BBC Vamp PluginsBBCA collection of audio feature extraction algorithms from BBC Research and Development.
EnergyLow Level FeaturesCalculates the RMS energy and low energy ratio of a signal.- IntensityLow Level FeaturesCalculates the intensity of a signal and the intensity ratio for a number of sub-bands.
- PeaksLow Level FeaturesCalculates peak and trough values of a signal.
- RhythmLow Level FeaturesCalculates rhythmic features of a signal, including onsets and tempo.
- Spectral ContrastLow Level FeaturesCalculates the peak and valleys of the spectral contrast feature.
- Spectral FluxLow Level FeaturesCalculates the spectral flux.
- Speech/Music segmenterClassificationCalculates boundaries between speech and music.
BeatRootSimon Dixon and Chris Cannam- Cepstral Pitch TrackerChris Cannam
- Chordino and NNLS ChromaMatthias MauchHarmony and chord extraction plugins by Matthias Mauch at C4DM.
- ChordinoNotesChordino provides a simple chord transcription based on NNLS Chroma (as in the NNLS Chroma plugin). Chord profiles given by the user in the file chord.dict are used to calculate frame-wise chord similarities. Two simple (non-state-of-the-art!) algorithms are available that smooth these to provide a chord transcription: a simple chord change method, and a standard HMM/Viterbi approach.
- NNLS ChromaVisualisationThis plugin provides a number of features derived from a DFT-based log-frequency amplitude spectrum: some variants of the log-frequency spectrum, including a semitone spectrum derived from approximate transcription using the NNLS algorithm; and based on this semitone spectrum, different chroma features.
- TuningKey and TonalityThe tuning plugin can estimate the local and global tuning of piece. The same tuning method is used for the NNLS Chroma and Chordino plugins.
- Constant-QQueen Mary, University of LondonA plugin implementing the Constant-Q transform of a time-domain signal.
- ChromagramVisualisationExtract a Constant-Q spectrogram with constant ratio of centre frequency to resolution from the audio, then wrap it around into a single-octave chromagram.
- Constant-Q Spectrogram (Hz range)VisualisationExtract a spectrogram with constant ratio of centre frequency to resolution from the input audio, specifying the frequency range in Hz.
- Constant-Q Spectrogram (MIDI pitch range)VisualisationExtract a spectrogram with constant ratio of centre frequency to resolution from the input audio, specifying the frequency range in MIDI pitch units.
- Expressive MeansFrithjof Vollmer and Chris CannamPlugins for analysis of articulation, vibrato, and portamento in audio recordings of musical performances, with a primary focus on bowed stringed instruments.
- Expressive MeansTime → OnsetsFinds note onsets and durations in monophonic recordings based on changes in spectral content, power, and pitch.
- Expressive Means (advanced)Time → OnsetsFinds note onsets and durations in monophonic recordings based on changes in spectral content, power, and pitch.
- Expressive Means (advanced): ArticulationNotes → ExpressionIdentifies types and intensities of articulation in monophonic recordings.
- Expressive Means (advanced): PortamentoNotes → ExpressionIdentifies types and intensities of portamento instances in monophonic recordings.
- Expressive Means: ArticulationNotes → ExpressionIdentifies types and intensity of articulation in monophonic recordings.
- Expressive Means: PortamentoNotes → ExpressionIdentifies types and intensities of portamento instances in monophonic recordings.
- Expressive Means (advanced): Pitch VibratoNotes → ExpressionIdentifies types and intensities of pitch vibrato instances in monophonic recordings.
- Expressive Means: Pitch VibratoNotes → ExpressionIdentifies types and intensities of pitch vibrato instances in monophonic recordings.
Fan Chirp F0gramAudio Processing Group, Universidad de la Republica, UruguayFundamental-frequency salience visualisation based on the Fan Chirp Transform, from the Universidad de la Republica, Uruguay.- Fan Chirp Transform F0gramVisualisationThis plug-in produces a representation, called F0gram, which exhibits the salience of the fundamental frequency of the sound sources in the audio file. The computation of the F0gram makes use of the Fan Chirp Transform analysis. It is based on the article "Fan chirp transform for music representation" P. Cancela, E. Lopez, M. Rocamora, International Conference on Digital Audio Effects, 13th. DAFx-10. Graz, Austria - 6-10 Sep 2010.
