Mercurial > hg > vamp-website

annotate plugin-doc/qm-vamp-plugins.html @ 101:bb33b0c75481 rdfquery

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Remove download.html (now dynamic)
author Chris Cannam
date Tue, 24 Jun 2014 13:32:30 +0100
parents bcb5a0818120
children 678a88672953
rev   line source
cannam@16 1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
cannam@16 2 <html>
cannam@16 3 <head>
cannam@16 4 <link rel="stylesheet" media="screen" type="text/css" href="/screen.css"/>
cannam@16 5 <link rel="icon" type="image/png" href="/images/waveform.png"/>
cannam@16 6 <link rel="shortcut" type="image/png" href="/images/waveform.png"/>
cannam@16 7 <title>QM Vamp Plugins: User Documentation</title>
cannam@16 8 <meta name="robots" content="index"/>
cannam@16 9 </head>
cannam@16 10 <body>
cannam@16 11 <h1 id="header"><span>Vamp Plugins</span></h1>
cannam@16 12
cannam@16 13 <h2>QM Vamp Plugins</h2>
cannam@16 14
cannam@16 15 <p>The QM Vamp Plugin set is a library of Vamp audio feature
cannam@16 16 extraction plugins developed at the <a
cannam@16 17 href="http://www.elec.qmul.ac.uk/digitalmusic/">Centre for Digital
cannam@16 18 Music</a> at Queen Mary, University of London. These plugins are
cannam@16 19 provided as a single library file, made available in binary form for
cannam@16 20 Windows, OS/X, and Linux from the Centre for Digital Music's <a
cannam@32 21 href="http://isophonics.net/QMVampPlugins">download
cannam@16 22 page</a>.
cannam@16 23 </p>
cannam@16 24 <p>For more information about Vamp plugins, see <a href="http://www.vamp-plugins.org/">http://www.vamp-plugins.org/</a> .
cannam@16 25 </p>
cannam@16 26
cannam@16 27 <div class="toc2">1. &nbsp;<a href="#qm-onsetdetector">Note Onset Detector</a></div>
cannam@16 28 <div class="toc2">2. &nbsp;<a href="#qm-tempotracker">Tempo and Beat Tracker</a></div>
cannam@29 29 <div class="toc2">3. &nbsp;<a href="#qm-barbeattracker">Bar and Beat Tracker</a></div>
cannam@29 30 <div class="toc2">4. &nbsp;<a href="#qm-keydetector">Key Detector</a></div>
cannam@29 31 <div class="toc2">5. &nbsp;<a href="#qm-tonalchange">Tonal Change</a></div>
cannam@29 32 <div class="toc2">6. &nbsp;<a href="#qm-adaptivespectrogram">Adaptive Spectrogram</a></div>
cannam@29 33 <div class="toc2">7. &nbsp;<a href="#qm-transcription">Polyphonic Transcription</a></div>
cannam@29 34 <div class="toc2">8. &nbsp;<a href="#qm-segmenter">Segmenter</a></div>
cannam@29 35 <div class="toc2">9. &nbsp;<a href="#qm-similarity">Similarity</a></div>
cannam@29 36 <div class="toc2">10. &nbsp;<a href="#qm-dwt">Discrete Wavelet Transform</a></div>
cannam@29 37 <div class="toc2">11. &nbsp;<a href="#qm-constantq">Constant-Q Spectrogram</a></div>
cannam@29 38 <div class="toc2">12. &nbsp;<a href="#qm-chromagram">Chromagram</a></div>
cannam@29 39 <div class="toc2">13. &nbsp;<a href="#qm-mfcc">Mel-Frequency Cepstral Coefficients</a></div>
cannam@16 40
cannam@29 41 <a name="qm-onsetdetector"></a><h2>1. Note Onset Detector</h2>
cannam@16 42
cannam@16 43 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-onsetdetector</code>
cannam@16 44 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-onsetdetector">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-onsetdetector</a>
cannam@16 45 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 46 </p>
cannam@16 47 <p>Note Onset Detector analyses a single channel of audio and estimates
cannam@16 48 the onset times of notes within the music &ndash; that is, the times at
cannam@16 49 which notes and other audible events begin.
cannam@16 50 </p>
cannam@16 51 <p>It calculates an onset likelihood function for each spectral frame,
cannam@16 52 and picks peaks in a smoothed version of this function. The plugin is
cannam@16 53 non-causal, returning all results at the end of processing.
cannam@16 54 </p>
cannam@16 55 <h3>Parameters</h3>
cannam@16 56
cannam@16 57 <p><b>Onset Detection Function Type</b> &ndash; The method used to calculate the
cannam@16 58 onset likelihood function. The most versatile method is the default,
cannam@16 59 "Complex Domain" (see reference, Duxbury et al 2003). "Spectral
cannam@16 60 Difference" may be appropriate for percussive recordings, "Phase
cannam@16 61 Deviation" for non-percussive music, and "Broadband Energy Rise" (see
cannam@16 62 reference, Barry et al 2005) for identifying percussive onsets in
cannam@16 63 mixed music.
cannam@16 64 </p>
cannam@16 65 <p><b>Onset Detector Sensitivity</b> &ndash; Sensitivity level for peak detection
cannam@16 66 in the onset likelihood function. The higher the sensitivity, the
cannam@16 67 more onsets will (rightly or wrongly) be detected. The peak picker
cannam@16 68 does not have a simple threshold level; instead, this parameter
cannam@16 69 controls the required "steepness" of the slopes in the smoothed
cannam@16 70 detection function either side of a peak value, in order for that peak
cannam@16 71 to be accepted as an onset.
cannam@16 72 </p>
cannam@16 73 <p><b>Adaptive Whitening</b> &ndash; This option evens out the temporal and
cannam@16 74 frequency variation in the signal, which can yield improved
cannam@16 75 performance in onset detection, for example in audio with big
cannam@16 76 variations in dynamics.
cannam@16 77 </p>
cannam@16 78 <h3>Outputs</h3>
cannam@16 79
cannam@16 80 <p><b>Note Onsets</b> &ndash; The detected note onset times, returned as a single
cannam@16 81 feature with timestamp but no value for each detected note.
cannam@16 82 </p>
cannam@16 83 <p><b>Onset Detection Function</b> &ndash; The raw note onset likelihood function
cannam@16 84 that was calculated as the first step of the detection process.
cannam@16 85 </p>
cannam@16 86 <p><b>Smoothed Detection Function</b> &ndash; The note onset likelihood function
cannam@16 87 following median filtering. This is the function from which
cannam@16 88 sufficiently steep peak values are picked and classified as onsets.
cannam@16 89 </p>
cannam@16 90 <h3>References and Credits</h3>
cannam@16 91
cannam@16 92 <p><b>Basic detection methods</b>: C. Duxbury, J. P. Bello, M. Davies and
cannam@16 93 M. Sandler, <i><a href="http://www.elec.qmul.ac.uk/dafx03/proceedings/pdfs/dafx81.pdf">Complex domain Onset Detection for Musical Signals</a></i>. In
cannam@16 94 Proceedings of the 6th Conference on Digital Audio Effects
cannam@16 95 (DAFx-03). London, UK. September 2003.
