Convert MP3 to OGG

Ogg is a container rather than a codec, and over its life it has held Vorbis, Opus, FLAC and Speex. For most of that time the practical meaning of ".ogg" was Vorbis, which is why a great many guidelines, tutorials and forum answers use the two words interchangeably.

What this page writes is Opus. It is the current Xiph.Org codec, it is better than Vorbis at every bitrate anyone would use, and every current decoder handles it. If the policy you are following names Vorbis by name rather than naming Ogg, read it carefully before submitting — that distinction occasionally matters to an automated check.

Free-software projects avoided MP3 for a specific legal reason: the format was covered by patents, and distributing an encoder or a decoder meant licensing them. Ogg Vorbis was designed as the unencumbered alternative, and a generation of project policies, packaging guidelines and upload rules were written around that difference.

Those patents have since expired, and Fraunhofer ended its MP3 licensing programme in 2017. The legal ground for the rule is gone. Plenty of policies remain in force anyway — through inertia, through a preference for open specifications, or because nobody has revisited the wording — and following the rule you were given is still the right move. It is worth knowing that the reason is historical.

Some of the argument survives its legal half. Opus and the Ogg container are specified in public RFCs — 6716 and 3533 — implemented by reference code anyone can read, and not controlled by a licensing body that could change its terms. For a project that intends to be readable in thirty years, that is a real property rather than a slogan.

The technical case has also inverted since the policies were written. In the 1990s choosing Ogg meant accepting a compatibility cost for an ideological benefit. Today Opus is straightforwardly the better codec at any bitrate a contributor would use, so the free format is also the one that sounds better at the same size.

This conversion decodes the MP3 and encodes it again with a different codec, so it is a second round of lossy compression on audio that has already been through one. That is unavoidable when the MP3 is what you have, and it is worth a moment of checking whether it is.

Recordings contributed to public archives frequently exist somewhere in better shape — a WAV from the session, a lossless copy from the person who made it, a higher-bitrate version on the original site. Converting from any of those gives a first-generation Opus file with nothing compounded, and the difference survives into every copy anyone makes afterwards.

Opus is efficient enough that a band at or slightly above the source bitrate makes the second encode essentially invisible. Going substantially higher is wasted, because the ceiling on quality was fixed when the MP3 was made and no setting here can lift it.

Going lower is the one genuinely damaging option, and it is tempting for a submission with a size limit. Two rounds of compression at a low target compound, and speech survives it far better than music does. If a size limit is forcing the issue, converting a stereo spoken recording to mono is usually a bigger saving than dropping the bitrate, and costs less.

Ogg splits a stream into pages, each carrying a position marker, which is what allows a player to seek without reading from the beginning and to report an accurate duration before playback starts. It also carries a block of named text fields for describing the file.

That puts it roughly on par with what a tagged MP3 already offered, so nothing is gained or lost in principle. The practical difference is that the Ogg fields are simple and consistently implemented, where ID3 accumulated three decades of versions and conventions that not every tool agrees on.

For the destinations these files are usually going to — a wiki page, a repository, a documentation site, a desktop application — Ogg support is complete and has been for years. Every current browser decodes Opus natively, so an audio file on a web page plays for practically everyone.

Hardware is the exception and it is worth naming so that nobody is caught out later: car head units, some televisions, older portable players and various hi-fi streamers will not play it. That is a problem for someone converting a music library and no problem at all for a contributor uploading a file to a project.

Files submitted to public projects outlive the context they were submitted in. A converted Ogg looks the same whether it came from a studio master or a low-bitrate download, and the next contributor to reuse it has no way to tell — except by the same spectrogram inspection that identifies any transcode.

So put the provenance somewhere durable: the upload form, the file description, a text field in the stream. It costs a sentence and it prevents a re-encoded MP3 from quietly becoming the archival copy of something that has a better source still in existence.

Opus is native to every browser, so this conversion begins immediately with nothing fetched — unlike an MP3 target, which needs a WebAssembly LAME build downloaded first because no browser can encode MP3 at all.

A drop takes up to a hundred files at 100 MB each, which for MP3 sources is well over an hour of audio per file, and returns them together as a ZIP. Nothing is transmitted at any point, so a large batch costs processor time and no bandwidth.