Centre for Digital Music, Queen Mary, University of London.
This file copyright 2006-2008 Chris Cannam and QMUL.
- 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.
+ This file is part of vamp-aubio-plugins.
+
+ vamp-aubio 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 3 of the License, or
+ (at your option) any later version.
+
+ vamp-aubio is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with aubio. If not, see <http://www.gnu.org/licenses/>.
*/
#include <math.h>
#include "Notes.h"
+#include <algorithm>
+
using std::string;
using std::vector;
using std::cerr;
using std::endl;
-Notes::Notes(float inputSampleRate, unsigned int apiVersion) :
+Notes::Notes(float inputSampleRate) :
Plugin(inputSampleRate),
- m_apiVersion(apiVersion),
m_ibuf(0),
- m_fftgrain(0),
m_onset(0),
- m_pv(0),
- m_peakpick(0),
+ m_pitch(0),
m_onsetdet(0),
- m_onsettype(aubio_onset_complex),
+ m_onsettype(OnsetComplex),
m_pitchdet(0),
- m_pitchtype(aubio_pitch_yinfft),
- m_pitchmode(aubio_pitchm_freq),
+ m_pitchtype(PitchYinFFT),
m_threshold(0.3),
- m_silence(-90),
+ m_silence(-70),
+ m_minioi(4),
m_median(6),
- m_minpitch(27),
+ m_minpitch(32),
m_maxpitch(95),
m_wrapRange(false),
m_avoidLeaps(false),
m_prevPitch(-1)
{
- if (apiVersion == 1) {
- cerr << "vamp-aubio: WARNING: using compatibility version 1 of the Vamp API for note\n"
- << "tracker plugin: upgrade your host to v2 for proper duration support" << endl;
- } else {
- cerr << "vamp-aubio: NOTE: using v2 API for true durations" << endl;
- }
}
Notes::~Notes()
{
- if (m_onsetdet) aubio_onsetdetection_free(m_onsetdet);
- if (m_pitchdet) del_aubio_pitchdetection(m_pitchdet);
+ if (m_onsetdet) del_aubio_onset(m_onsetdet);
+ if (m_pitchdet) del_aubio_pitch(m_pitchdet);
if (m_ibuf) del_fvec(m_ibuf);
if (m_onset) del_fvec(m_onset);
- if (m_fftgrain) del_cvec(m_fftgrain);
- if (m_pv) del_aubio_pvoc(m_pv);
- if (m_peakpick) del_aubio_peakpicker(m_peakpick);
+ if (m_pitch) del_fvec(m_pitch);
}
string
int
Notes::getPluginVersion() const
{
- if (m_apiVersion == 1) return 2;
- return 3;
+ return 4;
}
string
bool
Notes::initialise(size_t channels, size_t stepSize, size_t blockSize)
{
- m_channelCount = channels;
+ if (channels != 1) {
+ std::cerr << "Notes::initialise: channels must be 1" << std::endl;
+ return false;
+ }
+
m_stepSize = stepSize;
m_blockSize = blockSize;
- size_t processingBlockSize;
- if (m_onsettype == aubio_onset_energy ||
- m_onsettype == aubio_onset_hfc) {
- processingBlockSize = stepSize * 2;
- } else {
- processingBlockSize = stepSize * 4;
- }
+ m_ibuf = new_fvec(stepSize);
+ m_onset = new_fvec(1);
+ m_pitch = new_fvec(1);
- m_ibuf = new_fvec(stepSize, channels);
- m_onset = new_fvec(1, channels);
- m_fftgrain = new_cvec(processingBlockSize, channels);
- m_pv = new_aubio_pvoc(processingBlockSize, stepSize, channels);
- m_peakpick = new_aubio_peakpicker(m_threshold);
+ reset();
- m_onsetdet = new_aubio_onsetdetection(m_onsettype, processingBlockSize, channels);
+ return true;
+}
- m_pitchdet = new_aubio_pitchdetection(processingBlockSize * 4,
- stepSize,
- channels,
- lrintf(m_inputSampleRate),
- m_pitchtype,
- m_pitchmode);
+void
+Notes::reset()
+{
+ if (m_onsetdet) del_aubio_onset(m_onsetdet);
+ if (m_pitchdet) del_aubio_pitch(m_pitchdet);
+
+ m_onsetdet = new_aubio_onset
+ (const_cast<char *>(getAubioNameForOnsetType(m_onsettype)),
+ m_blockSize,
+ m_stepSize,
+ lrintf(m_inputSampleRate));
+
+ aubio_onset_set_threshold(m_onsetdet, m_threshold);
+ aubio_onset_set_silence(m_onsetdet, m_silence);
+ aubio_onset_set_minioi(m_onsetdet, m_minioi);
+