cannam@16 96 </p>
cannam@16 97 <p><b>Adaptive whitening</b>: D. Stowell and M. D. Plumbley, <i><a href="http://www.elec.qmul.ac.uk/digitalmusic/papers/2007/StowellPlumbley07-icmc.pdf">Adaptive whitening for improved real-time audio onset detection</a></i>. In
cannam@16 98 Proceedings of the International Computer Music Conference (ICMC'07),
cannam@16 99 August 2007.
cannam@16 100 </p>
cannam@16 101 <p><b>Percussion onset detector</b>: D. Barry, D. Fitzgerald, E. Coyle and
cannam@16 102 B. Lawlor, <i><a href="http://eleceng.dit.ie/papers/15.pdf">Drum Source Separation using Percussive Feature Detection and Spectral Modulation</a></i>. ISSC 2005.
cannam@16 103 </p>
cannam@16 104 <p>The Note Onset Detector Vamp plugin was written by Chris Duxbury, Juan
cannam@16 105 Pablo Bello and Christian Landone.
cannam@16 106 </p>
cannam@29 107
cannam@16 108 <a name="qm-tempotracker"></a><h2>2. Tempo and Beat Tracker</h2>
cannam@16 109
cannam@16 110 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-tempotracker</code>
cannam@16 111 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-tempotracker">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-tempotracker</a>
cannam@16 112 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 113 </p>
cannam@16 114 <p>Tempo and Beat Tracker analyses a single channel of audio and
cannam@16 115 estimates the positions of metrical beats within the music (the
cannam@16 116 equivalent of a human listener tapping their foot to the beat).
cannam@16 117 </p>
cannam@16 118 <h3>Parameters</h3>
cannam@16 119
cannam@46 120 <p><b>Beat Tracking Method</b> &ndash; The method used to track beats. The default, "New", uses a hybrid of the "Old" two-state beat tracking model
cannam@46 121 (see reference Davies 2007) and a dynamic programming method (see reference
cannam@46 122 Ellis 2007). A more detailed description is given below within the Bar and
cannam@46 123 Beat Tracker plugin. </p>
cannam@29 124
cannam@29 125 <p><b>Onset Detection Function Type</b> &ndash; The algorithm used to calculate the
cannam@16 126 onset likelihood function. The most versatile method is the default,
cannam@16 127 "Complex Domain" (see reference, Duxbury et al 2003). "Spectral
cannam@16 128 Difference" may be appropriate for percussive recordings, "Phase
cannam@16 129 Deviation" for non-percussive music, and "Broadband Energy Rise" (see
cannam@16 130 reference, Barry et al 2005) for identifying percussive onsets in
cannam@16 131 mixed music.
cannam@16 132 </p>
cannam@16 133 <p><b>Adaptive Whitening</b> &ndash; This option evens out the temporal and
cannam@16 134 frequency variation in the signal, which can yield improved
cannam@16 135 performance in onset detection, for example in audio with big
cannam@16 136 variations in dynamics.
cannam@16 137 </p>
cannam@16 138 <h3>Outputs</h3>
cannam@16 139
cannam@16 140 <p><b>Beats</b> &ndash; The estimated beat locations, returned as a single feature,
cannam@16 141 with timestamp but no value, for each beat, labelled with the
cannam@16 142 corresponding estimated tempo at that beat.
cannam@16 143 </p>
cannam@16 144 <p><b>Onset Detection Function</b> &ndash; The raw note onset likelihood function
cannam@16 145 used in beat estimation.
cannam@16 146 </p>
cannam@16 147 <p><b>Tempo</b> &ndash; The estimated tempo, returned as a feature each time the
cannam@16 148 estimated tempo changes, with a single value for the tempo in beats
cannam@16 149 per minute.
cannam@16 150 </p>
cannam@16 151 <h3>References and Credits</h3>
cannam@16 152
cannam@16 153 <p><b>Beat tracking method</b>: M. E. P. Davies and M. D. Plumbley.
cannam@16 154 <i><a href="http://www.elec.qmul.ac.uk/people/markp/2007/DaviesPlumbley07-taslp.pdf">Context-dependent beat tracking of musical audio</a></i>. In IEEE
cannam@16 155 Transactions on Audio, Speech and Language Processing. Vol. 15, No. 3,
cannam@46 156 pp1009-1020, 2007;<br>M. E. P. Davies and M. D. Plumbley.
cannam@16 157 <i><a href="http://www.elec.qmul.ac.uk/people/markp/2005/DaviesPlumbley05-icassp.pdf">Beat Tracking With A Two State Model</a></i>. In Proceedings of the IEEE
cannam@16 158 International Conference on Acoustics, Speech and Signal Processing
cannam@46 159 (ICASSP 2005), Vol. 3, pp241-244 Philadelphia, USA, March 19-23, 2005;
cannam@46 160 <br>D. P. W. Ellis. <i>Beat Tracking by Dynamic
cannam@46 161 Programming</i>. In Journal of New Music Research. Vol. 37, No. 1,
cannam@46 162 pp51-60, 2007.
cannam@16 163 </p>
cannam@16 164 <p><b>Onset detection methods</b>: C. Duxbury, J. P. Bello, M. Davies and
cannam@16 165 M. Sandler, <i><a href="http://www.elec.qmul.ac.uk/dafx03/proceedings/pdfs/dafx81.pdf">Complex domain Onset Detection for Musical Signals</a></i>. In
cannam@16 166 Proceedings of the 6th Conference on Digital Audio Effects
cannam@16 167 (DAFx-03). London, UK. September 2003.
cannam@16 168 </p>
cannam@16 169 <p><b>Adaptive whitening</b>: D. Stowell and M. D. Plumbley, <i><a href="http://www.elec.qmul.ac.uk/digitalmusic/papers/2007/StowellPlumbley07-icmc.pdf">Adaptive whitening for improved real-time audio onset detection</a></i>. In
cannam@16 170 Proceedings of the International Computer Music Conference (ICMC'07),
cannam@16 171 August 2007.
cannam@16 172 </p>
cannam@16 173 <p><b>Percussion onset detector</b>: D. Barry, D. Fitzgerald, E. Coyle and
cannam@16 174 B. Lawlor, <i><a href="http://eleceng.dit.ie/papers/15.pdf">Drum Source Separation using Percussive Feature Detection and Spectral Modulation</a></i>. ISSC 2005.
cannam@16 175 </p>
cannam@16 176 <p>The Tempo and Beat Tracker Vamp plugin was written by Matthew Davies
cannam@16 177 and Christian Landone.
cannam@16 178 </p>
cannam@29 179
cannam@29 180
cannam@29 181 <a name="qm-barbeattracker"></a><h2>3. Bar and Beat Tracker</h2>
cannam@29 182
cannam@29 183 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-barbeattracker</code>
cannam@29 184 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-barbeattracker">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-barbeattracker</a>
cannam@29 185 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@29 186 </p>
cannam@29 187
cannam@29 188 <p>Bar and Beat Tracker analyses a single channel of audio and
cannam@29 189 estimates the positions of bar lines and the resulting counted
cannam@29 190 metrical beat positions within the music (where the first beat of
cannam@29 191 each bar is "1", the equivalent of counting in time to the music).
