view src/DistanceMetric.h @ 246:aac9ad4064ea subsequence tip

Fix incorrect handling of silent tail in the non-subsequence MATCH phase; some debug output changes

/* -*- c-basic-offset: 4 indent-tabs-mode: nil -*-  vi:set ts=8 sts=4 sw=4: */

/*
    Vamp feature extraction plugin using the MATCH audio alignment
    algorithm.

    Centre for Digital Music, Queen Mary, University of London.
    Copyright (c) 2007-2020 Simon Dixon, Chris Cannam, and Queen Mary
    University of London, Copyright (c) 2014-2015 Tido GmbH.
    
    This program is free software; you can redistribute it and/or
    modify it under the terms of the GNU General Public License as
    published by the Free Software Foundation; either version 2 of the
    License, or (at your option) any later version.  See the file
    COPYING included with this distribution for more information.
*/

#ifndef DISTANCE_METRIC_H
#define DISTANCE_METRIC_H

#include "MatchTypes.h"

class DistanceMetric
{
public:
    enum Metric {

        /** Calculate the Manhattan distance between feature
         *  vectors. If the vectors contain energy, as the default
         *  MATCH feature does, this could be considered as a squared
         *  Euclidean distance metric. */
        Manhattan,

        /** Calculate the Euclidean distance between feature vectors. */
        Euclidean,

        /** Calculate the cosine distance between feature vectors. The
         *  normalisation setting will be ignored as the result is
         *  already magnitude-independent. */
        Cosine,
    };

    enum DistanceNormalisation {
            
        /** Do not normalise distance metrics */
        NoDistanceNormalisation,

        /** Normalise distance metric for pairs of frames by the sum
         *  of the two frames. */
        NormaliseDistanceToSum,

        /** Normalise distance metric for pairs of frames by the log
         *  of the sum of the frames. */
        NormaliseDistanceToLogSum,
    };
    
    enum NoiseAddition {

        /** Don't add noise. */
        NoNoise,

        /** Add a constant noise term. This can help avoid
         *  mis-tracking when one file contains a lot of silence. */
        AddNoise,
    };
    
    struct Parameters {

        Parameters() :
            metric(Manhattan),
            norm(NormaliseDistanceToLogSum),
            noise(AddNoise),
            scale(150.)
        {}

        Metric metric;
        DistanceNormalisation norm;
        NoiseAddition noise;
        double scale;
    };
    
    DistanceMetric(Parameters params);

    ~DistanceMetric();
    
    /** Calculates the distance in some metric between two vectors,
     *  with an optional normalisation by the combined values in the
     *  vectors. Note that normalisation assumes the values are all
     *  non-negative.
     *
     *  @param f1 one of the vectors involved in the distance calculation
     *  @param f2 one of the vectors involved in the distance calculation
     *  @return the distance
     */
    distance_t calcDistance(const feature_t &f1,
                            const feature_t &f2);

    /**
     * Mostly for internal use and testing
     */
    distance_t scaleValueIntoDistanceRange(double value);
    
private:
    Parameters m_params;

    distance_t scaleAndTally(double);
    template <typename T> T scaleIntoRange(double);

    distance_t m_max;
    int m_overcount;
};

#endif