+ m_pitchdet = new_aubio_pitch
+ (const_cast<char *>(getAubioNameForPitchType(m_pitchtype)),
+ m_blockSize,
+ m_stepSize,
+ lrintf(m_inputSampleRate));
+
+ aubio_pitch_set_unit(m_pitchdet, const_cast<char *>("freq"));
m_count = 0;
m_delay = Vamp::RealTime::frame2RealTime((4 + m_median) * m_stepSize,
m_currentOnset = Vamp::RealTime::zeroTime;
m_haveCurrent = false;
m_prevPitch = -1;
-
- return true;
-}
-
-void
-Notes::reset()
-{
}
size_t
ParameterDescriptor desc;
desc.identifier = "onsettype";
desc.name = "Onset Detection Function Type";
+ desc.description = "Type of onset detection function to use";
desc.minValue = 0;
- desc.maxValue = 6;
- desc.defaultValue = (int)aubio_onset_complex;
+ desc.maxValue = 7;
+ desc.defaultValue = (int)OnsetComplex;
desc.isQuantized = true;
desc.quantizeStep = 1;
desc.valueNames.push_back("Energy Based");
desc.valueNames.push_back("Phase Deviation");
desc.valueNames.push_back("Kullback-Liebler");
desc.valueNames.push_back("Modified Kullback-Liebler");
+ desc.valueNames.push_back("Spectral Flux");
list.push_back(desc);
desc = ParameterDescriptor();
desc.identifier = "pitchtype";
desc.name = "Pitch Detection Function Type";
+ desc.description = "Type of pitch detection function to use";
desc.minValue = 0;
desc.maxValue = 4;
- desc.defaultValue = (int)aubio_pitch_yinfft;
+ desc.defaultValue = (int)PitchYinFFT;
desc.isQuantized = true;
desc.quantizeStep = 1;
desc.valueNames.push_back("YIN Frequency Estimator");
desc = ParameterDescriptor();
desc.identifier = "minpitch";
desc.name = "Minimum Pitch";
+ desc.description = "Lowest pitch value to look for";
desc.minValue = 0;
desc.maxValue = 127;
desc.defaultValue = 32;
desc = ParameterDescriptor();
desc.identifier = "maxpitch";
desc.name = "Maximum Pitch";
+ desc.description = "Highest pitch value to look for";
desc.minValue = 0;
desc.maxValue = 127;
desc.defaultValue = 95;
desc = ParameterDescriptor();
desc.identifier = "wraprange";
desc.name = "Fold Higher or Lower Notes into Range";
+ desc.description = "Notes detected outside the range will be transposed to higher or lower octaves";
desc.minValue = 0;
desc.maxValue = 1;
desc.defaultValue = 0;
desc = ParameterDescriptor();
desc.identifier = "avoidleaps";
desc.name = "Avoid Multi-Octave Jumps";
+ desc.description = "Minimize octave jumps by transposing to the octave of the previously detected note";
desc.minValue = 0;
desc.maxValue = 1;
desc.defaultValue = 0;
desc = ParameterDescriptor();
desc.identifier = "peakpickthreshold";
desc.name = "Peak Picker Threshold";
+ desc.description = "Peak picking threshold, the higher the least detection";
desc.minValue = 0;
desc.maxValue = 1;
desc.defaultValue = 0.3;
desc = ParameterDescriptor();
desc.identifier = "silencethreshold";
desc.name = "Silence Threshold";
+ desc.description = "Silence threshold, the higher the least detection";
desc.minValue = -120;
desc.maxValue = 0;
- desc.defaultValue = -90;
+ desc.defaultValue = -70;
desc.unit = "dB";
desc.isQuantized = false;
list.push_back(desc);
+ desc = ParameterDescriptor();
+ desc.identifier = "minioi";
+ desc.name = "Minimum Inter-Onset Interval";
+ desc.description = "Time interval below which two consecutive onsets should be merged";
+ desc.minValue = 0;
+ desc.maxValue = 40;
+ desc.defaultValue = 4;
+ desc.unit = "ms";
+ desc.isQuantized = true;
+ desc.quantizeStep = 1;
+ list.push_back(desc);
+
return list;
}
return m_wrapRange ? 1.0 : 0.0;
} else if (param == "avoidleaps") {
return m_avoidLeaps ? 1.0 : 0.0;
+ } else if (param == "minioi") {
+ return m_minioi;
} else {
return 0.0;
}
{