cannam@29 192 It is closely related to the <a href="#qm-tempotracker">Tempo and
cannam@29 193 Beat Tracker</a>, producing the same results for beat position as
cannam@29 194 that plugin's "New" beat tracking method.
cannam@29 195
cannam@29 196 </p>
cannam@29 197
cannam@29 198 <h3>Method</h3>
cannam@29 199
cannam@29 200 <p>The plugin first calculates an onset detection function using the
cannam@29 201 "Complex Domain" method (see <a href="#qm-tempotracker">Tempo and Beat
cannam@29 202 Tracker</a>).</p>
cannam@29 203
cannam@29 204 <p>The beat tracking method performs two passes over the onset
cannam@29 205 detection function, first to estimate the tempo contour, and then
cannam@29 206 given the tempo, to recover the beat locations.</p>
cannam@29 207
cannam@29 208 <p>To identify the tempo, the onset detection function is partitioned
cannam@29 209 into 6-second frames with a 1.5-second increment. The autocorrelation
cannam@29 210 function of each 6-second onset detection function is found and this
cannam@29 211 is then passed through a perceptually weighted comb filterbank (see
cannam@29 212 reference Davies 2007). The successive comb filterbank output signals
cannam@29 213 are grouped together into a matrix of observations of periodicity
cannam@29 214 through time. The best path of periodicity through these observations
cannam@29 215 is found using the Viterbi algorithm, where the transition matrix is
cannam@29 216 defined as a diagonal Gaussian.</p>
cannam@29 217
cannam@29 218 <p>Given the estimates of periodicity, the beat locations are recovered
cannam@29 219 by applying the dynamic programming algorithm (see reference Ellis
cannam@29 220 2007). This process involves the calculation of a recursive cumulative
cannam@29 221 score function and backtrace signal. The cumulative score indicates
cannam@29 222 the likelihood of a beat existing at each sample of the onset
cannam@29 223 detection function input, and the backtrace gives the location of the
cannam@29 224 best previous beat given this point in time. Once the cumulative score
cannam@29 225 and backtrace have been calculated for the whole input signal, the
cannam@29 226 best path through beat locations is found by recursively sampling the
cannam@29 227 backtrace signal from the end of the input signal back to the
cannam@29 228 beginning. See reference Stark et al. 2009 for a description of the
cannam@29 229 real-time implementation of the beat tracking algorithm.</p>
cannam@29 230
cannam@29 231 <p>Once the beat locations have been identified, the plugin makes a
cannam@29 232 second pass over the input audio signal, partitioning it into beat
cannam@29 233 synchronous frames. The audio within each beat frame is down-sampled
cannam@29 234 to give a new sampling frequency of 2.8kHz. A beat-synchronous
cannam@29 235 spectral representation is then calculated within each frame, from
cannam@29 236 which a measure of beat spectral difference is calculated using
cannam@29 237 Jensen-Shannon divergence. The bar boundaries are identified as those
cannam@29 238 beat transitions leading to most consistent spectral change given the
cannam@29 239 specified number of beats per bar.</p>
cannam@29 240
cannam@29 241 <h3>Parameters</h3>
cannam@29 242
cannam@29 243 <p><b>Beats per Bar</b> &ndash; The number of beats per bar (or measure). The
cannam@29 244 plugin assumes that the number of beats per bar is fixed throughout
cannam@29 245 the music.
cannam@29 246 </p>
cannam@29 247 <h3>Outputs</h3>
cannam@29 248
cannam@29 249 <p><b>Beats</b> &ndash; The estimated beat locations, returned as a single feature,
cannam@29 250 with timestamp but no value, for each beat, labelled with the
cannam@29 251 number of that beat within the bar (e.g. consecutively 1, 2, 3, 4 for 4 beats to the bar).
cannam@29 252 </p>
cannam@29 253 <p><b>Bars</b> &ndash; The estimated bar line locations, returned as a single feature,
cannam@29 254 with timestamp but no value, for each bar.
cannam@29 255 </p>
cannam@29 256 <p><b>Beat Count</b> &ndash; The estimated beat locations, returned as a single feature,
cannam@29 257 with timestamp and a value corresponding to the
cannam@29 258 number of that beat within the bar. This is similar to the Beats output except that it returns a counting function rather than a series of instants.
cannam@29 259 </p>
cannam@29 260 <p><b>Beat Spectral Difference</b> &ndash; The new-bar likelihood function used in bar line estimation.
cannam@29 261 </p>
cannam@29 262
cannam@29 263 <h3>References and Credits</h3>
cannam@29 264
cannam@29 265 <p><b>Beat tracking method</b>: A. M. Stark, M. E. P. Davies and
cannam@29 266 M. D. Plumbley. <i>Real-time beat-synchronous analysis of musical
cannam@29 267 audio</i>. To appear in Proceedings of 12th International Conference
cannam@29 268 on Digital Audio Effects (DAFx). 2009;<br>M. E. P. Davies and
cannam@29 269 M. D. Plumbley. <i><a
cannam@29 270 href="http://www.elec.qmul.ac.uk/people/markp/2007/DaviesPlumbley07-taslp.pdf">Context-dependent
cannam@29 271 beat tracking of musical audio</a></i>. In IEEE Transactions on
cannam@29 272 Audio, Speech and Language Processing. Vol. 15, No. 3, pp1009-1020,
cannam@29 273 2007;<br>D. P. W. Ellis. <i>Beat Tracking by Dynamic
cannam@29 274 Programming</i>. In Journal of New Music Research. Vol. 37, No. 1,
cannam@29 275 pp51-60, 2007.</p>
cannam@29 276
cannam@29 277 <p><b>Bar finding method</b>: M. E. P. Davies and M. D. Plumbley. <i>A
cannam@29 278 spectral difference approach to extracting downbeats in musical
cannam@29 279 audio</i>. In Proceedings of 14th European Signal Processing Conference
cannam@29 280 (EUSIPCO), Italy, 2006.</p>
cannam@29 281
cannam@29 282 <p>The Bar and Beat Tracker Vamp plugin was written by Matthew Davies and Adam Stark.
cannam@29 283 </p>
cannam@29 284
cannam@29 285
cannam@29 286
cannam@29 287 <a name="qm-keydetector"></a><h2>4. Key Detector</h2>
cannam@16 288
cannam@16 289 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-keydetector</code>
cannam@16 290 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-keydetector">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-keydetector</a>
cannam@16 291 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 292 </p>
cannam@16 293 <p>Key Detector analyses a single channel of audio and continuously
cannam@16 294 estimates the key of the music by comparing the degree to which a
cannam@16 295 block-by-block chromagram correlates to the stored key profiles for
cannam@16 296 each major and minor key.
cannam@16 297 </p>
cannam@16 298 <p>The key profiles are drawn from analysis of Book I of the Well
cannam@16 299 Tempered Klavier by J S Bach, recorded at A=440 equal temperament.
