. Because everything happens in browser memory, files up to 200 MB are supported, which covers most phone clips but not long high resolution recordings. Only use this on videos you have the right to use. For WebM recordings see WebM to MP3 Converter, to make the video itself smaller use Video Compressor, and for a written transcript use Speech to Text Audio Transcriber.
The split that matters most in audio formats is lossy against lossless. A lossless format such as WAV or FLAC stores enough information to reconstruct the original waveform exactly, which is why WAV files run around 10MB per minute of stereo CD quality audio. Lossy formats such as MP3, OGG Vorbis, AAC, and Opus use psychoacoustic models to discard sound that listeners are unlikely to perceive, typically hitting a tenth of the size for audio most people cannot distinguish in normal listening. The important consequence is that the discarded information is gone permanently. Converting MP3 to WAV inflates the file back to full size without recovering anything, and each further lossy conversion compounds the loss, an effect known as generation loss. The practical rule is to keep a lossless master and export lossy copies from it, rather than converting between lossy formats repeatedly. Bitrate is where quality and size are actually traded. At 128 kbps most listeners cannot reliably distinguish MP3 from the source on laptop speakers or earbuds, though the gap becomes audible on good headphones with dense music. At 256 to 320 kbps MP3 is effectively transparent for nearly everyone. OGG Vorbis reaches similar quality at a somewhat lower bitrate, because it is a newer design than MP3, and Opus does better still, which is why messaging apps and voice chat use it. Sample rate and bit depth apply to the uncompressed side. This tool writes WAV at 16 bit and 44.1 kHz, the CD standard, which is the format editing software and mastering workflows expect, and which every audio application on every platform can open without conversion.