if (param == "onsettype") {
switch (lrintf(value)) {
- case 0: m_onsettype = aubio_onset_energy; break;
- case 1: m_onsettype = aubio_onset_specdiff; break;
- case 2: m_onsettype = aubio_onset_hfc; break;
- case 3: m_onsettype = aubio_onset_complex; break;
- case 4: m_onsettype = aubio_onset_phase; break;
- case 5: m_onsettype = aubio_onset_kl; break;
- case 6: m_onsettype = aubio_onset_mkl; break;
+ case 0: m_onsettype = OnsetEnergy; break;
+ case 1: m_onsettype = OnsetSpecDiff; break;
+ case 2: m_onsettype = OnsetHFC; break;
+ case 3: m_onsettype = OnsetComplex; break;
+ case 4: m_onsettype = OnsetPhase; break;
+ case 5: m_onsettype = OnsetKL; break;
+ case 6: m_onsettype = OnsetMKL; break;
+ case 7: m_onsettype = OnsetSpecFlux; break;
}
} else if (param == "pitchtype") {
switch (lrintf(value)) {
- case 0: m_pitchtype = aubio_pitch_yin; break;
- case 1: m_pitchtype = aubio_pitch_mcomb; break;
- case 2: m_pitchtype = aubio_pitch_schmitt; break;
- case 3: m_pitchtype = aubio_pitch_fcomb; break;
- case 4: m_pitchtype = aubio_pitch_yinfft; break;
+ case 0: m_pitchtype = PitchYin; break;
+ case 1: m_pitchtype = PitchMComb; break;
+ case 2: m_pitchtype = PitchSchmitt; break;
+ case 3: m_pitchtype = PitchFComb; break;
+ case 4: m_pitchtype = PitchYinFFT; break;
}
} else if (param == "peakpickthreshold") {
m_threshold = value;
m_wrapRange = (value > 0.5);
} else if (param == "avoidleaps") {
m_avoidLeaps = (value > 0.5);
+ } else if (param == "minioi") {
+ m_minioi = value;
}
}
OutputDescriptor d;
d.identifier = "notes";
d.name = "Notes";
+ d.description = "List of notes detected, with their frequency and velocity";
d.unit = "Hz";
d.hasFixedBinCount = true;
- if (m_apiVersion == 1) {
- d.binCount = 3;
- d.binNames.push_back("Frequency");
- d.binNames.push_back("Duration");
- d.binNames.push_back("Velocity");
- } else {
- d.binCount = 2;
- d.binNames.push_back("Frequency");
- d.binNames.push_back("Velocity");
- d.hasDuration = true;
- }
+ d.binCount = 2;
+ d.binNames.push_back("Frequency");
+ d.binNames.push_back("Velocity");
+ d.hasDuration = true;
d.hasKnownExtents = false;
d.isQuantized = false;
Notes::process(const float *const *inputBuffers, Vamp::RealTime timestamp)
{
for (size_t i = 0; i < m_stepSize; ++i) {
- for (size_t j = 0; j < m_channelCount; ++j) {
- fvec_write_sample(m_ibuf, inputBuffers[j][i], j, i);
- }
+ fvec_set_sample(m_ibuf, inputBuffers[0][i], i);
}
- aubio_pvoc_do(m_pv, m_ibuf, m_fftgrain);
- aubio_onsetdetection(m_onsetdet, m_fftgrain, m_onset);
-
- bool isonset = aubio_peakpick_pimrt(m_onset, m_peakpick);
+ aubio_onset_do(m_onsetdet, m_ibuf, m_onset);
+ aubio_pitch_do(m_pitchdet, m_ibuf, m_pitch);
- float frequency = aubio_pitchdetection(m_pitchdet, m_ibuf);
+ bool isonset = m_onset->data[0];
+ float frequency = m_pitch->data[0];
m_notebuf.push_back(frequency);
if (m_notebuf.size() > m_median) m_notebuf.pop_front();
feature.timestamp = m_currentOnset - m_delay;
feature.values.push_back(freq);
- if (m_apiVersion == 1) {
- feature.values.push_back
- (Vamp::RealTime::realTime2Frame
- (offTime, lrintf(m_inputSampleRate)) -
- Vamp::RealTime::realTime2Frame
- (m_currentOnset, lrintf(m_inputSampleRate)));
- feature.hasDuration = false;
- } else {
- feature.hasDuration = true;
- feature.duration = offTime - m_currentOnset;
- }
+ feature.values.push_back
+ (Vamp::RealTime::realTime2Frame
+ (offTime, lrintf(m_inputSampleRate)) -
+ Vamp::RealTime::realTime2Frame
+ (m_currentOnset, lrintf(m_inputSampleRate)));
+ feature.hasDuration = false;
feature.values.push_back(m_currentLevel);
fs[0].push_back(feature);