cannam@16 300 </p>
cannam@16 301 <h3>Parameters</h3>
cannam@16 302
cannam@16 303 <p><b>Tuning Frequency</b> &ndash; The frequency of concert A in the music under
cannam@16 304 analysis.
cannam@16 305 </p>
cannam@16 306 <p><b>Window Length</b> &ndash; The number of chroma analysis frames taken into
cannam@16 307 account for key estimation. This controls how eager the key detector
cannam@16 308 will be to return short-duration tonal changes as new key changes (the
cannam@16 309 shorter the window, the more likely it is to detect a new key change).
cannam@16 310 </p>
cannam@16 311 <h3>Outputs</h3>
cannam@16 312
cannam@16 313 <p><b>Tonic Pitch</b> &ndash; The tonic pitch of each estimated key change,
cannam@16 314 returned as a single-valued feature at the point where the key change
cannam@16 315 is detected, with value counted from 1 to 12 where C is 1, C# or Db is
cannam@16 316 2, and so on up to B which is 12.
cannam@16 317 </p>
cannam@16 318 <p><b>Key Mode</b> &ndash; The major or minor mode of the estimated key, where
cannam@16 319 major is 0 and minor is 1.
cannam@16 320 </p>
cannam@16 321 <p><b>Key</b> &ndash; The estimated key for each key change, returned as a
cannam@16 322 single-valued feature at the point where the key change is detected,
cannam@16 323 with value counted from 1 to 24 where 1-12 are the major keys and
cannam@16 324 13-24 are the minor keys, such that C major is 1, C# major is 2, and
cannam@16 325 so on up to B major which is 12; then C minor is 13, Db minor is 14,
cannam@16 326 and so on up to B minor which is 24.
cannam@16 327 </p>
cannam@16 328 <p><b>Key Strength Plot</b> &ndash; A grid representing the ongoing key
cannam@16 329 "probability" throughout the music. This is returned as a feature for
cannam@16 330 each chroma frame, containing 25 bins. Bins 1-12 are the major keys
cannam@16 331 from C upwards; bins 14-25 are the minor keys from C upwards. The
cannam@16 332 13th bin is unused: it just provides space between the first and
cannam@16 333 second halves of the feature if displayed in a single plot.
cannam@16 334 </p>
cannam@16 335 <p>The outputs are also labelled with pitch or key as text.
cannam@16 336 </p>
cannam@16 337 <h3>References and Credits</h3>
cannam@16 338
cannam@16 339 <p><b>Method</b>: see K. Noland and M. Sandler. <i><a href="http://www.aes.org/e-lib/browse.cfm?elib=14140">Signal Processing Parameters for Tonality Estimation</a></i>. In Proceedings of Audio Engineering Society
cannam@16 340 122nd Convention, Vienna, 2007.
cannam@16 341 </p>
cannam@16 342 <p>The Key Detector Vamp plugin was written by Katy Noland and Christian
cannam@16 343 Landone.
cannam@16 344 </p>
cannam@29 345
cannam@29 346 <a name="qm-tonalchange"></a><h2>5. Tonal Change</h2>
cannam@16 347
cannam@16 348 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-tonalchange</code>
cannam@16 349 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-tonalchange">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-tonalchange</a>
cannam@16 350 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 351 </p>
cannam@16 352 <p>Tonal Change analyses a single channel of audio, detecting harmonic
cannam@16 353 changes such as chord boundaries.
cannam@16 354 </p>
cannam@16 355 <h3>Parameters</h3>
cannam@16 356
cannam@16 357 <p><b>Gaussian smoothing</b> &ndash; The window length for the internal smoothing
cannam@16 358 operation, in chroma analysis frames. This controls how eager the
cannam@16 359 tonal change detector will be to identify very short-term tonal
cannam@16 360 changes. The default value of 5 is quite short, and may lead to more
cannam@16 361 (not always meaningful) results being returned; for many purposes a
cannam@16 362 larger value, closer to the maximum of 20, may be appropriate.
cannam@16 363 </p>
cannam@16 364 <p><b>Chromagram minimum pitch</b> &ndash; The MIDI pitch value (0-127) of the
cannam@16 365 minimum pitch included in the internal chromagram analyis.
cannam@16 366 </p>
cannam@16 367 <p><b>Chromagram maximum pitch</b> &ndash; The MIDI pitch value (0-127) of the
cannam@16 368 maximum pitch included in the internal chromagram analyis.
cannam@16 369 </p>
cannam@16 370 <p><b>Chromagram tuning frequency</b> &ndash; The frequency of concert A in the
cannam@16 371 music under analysis.
cannam@16 372 </p>
cannam@16 373 <h3>Outputs</h3>
cannam@16 374
cannam@16 375 <p><b>Transform to 6D Tonal Content Space</b> &ndash; A representation of the
cannam@16 376 musical content in a six-dimensional tonal space onto which the
cannam@16 377 algorithm maps 12-bin chroma vectors extracted from the audio.
cannam@16 378 </p>
cannam@16 379 <p><b>Tonal Change Detection Function</b> &ndash; A function representing the
cannam@16 380 estimated likelihood of a tonal change occurring in each spectral
cannam@16 381 frame.
cannam@16 382 </p>
cannam@16 383 <p><b>Tonal Change Positions</b> &ndash; The resulting estimated positions of tonal
cannam@16 384 changes.
cannam@16 385 </p>
cannam@16 386 <h3>References and Credits</h3>
cannam@16 387
cannam@16 388 <p><b>Method</b>: C. A. Harte, M. Gasser, and M. Sandler. <i><a href="http://portal.acm.org/citation.cfm?id=1178723.1178727">Detecting harmonic change in musical audio</a></i>. In Proceedings of the 1st ACM workshop on
cannam@16 389 Audio and Music Computing Multimedia, Santa Barbara, 2006.
cannam@16 390 </p>
cannam@29 391 <p>The Tonal Change Vamp plugin was written by Chris Harte and Martin
cannam@16 392 Gasser.
cannam@16 393 </p>
cannam@29 394
cannam@29 395
cannam@29 396 <a name="qm-adaptivespectrogram"></a><h2>6. Adaptive Spectrogram</h2>
cannam@29 397
cannam@29 398 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-adaptivespectrogram</code>
cannam@29 399 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-adaptivespectrogram">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-adaptivespectrogram</a>
cannam@29 400 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@29 401 </p>
cannam@29 402
cannam@29 403 <p>Adaptive Spectrogram produces a composite spectrogram from a set of
cannam@29 404 series of short-time Fourier transforms at differing resolutions.
