Frequently Asked Questions

Q: Does the AcousticBrainz client upload my audio files? Do you store audio files?
A: No to both. The AcousticBrainz client analyzes your files and calculates the audio features from it, but it does not alter your files in any way. These audio features are then uploaded to the server. No actual audio files are uploaded or stored on the server. Never ever.
We invite you to examine the code that runs the server. There is only one API end-point that allows submission of anything and it has to be a JSON document of a certain format. Audio files are rejected by the server.

Q: Why does AcousticBrainz prefer that people use the provided static extractor builds, rather than letting others build their own versions?
A: Besides building your own version being a major pain, it is possible for different combinations of libraries and compilers (gcc vs clang, libffmpeg vs libav) to produce slightly different outputs. We're aiming to reduce the variances of output between the extractors in an effort to create the most consistent and stable dataset. We're going to build a test suite in the future that can evaluate the performance of a compiled extractor, but we're not quite there yet. So, for the time being, if it is possible in any way for you to use one of our static builds, we would greatly appreciate that.

Q: What is the difference between AcousticBrainz and AcoustID?
A: AcousticBrainz does a deep analysis of music in an effort to discover as many acoustic qualities of the music as possible. The generated data is then made available for other people to use this data in their own projects or companies. AcoustID, however, does a much lighter music analysis with one single goal of identifying a recording of music. AcoustID is also designed to be as fast as possible, so it only does the necessary analysis on the recordings. While on the surface the two projects look similar, they have very different goals.