M4A to MP3 converter

What an M4A file actually is

An M4A is an MPEG-4 file, the same container family as an .mp4 video, holding audio and nothing else. The .m4a extension was popularised by Apple through iTunes to mark those audio-only files, and it is still what the iPhone, the Mac and Apple’s music store use for audio.

The container is not the codec. Inside an M4A is usually AAC, Advanced Audio Coding, the lossy codec designed as MP3’s successor, which generally sounds better than MP3 at the same bitrate. Less often it is Apple Lossless (ALAC), which is compressed without discarding anything. Nothing in the file name tells you which of the two you have, and that turns out to matter, as the voice memo section below explains.

Why convert M4A to MP3 at all

Compatibility, and nothing else. MP3 has been the lowest common denominator of digital audio for decades: older car stereos and portable players, upload forms that list one accepted format, other people’s editing software and the odd piece of office equipment will all take an MP3 when they will not take an M4A. Modern phones, computers and browsers play M4A without complaint, so the problem is almost always at the other end: a device or a service you do not control.

That cuts both ways. If everything you use already plays your M4A files, keep them as they are. AAC is the more efficient codec, and converting costs a little quality for no gain. Convert when something specific refuses the M4A, and convert once.

Converting iPhone voice memos to MP3

Voice Memos saves every recording as an .m4a. What is inside depends on the Audio Quality option under Voice Memos in the iPhone’s Settings app: Compressed, the default, records AAC; Lossless records Apple Lossless into a file with exactly the same extension.

To get a memo off the phone, open it in Voice Memos, use Share, and either save it to Files, AirDrop it to a Mac, or send it to yourself. On a Mac you can drag a recording out of the Voice Memos app into a Finder window, which leaves a plain .m4a file behind. Then drop any number of them onto the converter above.

One catch is worth knowing before you start. MDN’s browser codec guide lists Safari as the only major browser that supports ALAC, and in our own testing Chrome refused an Apple Lossless .m4a outright. So a memo recorded on the Lossless setting will fail in Chrome, Edge and Firefox with a decoding error on its row. Open this page in Safari and try the same file there. A compressed memo converts in Chrome, Edge, Firefox and Safari alike.

For speech, 128 kbps and the Mono option are plenty. Most voice memos are one person talking, so a second channel adds size rather than information, and Mono suits them. A bitrate far above the memo’s own makes a larger file, not a clearer one.

iTunes purchases, Apple Music and DRM

Since 2009 the iTunes Store has sold music without DRM, as “iTunes Plus” 256 kbps AAC in an ordinary .m4a. Those files are yours to convert, and they work here like any other M4A, with the caveat about lossy-to-lossy conversion below.

Two kinds of Apple file will not convert. Songs bought before the store dropped DRM, and never replaced since, can still be .m4p files, the p standing for protected, locked with Apple’s FairPlay system. Tracks downloaded for offline listening through an Apple Music subscription are protected too. Browsers cannot decode copy-protected audio, so these fail on their row with a message that says so. This page does not try to get round the protection.

Choosing an MP3 bitrate for speech and for music

The encoder writes constant bitrate MP3, so the size is simple arithmetic: bitrate times duration. A minute costs just under 1 MB at 128 kbps and a little over 2 MB at 320 kbps, and every row shows its estimate before you press convert.

• 128 kbps for speech: voice memos, lectures, interviews, meetings. Choose Mono as well for a single speaker: one voice needs one channel, and a mono MP3 spends its whole bitrate on it.
• 192 kbps, the default, is a sound choice for general music and mixed content, and the one to pick when you do not know what is in the batch.
• 256 kbps for music you care about, and the natural match for iTunes Plus purchases, which are 256 kbps AAC to begin with.
• 320 kbps is the highest bitrate the MP3 format allows. It is worth using only when the source is lossless or of a similar quality.

Because AAC is the more efficient codec, an MP3 at the same bitrate as its AAC source will sound slightly worse, never better. Matching the source or going one step above it is the sensible range. Past that point the extra bits only describe the existing compression artefacts in more detail.

What converting lossy to lossy costs

An AAC encoder has already thrown away the parts of the sound it judged you would not hear. To make an MP3, the file is decoded back to plain samples and handed to the MP3 encoder, which runs its own model of hearing and throws material away again, making different decisions and leaving different artefacts. That is generational loss. One generation at a sensible bitrate is usually hard to hear on speech. Several (M4A to MP3, MP3 back to M4A for a phone, then to MP3 again for an editor) add up to the smeared cymbals and the watery, swirling top end that give heavily transcoded audio away.