HPCP - Harmonic Pitch Class ProfileMusic Technology Group, Universitat Pompeu Fabra
INESC Porto Beat TrackerSMC Group at INESC Porto- libxtract Vamp pluginsChris Cannam and Jamie BullockLow-level feature extraction plugins using Jamie Bullock's libxtract library to provide around 50 spectral and other features.
- AutocorrelationLow Level FeaturesExtract the autocorrelation of an audio signal.
- Average DeviationLow Level FeaturesExtract the average deviation of a range of values.
- Average Magnitude Difference FunctionLow Level FeaturesExtract the AMDF of an audio signal.
- Average Squared Difference FunctionLow Level FeaturesExtract the ASDF of an audio signal.
- Bark CoefficientsLow Level FeaturesExtract bark coefficients from an audio spectrum.
- Discrete Cosine TransformVisualisationExtract the DCT of an audio signal.
- Fundamental FrequencyLow Level FeaturesExtract the fundamental frequency of an audio signal.
- Fundamental Frequency (failsafe)Low Level FeaturesExtract the fundamental frequency of an audio signal (failsafe).
- Harmonic SpectrumVisualisationExtract the harmonics from an audio spectrum.
- Highest ValueLow Level FeaturesExtract the highest value from a given range.
- InharmonicityLow Level FeaturesExtract the inharmonicity of an audio spectrum.
- Irregularity ILow Level FeaturesExtract the irregularity (type I) of an audio spectrum.
- Irregularity IILow Level FeaturesExtract the irregularity (type II) of an audio spectrum.
- KurtosisLow Level FeaturesExtract the kurtosis of a range of values.
- LoudnessLow Level FeaturesExtract the loudness of an audio signal from its spectrum.
- Lowest ValueLow Level FeaturesExtract the lowest value from a given range.
- MeanLow Level FeaturesExtract the mean of a range of values.
- Mel-Frequency Cepstral CoefficientsLow Level FeaturesExtract MFCC from an audio spectrum.
- NoisinessLow Level FeaturesExtract the noisiness of an audio spectrum.
- Non-zero countLow Level FeaturesExtract the number of non-zero elements in an input spectrum.
- Odd/even Harmonic RatioLow Level FeaturesExtract the odd-to-even harmonic ratio of an audio spectrum.
- Peak SpectrumVisualisationExtract the spectral peaks from an audio spectrum.
- RMS AmplitudeLow Level FeaturesExtract the RMS amplitude of an audio signal.
- SkewnessLow Level FeaturesExtract the skewness of a range of values.
- Spectral Average DeviationExtract the average deviation of an audio spectrum.
- Spectral CentroidLow Level FeaturesExtract the spectral centroid of an audio spectrum.
- Spectral Crest MeasureLow Level FeaturesExtract the spectral crest measure of an audio spectrum.
- Spectral FlatnessLow Level FeaturesExtract the spectral flatness of an audio spectrum.
- Spectral KurtosisLow Level FeaturesExtract the kurtosis of an audio spectrum.
- Spectral RolloffLow Level FeaturesExtract the rolloff point of an audio spectrum.
- Spectral SharpnessLow Level FeaturesExtract the spectral sharpness of an audio spectrum.
- Spectral SkewnessLow Level FeaturesExtract the skewness of an audio spectrum.
- Spectral SlopeLow Level FeaturesExtract the spectral slope of an audio spectrum.
- Spectral SmoothnessLow Level FeaturesExtract the spectral smoothness of an audio spectrum.
- Spectral SpreadLow Level FeaturesExtract the spectral spread of an audio spectrum.
- Spectral Standard DeviationLow Level FeaturesExtract the standard deviation of an audio spectrum.
- Spectral VarianceLow Level FeaturesExtract the variance of an audio spectrum.
- SpectrumVisualisationExtract the spectrum of an audio signal.
- Standard DeviationLow Level FeaturesExtract the standard deviation of a range of values.
- Sum of ValuesLow Level FeaturesExtract the sum of the values in a given range.
- TonalityLow Level FeaturesExtract the tonality an audio spectrum.
- Tristimulus ILow Level FeaturesExtract the tristimulus (type I) of an audio spectrum.
- Tristimulus IILow Level FeaturesExtract the tristimulus (type II) of an audio spectrum.
- Tristimulus IIILow Level FeaturesExtract the tristimulus (type III) of an audio spectrum.