cannam@29 405 Values are selected from these spectrograms by repeated subdivision by
cannam@29 406 time and frequency in order to maximise an entropy function across
cannam@29 407 each column.</p>
cannam@29 408
cannam@29 409 <h3>Parameters</h3>
cannam@29 410
cannam@29 411 <p><b>Number of resolutions</b> &ndash; The number of distinct
cannam@29 412 resolutions to calculate and use. The resolutions will be consecutive
cannam@29 413 powers of two starting from the smallest resolution specified.</p>
cannam@29 414
cannam@29 415 <p><b>Smallest resolution</b> &ndash; The smallest of the set of
cannam@29 416 resolutions to use.</p>
cannam@29 417
cannam@29 418 <p><b>Omit alternate resolutions</b> &ndash; Causes the plugin to
cannam@29 419 ignore alternate resolutions (i.e. the smallest resolution multiplied
cannam@29 420 by 2, 8, 32, etc) when composing a spectrogram. The smallest
cannam@29 421 resolution specified, and its multiples by 4, 16, etc as applicable,
cannam@29 422 will be retained. The total number of resolutions actually included
cannam@29 423 in the resulting spectrogram will therefore be N/2 (for even N) or
cannam@29 424 (N+1)/2 (for odd N) where N is the value of the "number of
cannam@29 425 resolutions" parameter. This permits a wider range of resolutions to
cannam@29 426 be included with less processing, at obvious cost in quality.</p>
cannam@29 427
cannam@29 428 <p><b>Multi-threaded processing</b> &ndash; Enables multi-threading of
cannam@29 429 the spectrogram calculation. This usually results in somewhat faster
cannam@29 430 processing where multiple CPU cores are available.</p>
cannam@29 431
cannam@29 432 <p>As an example of the resolution parameters, if the "number of
cannam@29 433 resolutions" is set to 5, "smallest resolution" to 128, and "omit
cannam@29 434 alternate resolutions" is not used, the composite spectrogram will be
cannam@29 435 calculated using spectrograms from 128, 256, 512, 1024, and 2048 point
cannam@29 436 short-time Fourier transforms (with 50% overlap in each case). With
cannam@29 437 "omit alternate resolutions" set, the same parameters would result in
cannam@29 438 spectrograms from 128, 512, and 2048 point STFTs being used.</p>
cannam@29 439
cannam@29 440 <h3>References and Credits</h3>
cannam@29 441
cannam@29 442 <p><b>Method</b>: X. Wen and M. Sandler. <i><a href="http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=ISPECX000003000001000051000001">Composite spectrogram using multiple Fourier transforms</a></i>. IET Signal Processing, 3(1):51-63, 2009.
cannam@29 443 </p>
cannam@29 444
cannam@29 445 <p>The Adaptive Spectrogram Vamp plugin was written by Wen Xue and Chris Cannam.</p>
cannam@29 446
cannam@29 447 <a name="qm-transcription"></a><h2>7. Polyphonic Transcription</h2>
cannam@29 448
cannam@29 449 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-transcription</code>
cannam@29 450 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-transcription">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-transcription</a>
cannam@29 451 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@29 452
cannam@29 453 <p>The Polyphonic Transcription plugin estimates a note transcription
cannam@29 454 using MIDI pitch values from its input audio, returning a feature for
cannam@29 455 each note (with timestamp and duration) whose value is the MIDI pitch
cannam@29 456 number. Velocity is not estimated.</p>
cannam@29 457
cannam@29 458 <p>Although the published description of the method is described as
cannam@29 459 real-time, the implementation used in this plugin is non-causal; it
cannam@29 460 buffers its input to operate on in a single unit, doing all the real
cannam@29 461 work after its entire input has been received, and is very memory
cannam@29 462 intensive. However, it is relatively fast (faster than real-time)
cannam@29 463 compared to other polyphonic transcription methods.</p>
cannam@29 464
cannam@29 465 <p>The plugin works best at 44.1KHz input sample rate, and is tuned for
cannam@29 466 piano and guitar music.</p>
cannam@29 467
cannam@29 468
cannam@29 469 <h3>References and Credits</h3>
cannam@29 470
cannam@29 471 <p><b>Method</b>: R. Zhou and J. D. Reiss. <i>A Real-Time Polyphonic Music Transcription System</i>. In Proceedings of the Fourth Music Information Retrieval Evaluation eXchange (MIREX), Philadelphia, USA, 2008;<br>R. Zhou and J. D. Reiss. <i>A Real-Time Frame-Based Multiple Pitch Estimation Method Using the Resonator Time Frequency Image</i>. Third Music Information Retrieval Evaluation eXchange (MIREX), Vienna, Austria, 2007.</p>
cannam@29 472
cannam@29 473 <p>The Polyphonic Transcription Vamp plugin was written by Ruohua Zhou.</p>
cannam@29 474
cannam@29 475
cannam@29 476 <a name="qm-segmenter"></a><h2>8. Segmenter</h2>
cannam@16 477
cannam@16 478 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-segmenter</code>
cannam@16 479 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-segmenter">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-segmenter</a>
cannam@16 480 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 481 </p>
cannam@16 482 <p>Segmenter divides a single channel of music up into structurally
cannam@16 483 consistent segments. It returns a numeric value (the segment type)
cannam@16 484 for each moment at which a new segment starts.
cannam@16 485 </p>
cannam@16 486 <p>For music with clearly tonally distinguishable sections such as verse,
cannam@16 487 chorus, etc., segments with the same type may be expected to be
cannam@16 488 similar to one another in some structural sense. For example,
cannam@16 489 repetitions of the chorus are likely to share a segment type.
cannam@16 490 </p>
cannam@16 491 <p>The plugin only attempts to identify similar segments; it does not
cannam@16 492 attempt to label them. For example, it makes no attempt to tell you
cannam@16 493 which segment is the chorus.
cannam@16 494 </p>
cannam@16 495 <p>Note that this plugin does a substantial amount of processing after
cannam@16 496 receiving all of the input audio data, before it produces any results.
cannam@16 497 </p>
cannam@16 498 <h3>Method</h3>
cannam@16 499
cannam@16 500 <p>The method relies upon structural/timbral similarity to obtain the
cannam@16 501 high-level song structure. This is based on the assumption that the
cannam@16 502 distributions of timbre features are similar over corresponding
cannam@16 503 structural elements of the music.
cannam@16 504 </p>
cannam@16 505 <p>The algorithm works by obtaining a frequency-domain representation of
cannam@16 506 the audio signal using a Constant-Q transform, a Chromagram or
cannam@16 507 Mel-Frequency Cepstral Coefficients (MFCC) as underlying features (the
cannam@16 508 particular feature is selectable as a parameter). The extracted
cannam@16 509 features are normalised in accordance with the MPEG-7 standard (NASE
cannam@16 510 descriptor), which means the spectrum is converted to decibel scale
cannam@16 511 and each spectral vector is normalised by the RMS energy envelope.
cannam@16 512 The value of this envelope is stored for each processing block of
cannam@16 513 audio. This is followed by the extraction of 20 principal components
cannam@16 514 per block using PCA, yielding a sequence of 21 dimensional feature
cannam@16 515 vectors where the last element in each vector corresponds to the
cannam@16 516 energy envelope.
cannam@16 517 </p>
cannam@16 518 <p>A 40-state Hidden Markov Model is then trained on the whole sequence
cannam@16 519 of features, with each state of the HMM corresponding to a specific
cannam@16 520 timbre type. This process partitions the timbre-space of a given track
cannam@16 521 into 40 possible types. The important assumption of the model is that
cannam@16 522 the distribution of these features remain consistent over a structural
cannam@16 523 segment. After training and decoding the HMM, the song is assigned a
cannam@16 524 sequence of timbre-features according to specific timbre-type
cannam@16 525 distributions for each possible structural segment.
