cannam@255
|
1
|
cannam@255
|
2 Vamp Example Plugins
|
cannam@255
|
3 ====================
|
cannam@255
|
4
|
cannam@255
|
5 The vamp-example-plugins library contains a number of Vamp audio
|
cannam@255
|
6 analysis plugins provided as part of the Vamp plugin SDK.
|
cannam@255
|
7
|
cannam@255
|
8 These are simple, but sometimes useful, plugins whose source code you
|
cannam@255
|
9 are free to study and reuse in any proprietary or non-proprietary
|
cannam@255
|
10 plugins of your own without any licensing obligation.
|
cannam@255
|
11
|
cannam@255
|
12 User documentation for the individual plugins in this library follows.
|
cannam@255
|
13
|
cannam@255
|
14
|
cannam@255
|
15 Amplitude Follower
|
cannam@255
|
16 ==================
|
cannam@255
|
17
|
cannam@255
|
18 System identifier: vamp-example-plugins:amplitudefollower
|
cannam@255
|
19 RDF URI: http://vamp-plugins.org/rdf/plugins/vamp-example-plugins#amplitudefollower
|
cannam@255
|
20
|
cannam@255
|
21 Amplitude Follower tracks and returns the amplitude of the audio
|
cannam@255
|
22 signal, block by block. It uses a method from the SuperCollider audio
|
cannam@255
|
23 processing language, implemented as a Vamp plugin by Dan Stowell.
|
cannam@255
|
24
|
cannam@255
|
25 Parameters
|
cannam@255
|
26 ----------
|
cannam@255
|
27
|
cannam@255
|
28 Attack time (seconds)
|
cannam@255
|
29 Release time (seconds)
|
cannam@255
|
30
|
cannam@255
|
31 Outputs
|
cannam@255
|
32 -------
|
cannam@255
|
33
|
cannam@255
|
34 Amplitude
|
cannam@255
|
35 ~~~~~~~~~
|
cannam@255
|
36
|
cannam@255
|
37 The estimated peak amplitude (in volts) for the current processing block.
|
cannam@255
|
38
|
cannam@255
|
39
|
cannam@255
|
40 Simple Fixed Tempo Estimator
|
cannam@255
|
41 ============================
|
cannam@255
|
42
|
cannam@255
|
43 System identifier: vamp-example-plugins:fixedtempo
|
cannam@255
|
44 RDF URI: http://vamp-plugins.org/rdf/plugins/vamp-example-plugins#fixedtempo
|
cannam@255
|
45
|
cannam@255
|
46 Simple Fixed Tempo Estimator analyses a fragment of audio and
|
cannam@255
|
47 estimates its tempo. It assumes that its input is of fixed tempo, and
|
cannam@255
|
48 it analyses only the first (small but configurable number of) seconds
|
cannam@255
|
49 before returning a result, discarding all subsequent input.
|
cannam@255
|
50
|
cannam@255
|
51 The plugin calculates an overall energy rise function across a series
|
cannam@255
|
52 of short frequency-domain input frames, takes the autocorrelation of
|
cannam@255
|
53 this function, filters it to stress possible metrical patterns,
|
cannam@255
|
54 locates peaks, and converts from autocorrelation lag to the
|
cannam@255
|
55 corresponding tempo.
|
cannam@255
|
56
|
cannam@255
|
57 The filtering process involves searching for peaks at simple
|
cannam@255
|
58 metrically related intervals (at a given autocorrelation lag as well
|
cannam@255
|
59 as at 0.5, 2, and 4 times that lag), boosting each peak that shows
|
cannam@255
|
60 strong related peaks. A simplistic perceptual curve is also applied
|
cannam@255
|
61 in order to increase the probability of detecting a "likely" tempo.
|
cannam@255
|
62 For improved tempo precision, each tempo with strong related peaks is
|
cannam@255
|
63 averaged with the tempi calculated from those peaks.
|
cannam@255
|
64
|
cannam@255
|
65 The method is mainly tuned for 4/4 pop and dance rhythms.
|
cannam@255
|
66
|
cannam@255
|
67 This plugin returns many of its intermediate calculations as
|
cannam@255
|
68 additional outputs, as well as the most favoured tempo. Although as a
|
cannam@255
|
69 tempo estimator it's still fairly primitive, it is intended to provide
|
cannam@255
|
70 a useful example of a slightly more complex feature extraction plugin
|
cannam@255
|
71 than the other examples, as well as one that returns several different
|
cannam@255
|
72 types of output at a time.