- VarianceLow Level FeaturesExtract the variance of a range of values.
- Zero Crossing RateLow Level FeaturesExtract the zero crossing rate of an audio signal.
- MARSYAS Vamp pluginsMarsyas PluginsLow-level feature extraction plugins containing functionality from the MARSYAS batch feature extractor.
- Marsyas - Batch Feature Extract - CentroidLow Level FeaturesMarsyas - Batch Feature Extract - Centroid.
- Marsyas - Batch Feature Extract - Line Spectral PairsLow Level FeaturesMarsyas - Batch Feature Extract - Line Spectral Pairs.
- Marsyas - Batch Feature Extract - Linear Prediction Cepstral CoefficientsLow Level FeaturesMarsyas - Batch Feature Extract - Linear Prediction Cepstral Coefficients.
- Marsyas - Batch Feature Extract - Mel-Frequency Cepstral CoefficientsLow Level FeaturesMarsyas - Batch Feature Extract - Mel-Frequency Cepstral Coefficients.
- Marsyas - Batch Feature Extract - Spectral Crest FactorLow Level FeaturesMarsyas - Batch Feature Extract - Spectral Crest Factor.
- Marsyas - Batch Feature Extract - Spectral Flatness MeasureLow Level FeaturesMarsyas - Batch Feature Extract - Spectral Flatness Measure.
- Marsyas - Batch Feature Extract - Spectral RolloffLow Level FeaturesMarsyas - Batch Feature Extract - Spectral Rolloff.
- Marsyas - Batch Feature Extract - Zero CrossingsLow Level FeaturesMarsyas - Batch Feature Extract - Zero Crossings.
- MATCH Vamp PluginSimon Dixon and Chris Cannam
- Mazurka PluginsThe Mazurka ProjectSpectral visualisation and feature extraction plugins from the Mazurka project.
- ChronogramVisualisationChronogram.
- Harmonic SpectrogramVisualisationHarmonic Spectrogram.
- Nevermore SpectrogramVisualisationNevermore Spectrogram.
- Power CurveLow Level FeaturesPower Curve.
- Spectral FluxLow Level FeaturesSpectral Flux.
- Spectral RefluxLow Level FeaturesSpectral Reflux.
- MELODIA - Melody ExtractionMusic Technology Group, Universitat Pompeu FabraMelody pitch estimator for polyphonic music.
- MELODIA - Melody ExtractionPitchEstimates the melody pitch in polyphonic music (also good for homophonic and monophonic music). Segments without melody are indicated by zero or negative values. For further details please read:
J. Salamon and E. Gomez, "Melody Extraction from Polyphonic Music Signals using Pitch Contour Characteristics", IEEE Transactions on Audio, Speech and Language Processing, 20(6):1759-1770, 2012.
We would highly appreciate the above reference being cited in publications of work in which this plug-in was used. - MELODIA - Melody Extraction (intermediate steps)VisualisationProvides visualisations of the intermediate steps calculated by the melody extraction algorithm implemented in the MELODIA - Melody Extraction plug-in. For further details please read:
J. Salamon and E. Gomez, "Melody Extraction from Polyphonic Music Signals using Pitch Contour Characteristics", IEEE Transactions on Audio, Speech and Language Processing, 20(6):1759-1770, 2012.
We would highly appreciate the above reference being cited in publications of work in which this plug-in was used.
- MIR.EDUMIR.EDU by Justin SalamonAn open source Vamp plug-in library written in C++ which implements a basic set of descriptors useful for teaching MIR.
- MIR.EDU: Attack Start/End TimesCompute the start and end times of the attack of the signal. The start and end times of the attack are computed according to Peeters (2004), see code/documentation for further details. NOTE: the accuracy of the estimation depends on the step (hop) size of the analysis, the smaller the better. The times are returned as timestamps without any values.
- MIR.EDU: Log Attack TimeCompute the logarithm (base 10) of the duration of the attack of the signal (unit: log10(seconds)). The start and end times of the attack are computed according to Peeters (2004), see code/documentation for further details. NOTE: the accuracy of the estimation depends on the step (hop) size of the analysis, the smaller the better. The the timestamp of the returned value is the start time of the attack. For visualising the attack start and end times use the 'Attack Start/End Times' plug-in.