cannam@16 526 </p>
cannam@16 527 <p>The segmentation itself is computed by clustering timbre-type
cannam@16 528 histograms. A series of histograms are created over a sliding window
cannam@16 529 which are grouped into M clusters by an adapted soft k-means
cannam@16 530 algorithm. Each of these clusters will correspond to a specific
cannam@16 531 segment-type of the analyzed song. Reference histograms, iteratively
cannam@16 532 updated during clustering, describe the timbre distribution for each
cannam@16 533 segment. The segmentation arises from the final cluster assignments.
cannam@16 534 </p>
cannam@16 535 <h3>Parameters</h3>
cannam@16 536
cannam@16 537 <p><b>Number of segment-types</b> &ndash; The maximum number of clusters
cannam@16 538 (segment-types) to be returned. The default is 10. Unlike many
cannam@16 539 clustering algorithms, the constrained clustering used in this plugin
cannam@16 540 does not produce too many clusters or vary significantly even if this
cannam@16 541 is set too high. However, this parameter can be useful for limiting
cannam@16 542 the number of expected segment-types.
cannam@16 543 </p>
cannam@16 544 <p><b>Feature Type</b> &ndash; The type of spectral feature used for segmentation. The available features are:<ul><li>"Hybrid", the default, which uses a Constant-Q transform (see <a href="#qm-constantq">related
cannam@16 545 plugin</a>): this is generally effective for modern studio recordings;</li><li> "Chromatic", using a chromagram derived from the Constant-Q feature (see <a href="#qm-chromagram">related plugin</a>): this may be preferable for live, acoustic, or older recordings, in which repeated sections may be less consistent in
cannam@16 546 sound;</li><li>"Timbral", using Mel-Frequency
cannam@16 547 Cepstral Coefficients (see <a href="#qm-mfcc">related plugin</a>), which is more likely to
cannam@16 548 result in classification by instrumentation rather than musical
cannam@16 549 content.</li></ul>
cannam@16 550 </p>
cannam@16 551 <p><b>Minimum segment duration</b> &ndash; The approximate expected minimum
cannam@16 552 duration for a segment, from 1 to 15 seconds. Changing this parameter
cannam@16 553 may help the plugin to find musical sections rather than just
cannam@16 554 following changes in the sound of the music, and also avoid wasting a
cannam@16 555 segment-type cluster for timbrally distinct but too-short segments.
cannam@16 556 The default of 4 seconds usually produces good results.
cannam@16 557 </p>
cannam@16 558 <h3>Outputs</h3>
cannam@16 559
cannam@16 560 <p><b>Segmentation</b> &ndash; The estimated segment boundaries, returned as a
cannam@16 561 single feature with one value at each segment boundary, with the value
cannam@16 562 representing the segment type number for the segment starting at that
cannam@16 563 boundary.
cannam@16 564 </p>
cannam@16 565 <h3>References and Credits</h3>
cannam@16 566
cannam@16 567 <p><b>Method</b>: M. Levy and M. Sandler. <i><a href="http://ieeexplore.ieee.org/iel5/10376/4432632/04432648.pdf?arnumber=4432648">Structural segmentation of musical audio by constrained clustering</a></i>. IEEE Transactions on Audio, Speech, and Language Processing, February 2008.
cannam@16 568 </p>
cannam@16 569 <p>Note that this plugin does not implement the beat-sychronous aspect
cannam@16 570 of the segmentation method described in the paper.
cannam@16 571 </p>
cannam@16 572 <p>The Segmenter Vamp plugin was written by Mark Levy. Thanks to George
cannam@16 573 Fazekas for providing much of this documentation.
cannam@16 574 </p>
cannam@29 575 <a name="qm-similarity"></a><h2>9. Similarity</h2>
cannam@16 576
cannam@16 577 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-similarity</code>
cannam@16 578 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-similarity">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-similarity</a>
cannam@16 579 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 580 </p>
cannam@16 581 <p>Similarity treats each channel of its audio input as a separate
cannam@16 582 "track", and estimates how similar the tracks are to one another using
cannam@16 583 a selectable similarity measure.
cannam@16 584 </p>
cannam@16 585 <p>The plugin also returns the intermediate data used as a basis of the
cannam@16 586 similarity measure; it can therefore be used on a single channel of
cannam@16 587 input (with the resulting intermediate data then being applied in some
cannam@16 588 other similarity or clustering algorithm, for example) if desired, as
cannam@16 589 well as with multiple inputs.
cannam@16 590 </p>
cannam@16 591 <p>Because of the way this plugin handles multiple inputs, by assuming
cannam@16 592 that each channel represents a separate piece of music, it may not be
cannam@16 593 appropriate for use directly in a general-purpose host (unless you
cannam@16 594 actually want to do something like compare two stereo channels for
cannam@16 595 timbral similarity, which is unlikely).
cannam@16 596 </p>
cannam@16 597 <h3>Parameters</h3>
cannam@16 598
cannam@16 599 <p><b>Feature Type</b> &ndash; The underlying audio feature used for the similarity
cannam@16 600 measure. The available features are:
cannam@16 601 <ul><li>"Timbre", in which the distance
cannam@16 602 between tracks is a symmetrised Kullback-Leibler divergence between
cannam@16 603 Gaussian-modelled MFCC means and variances across each track, for the
cannam@16 604 first 20 MFCCs including C0 (see <a href="#qm-mfcc">related plugin</a>);</li><li>"Chroma", which uses Kullback-Leibler divergence of
cannam@16 605 mean chroma histogram (see <a href="#qm-chromagram">related plugin</a>);</li><li>"Rhythm", using the cosine distance between
cannam@16 606 "beat spectrum" measures derived from a short sampled section of the
cannam@16 607 track;</li><li>and combined "Timbre and Rhythm" and "Chroma and Rhythm"
cannam@16 608 features.</li></ul>
cannam@16 609 </p>
cannam@16 610 <h3>Outputs</h3>
cannam@16 611
cannam@16 612 <p><b>Distance Matrix</b> &ndash; A matrix of the distance measures between input
cannam@16 613 channels, returned as a series of vector features timestamped at
cannam@16 614 one-second intervals. The distance from channel i to channel j
cannam@16 615 appears as the j'th bin of the feature at time i.
cannam@16 616 </p>
cannam@16 617 <p><b>Distance from First Channel</b> &ndash; A single vector feature, timestamped
cannam@16 618 at time zero, containing the distances between the first input channel
cannam@16 619 and each of the input channels (including the first channel itself at
cannam@16 620 bin 0, which should have zero distance).
cannam@16 621 </p>
cannam@16 622 <p><b>Ordered Distances from First Channel</b> &ndash; A pair of vector features,
cannam@16 623 at times 0 and 1 second. The feature at time 0 contains the 1-based
cannam@16 624 indices of the input channels in the order of similarity to the first
cannam@16 625 input channel (so its first bin should always contain 1, as the first
cannam@16 626 channel is most similar to itself). The feature at time 1 contains,
cannam@16 627 in bin n, the distance between the first input channel and the channel
cannam@16 628 with index found at bin n of the feature at time 0.
