Module fingerprint

The acoustic fingerprint of a file: what the audio is, not what it says.

Why this exists

Every other identifier in this program comes out of the tags, and a badly tagged file therefore cannot be identified at all: track03.mp3 with no title and no artist is invisible to MusicBrainz, because there is nothing to ask about. A fingerprint is computed from the decoded audio, so it answers for a file whose tags say nothing, and it disagrees with a file whose tags say the wrong thing.

It is computed, not fetched — and stored beside the checksum

A fingerprint is Aède’s own conclusion about the bytes, like crate::audit::integrity and unlike anything in sources.json: nobody told us, we worked it out. So it lives on crate::model::AudioFile and is carried across a rescan by the same size-and-mtime rule the integrity verdict uses. What AcoustID answers when shown one is a different thing entirely, and goes in the attributed layer where every other source’s claim goes.

Two ways to compute one, and why ffmpeg comes first

Chromaprint is the algorithm; there are two ordinary ways to run it.

• ffmpeg, when it was built with --enable-chromaprint. Debian and Ubuntu build it that way, so on those machines this is no new dependency at all — Aède already requires ffmpeg for spectrum and copy --compress.
• fpcalc, the tool Chromaprint itself ships. Canonical, and the one AcoustID’s own documentation names.

Homebrew’s ffmpeg does not have chromaprint, which this module’s first version claimed it did — a guess written as a fact, corrected by a reader pasting the configuration: line of a fresh brew install ffmpeg. On macOS the answer is brew install chromaprint, and it is a small package. The lesson is not about ffmpeg: a build option is a property of somebody else’s packaging decision, and this program has no business asserting one it has not seen. So the availability check asks the muxer list rather than reading a version string, and the message names both ways out with the right one first for each platform.

The two agree on the whole input, not just a preview. fpcalc normally processes only the first 120 seconds; -length 0 removes that limit, matching ffmpeg’s complete-input path. On a short file its ordinary default agrees with ffmpeg -algorithm 1, byte for byte. fpcalc -algorithm 1 is a different algorithm and answers something else — the measurement is beside the ALGORITHM constant below.

ffmpeg is tried first because it is more likely to be there already; the fallback exists because “built with chromaprint” is not a promise anyone made. Neither is linked or vendored, exactly as ffmpeg is not: a checkout with neither installed builds, passes its tests, and says plainly what is missing.

The duration is part of the question

AcoustID is asked for a fingerprint and a length, and refuses without both — the length is how it tells a full track from an excerpt that fingerprints alike. fpcalc prints it; ffmpeg’s muxer does not, so the caller supplies it from the catalog, which read it out of the file’s own header at scan time.

Structs

Fingerprint

What one file’s audio amounts to, for a service that identifies audio.

Enums

By

Which program produced a fingerprint, for a message to name.

Functions

find

Whether either program is available, and which.

missing

What to tell somebody who has neither program, worded so they can act.

of

Computes the fingerprint of one file.

read_ffmpeg

Reads what ffmpeg’s chromaprint muxer wrote to its output.

read_fpcalc

Reads fpcalc’s two lines: DURATION= and FINGERPRINT=.