The practical rules follow from that:

• Keep the original M4A. The MP3 is a copy made for compatibility, not a replacement.
• Convert from the best source you have. A lossless memo or a WAV beats an AAC file of the same recording, and the WAV to MP3 converter exists for exactly that case.
• Do not re-encode an MP3 to “fix” it. The converter warns on the row when you add a file that is already an MP3.

Converting a batch of M4A files

Drop several files at once, or select several in the file picker. Each gets a row showing its size, its duration and roughly how big the MP3 will be. The bitrate and channel settings apply to the whole batch; press Convert all and the files are converted one at a time, each with its own progress bar.

One at a time is deliberate. Decoding turns compressed audio into 32-bit floating-point samples, so an hour of stereo takes more than a gigabyte of memory however small the file is on disk. Holding one decoded file at a time keeps a long batch from taking the tab down with it. For the same reason files over 200 MB are refused, and a row warns you when its file is over 50 MB or long enough to need more than a gigabyte.

Length has a limit of its own, whatever the size on disk. The decoded samples are copied to the encoder’s worker in one piece, and Chrome refuses a copy much over 2 GB. At 48 kHz, a common rate for audio hardware, that is about 85 minutes of stereo or about 170 minutes of mono, and a little more at 44.1 kHz. A recording over the limit gets a message on its row saying so. For a long stereo file, choosing Mono halves the amount, since the channels are mixed before the copy; past about 170 minutes, split the recording into parts first.

• Cancel stops the file being encoded at once, since the encoder’s worker is shut down rather than asked nicely, and leaves the rest of the list queued. Convert all picks up where you stopped.
• Change your mind about the bitrate after converting, and the rows made at the old setting say so. Convert all then redoes only those.
• Download on a row saves that MP3 under its original name with .mp3 on the end, tidied so it is safe to save: accents are dropped, spaces and punctuation become hyphens, and letters outside the Latin alphabet are left out. Download all saves every finished file separately rather than as a zip, and Chrome may ask once whether the page may download several files. Two inputs with the same name get -2, -3 and so on, so nothing is overwritten.

What happens to your files

Each file is read off your disk by the browser’s file API and decoded by the browser’s own audio decoder, the one it uses to play media. The samples are passed to a Web Worker running LAME, the long-established open-source MP3 encoder, compiled to WebAssembly. The finished MP3 is held in the tab’s memory, and the Download button points at that copy.

No part of that involves a server. The only network request the converter makes is for the encoder’s own code, fetched from this site the first time you press Convert all. Your audio never leaves the tab, and closing the tab discards everything. You can check this for yourself: open your browser’s network panel, convert a file, and look for an upload. There is none.

What this converter does not do

• It writes no tags. Titles, artist and album names and cover art stored in the M4A are not copied into ID3 tags, so a music player shows the MP3 by file name alone. A tag editor can add them back afterwards.
• It does not remove DRM. Protected .m4p files and Apple Music downloads fail to decode, and stay that way.
• It converts audio only. An .mp4 video decodes, but only its audio track becomes the MP3. A 5.1 soundtrack comes out as stereo, or mono if you choose it, mixed with the standard speaker downmix so the centre channel, where film dialogue usually sits, is kept. For browser recordings in WEBM, use the WEBM to MP3 converter.
• It does not choose the sample rate. The Web Audio API decodes at the rate of your sound hardware, usually 44.1 or 48 kHz, and the MP3 is written at that rate. It changes nothing you can hear.
• It does not edit. To cut a recording down first, use the audio trimmer; to combine several into one, the audio joiner.

Where VoiceSnap Pro fits

People convert voice memos to MP3 for all sorts of reasons, and a common one is getting at the words: sending a recording to a transcription service, or pasting its gist into an email. If the words are the point, the audio to text tool transcribes a file on your own device, and dictation removes the recording from the process altogether.

That is what we are building. VoiceSnap Pro is a voice-to-text dictation app for macOS and Windows: hold one keyboard shortcut, speak, and clean punctuated text appears in whatever field your cursor is already in, whether that is an email, a Slack message or a code editor. Every dictation is also saved to a searchable notes library, so the thought you would have put in a voice memo is text you can find again. It will be a one-time $39 purchase, not a subscription.

It has not shipped yet, so there is nothing to download today. Join the waitlist for one email on release day. Until then, the voice typing tool shows what dictated text reads like, in your browser.