- MIR.EDU: MFCCCompute the Mel Frequency Cepstral Coefficients (MFCC) for each frame. MFCCs provide a concise representation of the spectral envelope of a sound, which in turn is related to the sound's timbre. Please refer to the code in MFCC.cpp and the reference provided in MFCC.h for a detailed explanation of how MFCCs are computed.
- MIR.EDU: RMSCompute the root mean square of the signal for each frame.
- MIR.EDU: Spectral CentroidCompute the spectral centroid of the signal for each frame. The unit of the values returned is Hz. If the frame is completely silent a value of 0 is returned.
- MIR.EDU: Spectral CrestCompute the spectral crest of the signal for each frame. The crest is defined as the ratio between the maximum spectral magnitude in the frame and the arithmetical mean of the spectral magnitudes. The minimum possible value is 1 (flat spectrum) and it increases the peakier the spectrum is. If the frame is completely silent a value of 1 (flat spectrum) is returned.
- MIR.EDU: Spectral FlatnessCompute the spectral flatness of the signal for each frame. The flatness is defined as the ratio of the geometric and arithmetical means of the spectral magnitudes. The values returned range between 0 (peaky spectrum) and 1 (flat spectrum). If the frame is completely silent a value of 1 (flat spectrum) is returned.
- MIR.EDU: Spectral FluxCompute the spectral flux between every two consecutive frames of the signal. The flux is defined as 1 minus the normalized correlation between successive magnitude spectra. The values returned range between 0 (no change) and 1 (maximum change). For the first frame the flux is always zero. If both frames are silent flux = 0, if only one is silent flux = 1.
- MIR.EDU: Spectral KurtosisCompute the spectral kurtosis of the signal for each frame. The kurtosis is a measure of the peakedness of a distribution. For a gaussian distribution kurtosis = 3, for a flat distribution kurtosis < 3 and for a peakier distribution kurtosis > 3.If the frame is completely silent a value of 0 is returned.
- MIR.EDU: Spectral Roll-offCompute the spectral roll-off of the signal for each frame, defined as the frequency below which 95% of the signal energy is contained. The threshold (95%) is defined as a paramtere that can be changed by the user. The unit of the values returned is Hz. If the frame is completely silent a value of 0 is returned.
- MIR.EDU: Spectral SkewnessCompute the spectral skewness of the signal for each frame. If the frame is completely silent a value of 0 is returned.
- MIR.EDU: Spectral SpreadCompute the spectral spread of the signal for each frame. The unit of the values returned is Hz. If the frame is completely silent a value of 0 is returned.
- MIR.EDU: Temporal CentroidCompute the temporal centroid of the entire signal, which is the centre of gravity of the energy of the signal. Energy is represented by the RMS of the signal. The termporal centroid is computed between times n1 and n2 which are the first and last times the signal RMS is above 15% of its maximum value. The temporal centroid is returned as a timestamp with no corresponding value.
- MIR.EDU: Zero Crossing RateCompute the zero crossing rate of the signal for each frame (i.e. the number of times the signal changes sign). The unit of the values returned is crossings/second.
- OFA Vamp PluginChris CannamPlugin that performed audio fingerprinting and lookup using the no-longer-supported MusicIP OFA library. This plugin is provided for interest only and is no longer practically useful.
- MusicIP Audio FingerprinterCalculates an audio fingerprint using the MusicIP OFA fingerprinting library.
- MusicIP PUID LookupCalculates an audio fingerprint using the MusicIP OFA fingerprinting library and uses it to look up a MusicDNS PUID.
- OnsetsDS pluginDan Stowell
- pYINMatthias MauchpYIN is a modification of the well-loved YIN algorithm for fundamental frequency (F0) estimation in monophonic audio.
- Local Candidate PYINMonophonic pitch and note tracking based on a probabilistic Yin extension.
- pYinPitchMonophonic pitch and note tracking based on a probabilistic Yin extension.
- YinPitchA vamp implementation of the Yin algorithm for monophonic frequency estimation.
Queen Mary plugin setQueen Mary, University of LondonPlugins from the Centre for Digital Music at Queen Mary, University of London.- Adaptive SpectrogramVisualisationProduce an adaptive spectrogram by adaptive selection from spectrograms at multiple resolutions.
- Bar and Beat TrackerTime → TempoEstimate bar and beat locations.
