MP3 to AAC: how to convert it, and when you shouldn't

An MP3 encoder works by deciding which parts of the signal you are unlikely to hear and deleting them. By the time you have an MP3, that decision is permanent. Feeding it to an AAC encoder means a second encoder now makes its own set of cuts — on audio that has already been thinned out, and using a different psychoacoustic model, so it discards different things. This is generation loss, and it is why a 192 kbps AAC made from a 192 kbps MP3 sounds worse than either a 192 kbps MP3 or a 192 kbps AAC made from the CD.

The damage is usually mild on a single pass and mostly shows up in the places lossy codecs already struggle: cymbals, applause, reverb tails, sibilance in vocals. If you are converting a podcast for a car stereo, you will probably never notice. If you are converting a music library you intend to keep for a decade, you are permanently downgrading your masters — do it on copies, and keep the MP3s.

One thing raising the bitrate does not do is repair anything. Encoding a 128 kbps MP3 to 320 kbps AAC gives the second encoder plenty of room and produces a file that is faithful to the 128 kbps MP3, missing detail included, at two and a half times the size.

There are real reasons to do it, and they are all about compatibility or size:

• Something only accepts AAC or .m4a. iPhone ringtones (.m4r) must be AAC. Some video editors, DAWs and streaming pipelines want AAC in an MP4 container and reject MP3 outright.
• You are shrinking files for speech. AAC beats MP3 clearly below 128 kbps, so re-encoding a 128 kbps MP3 lecture recording to 64 kbps AAC halves the size with less harm than a 64 kbps MP3 would do.
• Consistency in a library. If everything else is AAC and a handful of MP3s break your tagging, artwork or gapless playback, converting them can be worth a small quality hit.

What is not a good reason: “AAC is a better codec.” It is, at the same bitrate, from the same source. That advantage is spent at encoding time and cannot be claimed retroactively on a file that is already MP3. The AAC vs MP3 comparison spells out where the gap is real and where it has closed.

For batches, scripts, or anything you would rather not click through, ffmpeg is the standard answer and runs on Windows, macOS and Linux:

ffmpeg -i input.mp3 -c:a aac -b:a 192k output.m4a

To convert a whole folder in one go, loop over it in bash or zsh:

for f in *.mp3; do ffmpeg -i "$f" -c:a aac -b:a 192k "${f%.mp3}.m4a"; done

Add -vn if the MP3s carry embedded cover art that trips the muxer up, and note that the .m4a extension is what tells FFmpeg to use an MP4 container; naming the output .aac gives you a raw ADTS stream instead.

On macOS you already have Apple's encoder available without installing anything, through afconvert:

afconvert -f m4af -d aac -b 192000 input.mp3 output.m4a

“MP3 to AAC encoder” is a fair thing to search for, because AAC encoders differ far more than MP3 encoders do. In rough order of quality at typical music bitrates:

1. Apple AAC — via Music, iTunes or afconvert. The reference point for AAC-LC quality, and free on every Mac and on Windows with iTunes installed.
2. libfdk_aac — Fraunhofer's encoder, excellent, but its licence means it is compiled out of nearly every FFmpeg binary you can download. You need a custom build to use it.
3. FFmpeg's native aac — mediocre years ago, solid since roughly 2016. At 192 kbps and up it is a perfectly reasonable choice, and it is the one you will actually have.
4. VLC and generic GUI converters — convenient (Media → Convert/Save → an Audio - MP4 profile), but the encoder behind them is usually older and the bitrate controls are coarser.

Avoid HE-AAC and HE-AAC v2 profiles unless you are targeting very low bitrates for speech. They are efficient below about 64 kbps and counterproductive above it, and some older hardware decodes them badly or not at all.

The rule for any lossy-to-lossy transcode is to match the source or go a little above it, never below. Some practical pairings:

If the encoder offers variable bitrate, use it. VBR spends bits where the music is dense and saves them during quiet passages, which matters more on a transcode than on a first encode.

Plenty of people land on “MP3 to AAC” when the problem is the reverse: a file downloaded from iTunes, Apple Music, YouTube or a voice recorder that a car stereo, DJ deck or old MP3 player refuses to open. Those files are AAC — usually with a .m4a, .m4b or .aac extension — and the fix is to go the other way.

That direction this site does handle, in the browser and without an upload. Drop the files on the AAC to MP3 converter and they are decoded and re-encoded on your own device at 128 to 320 kbps, 20 files per batch. If your files are specifically .m4a, the M4A to MP3 page covers that case, iTunes and Voice Memos quirks included.

One honest caveat about direction: browsers can decode AAC natively but cannot encode it, so no client-side web tool converts MP3 to AAC without either a server or a multi-megabyte encoder download. That is why the recommendations above are desktop tools rather than a link to a page here.