cannam@16 629 </p>
cannam@16 630 <p><b>Feature Means</b> &ndash; A series of vector features containing the mean
cannam@16 631 values of each of the feature bins across the duration of each of the
cannam@16 632 input channels. This output returns one feature for each input
cannam@16 633 channel, timestamped at one-second intervals. The number of bins for
cannam@16 634 each feature depends on the feature type; it will be 20 for MFCC
cannam@16 635 features and 12 for chroma features. No features will be returned on
cannam@16 636 this output if the feature type is purely rhythmic.
cannam@16 637 </p>
cannam@16 638 <p><b>Feature Variances</b> &ndash; Just as Feature Means, but variances.
cannam@16 639 </p>
cannam@16 640 <p><b>Beat Spectra</b> &ndash; A series of vector features containing the rhythmic
cannam@16 641 autocorrelation profiles (beat spectra) for each of the input
cannam@16 642 channels. This output returns one 512-bin feature for each input
cannam@16 643 channel, timestamped at one-second intervals. No features will be
cannam@16 644 returned on this output if the feature type contains no rhythm
cannam@16 645 component.
cannam@16 646 </p>
cannam@16 647 <h3>References and Credits</h3>
cannam@16 648
cannam@16 649 <p><b>Timbral similarity</b>: M. Levy and M. Sandler. <i><a href="http://www.elec.qmul.ac.uk/easaier/papers/mlevytimbralsimilarity.pdf">Lightweight measures for timbral similarity of musical audio</a></i>. In Proceedings of the 1st
cannam@16 650 ACM workshop on Audio and Music Computing Multimedia, Santa Barbara,
cannam@16 651 2006.
cannam@16 652 </p>
cannam@16 653 <p><b>Combined rhythmic and timbral similarity</b>: K. Jacobson. <i><a href="http://ismir2006.ismir.net/PAPERS/ISMIR0696_Paper.pdf">A Multifaceted Approach to Music Similarity</a></i>. In Proceedings of the
cannam@16 654 Seventh International Conference on Music Information Retrieval
cannam@16 655 (ISMIR), 2006.
cannam@16 656 </p>
cannam@16 657 <p>The Similarity Vamp plugin was written by Mark Levy, Kurt Jacobson and
cannam@16 658 Chris Cannam.
cannam@16 659 </p>
cannam@29 660
cannam@29 661
cannam@29 662 <a name="qm-dwt"></a><h2>10. Discrete Wavelet Transform</h2>
cannam@29 663
cannam@29 664 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-dwt</code>
cannam@29 665 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-dwt">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-dwt</a>
cannam@29 666 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@29 667
cannam@29 668 <p>Discrete Wavelet Transform plugin performs the forward DWT on the
cannam@29 669 signal. The wavelet coefficients are derived from a fast segmented DWT
cannam@29 670 algorithm without block end effects. The DWT can be performed with
cannam@29 671 various functions from a selection of wavelets up to the 16th scale.<p>
cannam@29 672
cannam@29 673 <p>The wavelet coefficients are returned as feature columns at a rate of
cannam@29 674 half the sample rate of the signal to be analysed. To simulate
cannam@29 675 multiresolution in the layer data table, the coefficient values at
cannam@29 676 higher scales are copied multiple times according to the number of the
cannam@29 677 scale. For example, for scale 2 each value will appear twice, at scale
cannam@29 678 3 they will be appear four times, at scale 4 there will be 8 times the
cannam@29 679 same coefficient value in order to simulate the lower resolution at
cannam@29 680 higher scales.</p>
cannam@29 681
cannam@29 682 <h3>Parameters</h3>
cannam@29 683
cannam@29 684 <p><b>Scales</b> &ndash; Adjusts the number of scales of the DWT. The
cannam@29 685 processing block size needs to be set to at least 2<sup>n</sup>, where n =
cannam@29 686 number of scales.</p>
cannam@29 687
cannam@29 688 <p><b>Wavelet</b> &ndash; Selects the wavelet function to be used for
cannam@29 689 the transform. Wavelets from the following families are available:
cannam@29 690 Daubechies, Symlets, Coiflets, Biorthogonal, Meyer.</p>
cannam@29 691
cannam@29 692 <h3>References and Credits</h3>
cannam@29 693
cannam@29 694 <p><b>Principles</b>: S. Mallat. <i>A theory for multiresolution signal decomposition: the wavelet representation</i>. In IEEE Transactions on Pattern Analysis and Machine Intelligence, 11 (1989), pp. 674-693;<br>
cannam@29 695 P. Rajmic and J. Vlach. <i>Real-Time Audio Processing via Segmented Wavelet Transform</i>. In Proceedings of the 10th Int. Conference on Digital Audio Effects (DAFx-07), Bordeaux, France, September 10-15, 2007.</p>
cannam@29 696
cannam@29 697 <p>The Discrete Wavelet Transform plugin was written by Thomas Wilmering.</p>
cannam@29 698
cannam@29 699 <a name="qm-constantq"></a><h2>11. Constant-Q Spectrogram</h2>
cannam@16 700
cannam@16 701 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-constantq</code>
cannam@16 702 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-constantq">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-constantq</a>
cannam@16 703 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 704 </p>
cannam@16 705 <p>Constant-Q Spectrogram calculates a spectrogram based on a short-time
cannam@16 706 windowed constant Q spectral transform. This is a spectrogram in
cannam@16 707 which the ratio of centre frequency to resolution is constant for each
cannam@16 708 frequency bin. The frequency bins correspond to the frequencies of
cannam@16 709 "musical notes" rather than being linearly spaced in frequency as they
cannam@16 710 are for the conventional DFT spectrogram.
cannam@16 711 </p>
cannam@16 712 <p>The pitch range and the number of frequency bins per octave may be
cannam@16 713 adjusted using the plugin's parameters. Note that the plugin's
cannam@16 714 preferred step and block sizes are defined by these parameters, and
cannam@16 715 the plugin will not accept any other block size than its preferred
cannam@16 716 value.
cannam@16 717 </p>
cannam@16 718 <h3>Parameters</h3>
cannam@16 719
cannam@16 720 <p><b>Minimum Pitch</b> &ndash; The MIDI pitch value (0-127) corresponding to the lowest
cannam@16 721 frequency to be included in the constant-Q transform.
cannam@16 722 </p>
cannam@16 723 <p><b>Maximum Pitch</b> &ndash; The MIDI pitch value (0-127) corresponding to the
cannam@16 724 lowest frequency to be included in the constant-Q transform.
cannam@16 725 </p>
cannam@16 726 <p><b>Tuning Frequency</b> &ndash; The frequency of concert A in the
cannam@16 727 music under analysis.
cannam@16 728 </p>
cannam@16 729 <p><b>Bins per Octave</b> &ndash; The number of constant-Q transform bins to be
cannam@16 730 computed per octave.
cannam@16 731 </p>
cannam@16 732 <p><b>Normalized</b> &ndash; Whether to normalize each output column to unit
cannam@16 733 maximum.
