soundfingerprinting is a C# framework designed for developers, enthusiasts, researchers in the fields of audio and digital signal processing, data mining, and alike. It implements an efficient algorithm of digital signal processing which allows developing a system of audio fingerprinting and recognition.
Following is a code sample that shows you how to extract unique characteristics from an audio file and later use them as identifiers to recognize unknown snippets from a variety of sources. These characteristics known as sub-fingerprints (or fingerprints, these 2 terms used interchangeably) will be stored in the configurable backend. The interfaces for fingerprinting and querying audio files have been implemented as Fluent Interfaces with Builder and Command patterns in mind.
private readonly IModelService modelService = new InMemoryModelService(); // store fingerprints in memory
private readonly IAudioService audioService = new NAudioService(); // use NAudio audio processing library
private readonly IFingerprintCommandBuilder fingerprintCommandBuilder = new FingerprintCommandBuilder();
public void StoreAudioFileFingerprintsInStorageForLaterRetrieval(string pathToAudioFile)
{
TrackData track = new TrackData("GBBKS1200164", "Adele", "Skyfall", "Skyfall", 2012, 290);
// store track metadata in the database
var trackReference = modelService.InsertTrack(track);
// create sub-fingerprints and its hash representation
var hashedFingerprints = fingerprintCommandBuilder
.BuildFingerprintCommand()
.From(pathToAudioFile)
.UsingServices(audioService)
.Hash()
.Result;
// store sub-fingerprints and its hash representation in the database
modelService.InsertHashedFingerprintsForTrack(hashedFingerprints, trackReference);
}The default storage, which comes bundled with soundfingerprinting package, is a plain in memory storage, managed by InMemoryModelService. In case you would like to store fingerprints in a perstistent database you can take advantage of MSSQL integration available in SoundFingerprinting.SQL package via SqlModelService class. The MSSQL database initialization script as well as source files can be found here. Do not forget to add connection string FingerprintConnectionString in your app.config file, as described on the wiki.
Once you've inserted the fingerprints into the database, later you might want to query the storage in order to recognize the song those samples you have. The origin of query samples may vary: file, URL, microphone, radio tuner, etc. It's up to your application, where you get the samples from.
private readonly IQueryCommandBuilder queryCommandBuilder = new QueryCommandBuilder();
public TrackData GetBestMatchForSong(string queryAudioFile)
{
int secondsToAnalyze = 10; // number of seconds to analyze from query file
int startAtSecond = 0; // start at the begining
// query the underlying database for similar audio sub-fingerprints
var queryResult = queryCommandBuilder.BuildQueryCommand()
.From(queryAudioFile, secondsToAnalyze, startAtSecond)
.UsingServices(modelService, audioService)
.Query()
.Result;
if(queryResult.IsSuccessful)
{
return queryResult.BestMatch.Track; // successful match has been found
}
return null; // no match has been found
}See the Wiki Page for the operational details and information.
soundfingerprinting project structure, as well interface signatures has been drastically changed from release 1.0.1 to 2.0.0. This is due to the fact that git repository has been splitted to accommodate correct release cycles for all developed modules. This is a necessary evil which has been planned for a while, since releases were bloated with versioning issues.
- Unseen.Bass audio library integration soundfingerprinting.audio.bass
- MSSQL storage implementation (first choice for persistent storage) soundfingerprinrint.sql
- MongoDB storage implementation (bad performance still in alpha) soundfingerprinting.mongodb
- NeuralHasher recognition algorithm (still in alpha) soundfingerprinting.neuralhasher,
- and all demo apps are now located in separate git repositories, duplicates detector, sound tools.
This feature has been asked for a long time and is now available starting from release 2.0.0. QueryCommand interface has one additional method QueryWithTimeSequenceInformation which returns best candidate as well as query location in the resulting track.
Some of the interfaces which are used by the framework can be easily substituted according to your needs. In case you don't want to use NAudio as your audio library, you can take advantage of Bass.Net integration available through SoundFingerprinting.Audio.Bass package.
####Available integrations:
- SoundFingerprinting.Audio.Bass - Bass.Net audio library integration. Works faster, more accurate resampling, independent upon target OS. Sources available here
- SoundFingerprinting.SQL - implements integration with MSSQL storage. Sources available here
- SoundFingerprinting.MongoDb - implements integration with MongoDb, still in alpha phase. Source available here
Fingerprinting and Querying algorithms can be easily parametrized with corresponding configuration objects passed as parameters on command creation.
var hashDatas = fingerprintCommandBuilder
.BuildFingerprintCommand()
.From(samples)
.WithFingerprintConfig(
config =>
{
config.TopWavelets = 250; // increase number of top wavelets
config.Stride = new RandomStride(512, 256); // stride between sub-fingerprints
})
.UsingServices(audioService)
.Hash()
.Result;Each and every configuration parameter can influence the recognition rate, required storage, computational cost, etc. Stick with the defaults, unless you would like to experiment.
Links to the third party libraries used by soundfingerprinting project.
- NAudio
- FFTW - used as a default framework for FFT algorithm.
- Ninject - used to take advantage of dependency inversion principle.
Even though a couple of controversial topics are discussed recently in software community (here, and here) I'm still strongly committed to TDD practices. In case you'd like to contribute, your code has to come with well written unit or integration tests (when appropriate). Below are some coverage percentages for the released modules:
- SoundFingerprinting - 83%
- SoundFingerprinting.Audio.Bass - 82%
- SoundFingerprinting.Audio.NAudio - 78%
- SoundFingerprinting.SQL - 76%
- SoundFingerprinting.MongoDb - 98%
These coverage percentages are given only for the reference, they do not neceserally mean the code is without bugs (which is obvisouly not true).
- Can I apply this algorithm for speech recognition purposes? No. The granularity of one fingerprint is roughly ~1.86 seconds, thus any sound recording which is less than that will be disregarded.
- Can the algorithm detect exact query position in resulted track? Yes. Starting from version 2.0.0 it's possible to detect exact query snippet position within the best-matched track.
git clone git@github.com:AddictedCS/soundfingerprinting.git
In order to build latest version of the SoundFingerprinting assembly run the following command from repository root
.\build.cmd
Install-Package SoundFingerprinting
My description of the algorithm alogside with the demo project can be found on CodeProject The demo project is a Audio File Duplicates Detector. Its latest source code can be found here. Its a WPF MVVM project that uses the algorithm to detect what files are perceptually very similar.
If you want to contribute you are welcome to open issues or discuss on issues page. Feel free to contact me for any remarks, ideas, bug reports etc.
The framework is provided under GPLv3 licence agreement.
The framework implements the algorithm from Content Fingerprinting Using Wavelets paper.
© Soundfingerprinting, 2010-2015, ciumac.sergiu@gmail.com