Convert MOV to MP3

This is the thing worth understanding before choosing a setting. A camera records audio so that somebody can hear what was happening — to sync a proper recording against it, to know which take was which, to follow the words. Even on a phone, the microphone is a few millimetres of hole pointing wherever the lens is pointing.

So the MP3 you make is a working copy of a working copy. That is not a criticism of it: a transcriber does not need a studio recording, an editor cutting to dialogue needs to hear the dialogue, and a speaker who wants their own talk wants the talk rather than the acoustics. Knowing that stops you spending 320 kilobits a second describing an air conditioner.

Every other conversion here uses codecs the operating system already exposes to the browser. MP3 is the exception in one direction: browsers decode it universally and encode it nowhere, because the encoder was patent-encumbered for most of the format’s life and nobody added one afterwards.

So an MP3 encoder is fetched the first time you convert to this format and cached afterwards. It happens before the progress bar moves, which is why the first conversion has a pause in it that the second does not. Nothing about your file is involved in that request — it is a piece of software arriving, not a recording leaving.

The control offers small, balanced and high quality rather than a number, and behind it the three land at roughly 96, 160 and 320 kilobits a second. That translates directly: an hour of audio is about 43 MB at 96, 72 MB at 160 and 144 MB at 320.

For camera sound, 96 is genuinely enough and 160 is the setting to pick if you do not want to think about it. The case for 320 is music — a rehearsal filmed on a phone, a live set, a recital — where the extra bits go on something that was actually captured. On speech recorded by an on-board microphone they mostly go on room tone.

The MOV’s audio is already compressed. A phone writes AAC, and most cameras write either AAC or uncompressed PCM. Decoding that and encoding MP3 spends a second helping of detail on top of whatever the first pass removed, and there is no way around it short of not converting.

On speech it is inaudible at any sensible band. Where it shows is on sustained musical content and on anything with a lot of high-frequency texture — cymbals, applause, rain, crowd noise — which acquire a faint swirling quality if you compress them twice at a low bitrate. If the recording is music you care about, WAV or FLAC keep what is there instead of compressing it again.

An MP3 file has room for exactly one audio track, so whatever the QuickTime container held beyond the first is discarded. That is fine for a phone and it is a real trap on a camera, where a second track often carries a radio microphone or a separate feed from a mixing desk.

Check one file before you commit to a batch. If what comes out is the on-board microphone and you wanted the lapel, the ordering is in the source file and cannot be changed here — the fix is to export the track you want from an editor, or to convert the separate audio recording directly instead of pulling it out of the video.

No tags are written. The file has no title, no artist, no date and no artwork, and a media player will show it as whatever the filename says. That is invisible while the file sits in a folder and very visible once it is on somebody’s phone among two hundred other recordings.

The cheapest fix is the filename. A date, a subject and a take number cost nothing to type and survive every transfer. If the MP3 is going into a podcast client or a music library, add proper tags afterwards with a tagging tool — the conversion is not the place to do it and nothing here pretends to.

The arithmetic is worth seeing once. A minute of 1080p video from a phone is around 65 MB; a minute of 160 kilobit MP3 is about 1.2 MB. The audio is roughly one fiftieth of the file and it is the entire content for anyone who is going to transcribe, index or listen to it.

That is why this conversion is so much more useful than sending the video. A transcription service charges by the minute either way and handles audio faster; a colleague on a train can play an MP3 and cannot stream a video; an archive of talks stored as audio is a fraction of the size and searchable once transcribed.

The limit applies to the file you hand over, which is the video, and video is enormous compared with what comes out. An iPhone recording at 4K passes 100 MB in about thirty-five seconds; at 1080p it takes about a minute and a half. The MP3 from either is a couple of megabytes.

For anything longer, the sensible route is the software that already has the footage. Every editor can export an audio-only file, and doing it there also lets you choose which track and trim the ends. This page is for the case where the video is what you have, on the machine you are sitting at, and you want the sound out of it now.

Reading the container, decoding the audio and encoding the MP3 all happen in this tab. Nothing carrying the recording is sent anywhere, and the network tab shows the encoder arriving and nothing going the other way.

That is the difference that matters for this material. Interviews, meetings, medical appointments, client sites and rehearsals are exactly the recordings people convert casually and would never deliberately publish. Getting an MP3 out of one should not involve handing the video to a stranger, and here it does not.