|
cannam@255
|
73
|
cannam@255
|
74 Parameters
|
cannam@255
|
75 ----------
|
cannam@255
|
76
|
cannam@255
|
77 Minimum estimated tempo, Maximum estimated tempo (bpm) - These
|
cannam@255
|
78 parameters control the range of values within which the tempo
|
cannam@255
|
79 estimator will return its estimate.
|
cannam@255
|
80
|
cannam@255
|
81 Input duration to study (seconds) - The tempo estimator uses only the
|
cannam@255
|
82 first part of its input, discarding any that follows. This parameter
|
cannam@255
|
83 controls how much input it will use. There is no value in increasing
|
cannam@255
|
84 this beyond 8x the duration of the slowest returned beat. The default
|
cannam@255
|
85 of 10 seconds is likely to be appropriate for most purposes.
|
cannam@255
|
86
|
cannam@255
|
87 Outputs
|
cannam@255
|
88 -------
|
cannam@255
|
89
|
cannam@255
|
90 Tempo
|
cannam@255
|
91 ~~~~~
|
cannam@255
|
92
|
cannam@255
|
93 The tempo estimator's best guess at the tempo of its input, in beats
|
cannam@255
|
94 per minute.
|
cannam@255
|
95
|
cannam@255
|
96 This is returned as a feature whose timestamp and duration cover the
|
cannam@255
|
97 range of the input which was used in estimating the tempo, with a
|
cannam@255
|
98 single value containing the tempo.
|
cannam@255
|
99
|
cannam@255
|
100 Tempo candidates
|
cannam@255
|
101 ~~~~~~~~~~~~~~~~
|
cannam@255
|
102
|
cannam@255
|
103 Several guesses at the possible tempo. This output is returned as a
|
cannam@255
|
104 single feature whose timestamp and duration cover the range of the
|
cannam@255
|
105 input which was used in estimating the tempo, with up to 10 bins
|
cannam@255
|
106 containing one tempo value in each bin, with the "best guess" tempo in
|
cannam@255
|
107 bin 0.
|
cannam@255
|
108
|
cannam@255
|
109 Detection function
|
cannam@255
|
110 ~~~~~~~~~~~~~~~~~~
|
cannam@255
|
111
|
cannam@255
|
112 The basic onset detection function used in tempo estimation.
|
cannam@255
|
113
|
cannam@255
|
114 Autocorrelation function
|
cannam@255
|
115 ~~~~~~~~~~~~~~~~~~~~~~~~
|
cannam@255
|
116
|
cannam@255
|
117 The autocorrelation of the onset detection function.
|
cannam@255
|
118
|
cannam@255
|
119 Filtered Autocorrelation
|
cannam@255
|
120 ~~~~~~~~~~~~~~~~~~~~~~~~
|
cannam@255
|
121
|
cannam@255
|
122 The autocorrelation after filtering to boost values with possible
|
cannam@255
|
123 metrically related peaks and to apply perceptual weighting. The peak
|
cannam@255
|
124 value of this function is the one that will be used as the "best
|
cannam@255
|
125 guess".
|
cannam@255
|
126
|
cannam@255
|
127
|
cannam@255
|
128 Simple Percussion Onset Detector
|
cannam@255
|
129 ================================
|
cannam@255
|
130
|
cannam@255
|
131 System identifier: vamp-example-plugins:percussiononsets
|
cannam@255
|
132 RDF URI: http://vamp-plugins.org/rdf/plugins/vamp-example-plugins#percussiononsets
|
cannam@255
|
133
|
cannam@255
|
134 Simple Percussion Onset Detector estimates the locations of percussive
|
cannam@255
|
135 onsets in the audio signal. It uses a method described in "Drum
|
cannam@255
|
136 Source Separation using Percussive Feature Detection and Spectral
|
cannam@255
|
137 Modulation" by Dan Barry, Derry Fitzgerald, Eugene Coyle and Bob
|
cannam@255
|
138 Lawlor, ISSC 2005.
|
cannam@255
|
139
|
cannam@255
|
140 The principle is to exploit the broadband nature of noisy percussive
|
cannam@255
|
141 onsets by identifying only those frames in which the energy rise shows
|
cannam@255
|
142 a broadband profile.