cannam@16 734 </p>
cannam@16 735 <h3>Outputs</h3>
cannam@16 736
cannam@16 737 <p><b>Constant-Q Spectrogram</b> &ndash; The calculated spectrogram, as a single
cannam@16 738 feature per process block containing one bin for each pitch included
cannam@16 739 in the spectrogram's range.
cannam@16 740 </p>
cannam@16 741 <h3>References and Credits</h3>
cannam@16 742
cannam@16 743 <p><b>Principle</b>: J. Brown. <i><a href="http://www.wellesley.edu/Physics/brown/pubs/cq1stPaper.pdf">Calculation of a constant Q spectral transform</a></i>. Journal of the Acoustical Society of America, 89(1):
cannam@16 744 425-434, 1991.
cannam@16 745 </p>
cannam@16 746 <p>The Constant-Q Spectrogram Vamp plugin was written by Christian
cannam@16 747 Landone.
cannam@16 748 </p>
cannam@29 749 <a name="qm-chromagram"></a><h2>12. Chromagram</h2>
cannam@16 750
cannam@16 751 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-chromagram</code>
cannam@16 752 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-chromagram">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-chromagram</a>
cannam@16 753 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 754 </p>
cannam@16 755 <p>Chromagram calculates a constant Q spectral transform (as in the
cannam@16 756 Constant Q Spectrogram plugin) and then wraps the frequency bin values
cannam@16 757 into a single octave, with each bin containing the sum of the
cannam@16 758 magnitudes from the corresponding bin in all octaves. The number of
cannam@16 759 values in each feature vector returned by the plugin is therefore the
cannam@16 760 same as the number of bins per octave configured for the underlying
cannam@16 761 constant Q transform.
cannam@16 762 </p>
cannam@16 763 <p>The pitch range and the number of frequency bins per octave for the
cannam@16 764 transform may be adjusted using the plugin's parameters. Note that
cannam@16 765 the plugin's preferred step and block sizes depend on these
cannam@16 766 parameters, and the plugin will not accept any other block size than
cannam@16 767 its preferred value.
cannam@16 768 </p>
cannam@16 769 <h3>Parameters</h3>
cannam@16 770
cannam@16 771 <p><b>Minimum Pitch</b> &ndash; The MIDI pitch value (0-127) corresponding to the
cannam@16 772 lowest frequency to be included in the constant-Q transform used in
cannam@16 773 calculating the chromagram.
cannam@16 774 </p>
cannam@16 775 <p><b>Maximum Pitch</b> &ndash; The MIDI pitch value (0-127) corresponding to the
cannam@16 776 lowest frequency to be included in the constant-Q transform used in
cannam@16 777 calculating the chromagram.
cannam@16 778 </p>
cannam@16 779 <p><b>Tuning Frequency</b> &ndash; The frequency of concert A in the
cannam@16 780 music under analysis.
cannam@16 781 </p>
cannam@16 782 <p><b>Bins per Octave</b> &ndash; The number of constant-Q transform bins to be
cannam@16 783 computed per octave, and thus the total number of bins present in the
cannam@16 784 resulting chromagram.
cannam@16 785 </p>
cannam@16 786 <p><b>Normalized</b> &ndash; Whether to normalize each output column. Normalization
cannam@16 787 may be to unit sum or unit maximum.
cannam@16 788 </p>
cannam@16 789 <h3>Outputs</h3>
cannam@16 790
cannam@16 791 <p><b>Chromagram</b> &ndash; The calculated chromagram, as a single feature per
cannam@16 792 process block containing the number of bins given in the bins per
cannam@16 793 octave parameter.
cannam@16 794 </p>
cannam@16 795 <h3>References and Credits</h3>
cannam@16 796
cannam@16 797 <p>The Chromagram Vamp plugin was written by Christian Landone.
cannam@16 798 </p>
cannam@29 799 <a name="qm-mfcc"></a><h2>13. Mel-Frequency Cepstral Coefficients</h2>
cannam@16 800
cannam@16 801 <p><b>System identifier</b> &ndash; <code>vamp:qm-vamp-plugins:qm-mfcc</code>
cannam@16 802 <br><b>RDF URI</b> &ndash; <a href="http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-mfcc">http://vamp-plugins.org/rdf/plugins/qm-vamp-plugins#qm-mfcc</a>
cannam@16 803 <br><b>Links</b> &ndash; <a href="#">Back to top of library documentation</a> &ndash; <a href="http://www.elec.qmul.ac.uk/digitalmusic/downloads/index.html#qm-vamp-plugins">Download location</a>
cannam@16 804 </p>
cannam@16 805 <p>Mel-Frequency Cepstral Coefficients calculates MFCCs from a single
cannam@16 806 channel of audio. These coefficients, derived from a cosine transform
cannam@16 807 of the mapping of an audio spectrum onto a frequency scale modelled on
cannam@16 808 human auditory response, are widely used in speech recognition, music
cannam@16 809 classification and other tasks.
cannam@16 810 </p>
cannam@16 811 <h3>Parameters</h3>
cannam@16 812
cannam@16 813 <p><b>Number of Coefficients</b> &ndash; The number of MFCCs to return. Commonly
cannam@16 814 used values include 13 or the default 20. This number includes C0 if
cannam@16 815 requested (see Include C0 below).
cannam@16 816 </p>
cannam@16 817 <p><b>Power for Mel Amplitude Logs</b> &ndash; An optional power value to which the
cannam@16 818 spectral amplitudes should be raised before applying the cosine
cannam@16 819 transform. Values greater than 1 may in principle reduce the
cannam@16 820 contribution of noise to the results. The default is 1.
cannam@16 821 </p>
cannam@16 822 <p><b>Include C0</b> &ndash; Whether to include the "zero'th" coefficient, which
cannam@16 823 simply reflects the overall signal power across the Mel frequency
cannam@16 824 bands.
cannam@16 825 </p>
cannam@16 826 <h3>Outputs</h3>
cannam@16 827
cannam@16 828 <p><b>Coefficients</b> &ndash; The MFCC values, returned as one vector feature per
cannam@16 829 processing block.
cannam@16 830 </p>
cannam@16 831 <p><b>Means of Coefficients</b> &ndash; The overall means of the MFCC bins, as a
cannam@16 832 single vector feature with time 0 that is returned when processing is
cannam@16 833 complete.
cannam@16 834 </p>
cannam@16 835 <h3>References and Credits</h3>
cannam@16 836
cannam@16 837 <p><b>MFCCs in music</b>: See B. Logan. <i><a href="http://ismir2000.ismir.net/papers/logan_paper.pdf">Mel-Frequency Cepstral Coefficients for Music Modeling</a></i>. In Proceedings of the First International
cannam@16 838 Symposium on Music Information Retrieval (ISMIR), 2000.
cannam@16 839 </p>
cannam@16 840 <p>The Mel-Frequency Cepstral Coefficients Vamp plugin was written by
cannam@16 841 Nicolas Chetry and Chris Cannam.
cannam@16 842 </p>
cannam@16 843 <p></p>
cannam@16 844 </CONTENTS>
cannam@16 845 </body>
cannam@16 846 </html>