- ChromagramVisualisationExtract a series of tonal chroma vectors from the audio.
- Constant-Q SpectrogramVisualisationExtract a spectrogram with constant ratio of centre frequency to resolution from the input audio.
- Discrete Wavelet TransformVisualisationVisualisation by scalogram.
- Key DetectorKey and TonalityEstimate the key of the music.
- Mel-Frequency Cepstral CoefficientsLow Level FeaturesCalculate a series of MFCC vectors from the audio.
- Note Onset DetectorTime → OnsetsEstimate individual note onset positions.
- Polyphonic TranscriptionNotesTranscribe the input audio to estimated notes.
- SegmenterClassificationDivide the track into a sequence of consistent segments.
- SimilarityClassificationReturn a distance matrix for similarity between the input audio channels.
- Tempo and Beat TrackerTime → TempoEstimate beat locations and tempo.
- Tonal ChangeKey and TonalityDetect and return the positions of harmonic changes such as chord boundaries.
RGU Mel-Frequency SpectrumRobert Gordon University- SegmentinoQueen Mary, University of London
- Silvet Note TranscriptionQueen Mary, University of London
- Simple CepstrumChris CannamA simple Vamp plugin to calculate and return cepstrum values from DFT bins. Useful as a preliminary tool for looking at cepstral data for simple pitch or envelope methods.
- Simple CepstrumLow Level FeaturesReturn simple cepstral data from DFT bins. This plugin is intended for casual inspection of cepstral data. It returns a lot of different sorts of data and is quite slow; it's not a good way to extract a single feature rapidly.
- Stereo PlanChris CannamAzimuth-based stereo plan.
- Stereo PlanReturn a stereo plan decomposition of the audio. The returned feature grid covers the stereo plan from left-channel-only (first bin) to right-channel-only (last bin), with each value indicating what proportion of signal energy is found at that point on the plan at that moment. The input should consist of two channels containing left and right channel signals.
- TempogramCarl Bussey
- Tipic - Tibre-Invariant Pitch ChromaQueen Mary, University of London
- Tuning DifferenceChris Cannam
University of Alicante Vamp PluginsUniversity of AlicanteThe UAPlugins set is a library of Vamp plugins developed in the DRIMS project to perform onset detection and polyphonic transcription. The methods used in this library were developed by Antonio Pertusa and José Manuel Iñesta.- Note Onset DetectorTime → OnsetsNote onset detection using a one-semitone filterbank.
- Polyphonic TranscriptionNotesMultiple fundamental frequency estimation for polyphonic music transcription.
Vamp Aubio pluginsPaul BrossierThe Vamp Aubio Plugins provide methods from Paul Brossier's aubio annotation library, including beat and tempo tracking, onset detection, pitch detection, note tracking, silence detector, and Mel-frequency cepstral coefficients.- Aubio Beat TrackerTime → TempoEstimate the musical tempo and track beat positions.
- Aubio Mel-Frequency Bands ExtractorLow Level FeaturesCompute energy in each 40 mel-frequency bands.
- Aubio Mfcc ExtractorLow Level FeaturesExtract Mel-Frequency Cepstrum Coefficients.
- Aubio Note TrackerNotesEstimate note onset positions, pitches and durations.
- Aubio Onset DetectorTime → OnsetsEstimate note onset times.
- Aubio Pitch DetectorPitchTrack estimated note pitches.
- Aubio Silence DetectorLow Level FeaturesDetect levels below a certain threshold.
- Aubio Spectral DescriptorLow Level FeaturesCompute spectral description function.
- Vamp example pluginsVamp SDK Example PluginsA set of simple plugins as included with the Vamp developers kit. Amplitude tracker, simple percussion onset detector, tempo estimator, spectral centroid, power spectrum, and zero-crossing counter.
- Amplitude FollowerLow Level FeaturesTrack the amplitude of the audio signal.
- Simple Fixed Tempo EstimatorTime → TempoStudy a short section of audio and estimate its tempo, assuming the tempo is constant.
- Simple Percussion Onset DetectorTime → OnsetsDetect percussive note onsets by identifying broadband energy rises.
- Simple Power SpectrumVisualisationReturn the power spectrum of a signal.
- Spectral CentroidLow Level FeaturesCalculate the centroid frequency of the spectrum of the input signal.
- Zero CrossingsLow Level FeaturesDetect and count zero crossing points.