|
cannam@255
|
143
|
cannam@255
|
144 The plugin takes a series of frequency domain frames, and examines
|
cannam@255
|
145 each frame to count the number of bins whose energy content has
|
cannam@255
|
146 increased by more than a certain threshold since the prior frame.
|
cannam@255
|
147 Frames in which this number is at a peak relative to prior and
|
cannam@255
|
148 following frames and also exceeds another threshold value are
|
cannam@255
|
149 classified as percussive onsets.
|
cannam@255
|
150
|
cannam@255
|
151 Parameters
|
cannam@255
|
152 ----------
|
cannam@255
|
153
|
cannam@255
|
154 Energy rise threshold (dB) - The rise in energy within a bin from one
|
cannam@255
|
155 frame to the next that is required for a bin to be counted toward the
|
cannam@255
|
156 detection function's bin count. This roughly corresponds to how
|
cannam@255
|
157 "loud" a percussive sound must be in order to be detected.
|
cannam@255
|
158
|
cannam@255
|
159 Sensitivity (%) - The proportion of bins that must exceed the energy
|
cannam@255
|
160 rise threshold in order for an onset to be detected (at frames in
|
cannam@255
|
161 which the detection function peaks). This roughly corresponds to how
|
cannam@255
|
162 "noisy" a percussive sound must be in order to be detected.
|
cannam@255
|
163
|
cannam@255
|
164 Outputs
|
cannam@255
|
165 -------
|
cannam@255
|
166
|
cannam@255
|
167 Onsets
|
cannam@255
|
168 ~~~~~~
|
cannam@255
|
169
|
cannam@255
|
170 The estimated onset locations.
|
cannam@255
|
171
|
cannam@255
|
172 Detection Function
|
cannam@255
|
173 ~~~~~~~~~~~~~~~~~~
|
cannam@255
|
174
|
cannam@255
|
175 The energy rise detection function whose peaks were used to estimate
|
cannam@255
|
176 onset locations.
|
cannam@255
|
177
|
cannam@255
|
178
|
cannam@255
|
179 Simple Power Spectrum
|
cannam@255
|
180 =====================
|
cannam@255
|
181
|
cannam@255
|
182 System identifier: vamp-example-plugins:powerspectrum
|
cannam@255
|
183 RDF URI: http://vamp-plugins.org/rdf/plugins/vamp-example-plugins#powerspectrum
|
cannam@255
|
184
|
cannam@255
|
185 Simple Power Spectrum returns a power spectrum calculated from
|
cannam@255
|
186 windowed short-time Fourier transforms of the input audio. (The power
|
cannam@255
|
187 spectrum for a frame consists of a sequence of the squares of the
|
cannam@255
|
188 magnitudes of the complex values for each frequency bin in the result
|
cannam@255
|
189 of the Fourier transform.)
|
cannam@255
|
190
|
cannam@255
|
191 This very simple plugin is an illustration of the fact that if a
|
cannam@255
|
192 plugin requests frequency-domain input, its input will already be in
|
cannam@255
|
193 the form needed for a spectrum such as this. The plugin has no work
|
cannam@255
|
194 left to do except to calculate the squared magnitude from the
|
cannam@255
|
195 cartesian complex representation.
|
cannam@255
|
196
|
cannam@255
|
197 This plugin also illustrates how to return "grid-type" visualisation
|
cannam@255
|
198 data from a Vamp plugin.
|
cannam@255
|
199
|
cannam@255
|
200 Parameters
|
cannam@255
|
201 ----------
|
cannam@255
|
202
|
cannam@255
|
203 None.
|
cannam@255
|
204
|
cannam@255
|
205 Outputs
|
cannam@255
|
206 -------
|
cannam@255
|
207
|
cannam@255
|
208 Power Spectrum
|
cannam@255
|
209 ~~~~~~~~~~~~~~
|
cannam@255
|
210
|
cannam@255
|
211 The power spectrum calculated from the input frame. This output
|
cannam@255
|
212 returns a single feature per processing block, containing
|
cannam@255
|
213 blocksize/2+1 power values corresponding to the FFT bins from DC to
|
cannam@255
|
214 Nyquist inclusive. The DC bin is always returned.
|
cannam@255
|
215
|
cannam@255
|
216
|
cannam@255
|
217 Spectral Centroid
|
cannam@255
|
218 =================
|
cannam@255
|
219
|
cannam@255
|
220 System identifier: vamp-example-plugins:spectralcentroid
|
cannam@255
|
221 RDF URI: http://vamp-plugins.org/rdf/plugins/vamp-example-plugins#spectralcentroid
|
cannam@255
|
222
|
cannam@255
|
223 Spectral Centroid calculates the "centre of gravity" of the frequency
|
cannam@255
|
224 spectrum for each input frame.
|
cannam@255
|
225
|
cannam@255
|
226 Parameters
|
cannam@255
|
227 ----------
|
cannam@255
|
228
|
cannam@255
|
229 None.
|
cannam@255
|
230
|
cannam@255
|
231 Outputs
|
cannam@255
|
232 -------
|
cannam@255
|
233
|
cannam@255
|
234 Log Frequency Centroid
|
cannam@255
|
235 ~~~~~~~~~~~~~~~~~~~~~~
|
cannam@255
|
236
|
cannam@255
|
237 The centroid of the log-weighted frequency spectrum. That is, the sum
|
cannam@255
|
238 across Fourier transform output bins of the logarithm of the bin
|
cannam@255
|
239 frequency multiplied by the bin magnitude, divided by the sum of the
|
cannam@255
|
240 bin magnitudes, and the inverse logarithm taken so as to give the
|
cannam@255
|
241 result as a frequency in Hz.
|
cannam@255
|
242
|
cannam@255
|
243 Linear Frequency Centroid
|
cannam@255
|
244 ~~~~~~~~~~~~~~~~~~~~~~~~~
|
cannam@255
|
245
|
cannam@255
|
246 The centroid of the linear-weighted frequency spectrum. That is, the
|
cannam@255
|
247 sum across Fourier transform output bins of the bin frequency
|
cannam@255
|
248 multiplied by the bin magnitude, divided by the sum of the bin
|
cannam@255
|
249 magnitudes. The result is a frequency in Hz.
|
cannam@255
|
250
|
cannam@255
|
251
|
cannam@255
|
252 Zero Crossings
|
cannam@255
|
253 ==============
|
cannam@255
|
254
|
cannam@255
|
255 System identifier: vamp-example-plugins:zerocrossing
|
cannam@255
|
256 RDF URI: http://vamp-plugins.org/rdf/plugins/vamp-example-plugins#zerocrossing
|
cannam@255
|
257
|
cannam@255
|
258 Zero Crossings calculates the positions and density of "zero-crossing"
|
cannam@255
|
259 points in an audio waveform. For the purposes of this plugin, that
|
cannam@255
|
260 means those positions at which the sampled value switches from
|
cannam@255
|
261 zero-or-less to greater-than-zero, or vice versa.
|
cannam@255
|
262
|
cannam@255
|
263 Parameters
|
cannam@255
|
264 ----------
|
cannam@255
|
265
|
cannam@255
|
266 None.
|
cannam@255
|
267
|
cannam@255
|
268 Outputs
|
cannam@255
|
269 -------
|
cannam@255
|
270
|
cannam@255
|
271 Zero Crossing Counts
|
cannam@255
|
272 ~~~~~~~~~~~~~~~~~~~~
|
cannam@255
|
273
|
cannam@255
|
274 The number of zero-crossing points found in the current block of
|
cannam@255
|
275 samples, as a single-valued feature returned per processing block.
|
cannam@255
|
276
|
cannam@255
|
277 Zero Crossings
|
cannam@255
|
278 ~~~~~~~~~~~~~~
|
cannam@255
|
279
|
cannam@255
|
280 The locations of zero-crossing points, returning one feature
|
cannam@255
|
281 timestamped to the zero-crossing location, without values, for each
|
cannam@255
|
282 crossing point.
|
cannam@255
|
283
|