From: Acoustical feature analysis and optimization for aesthetic recognition of Chinese traditional music
Category | Feature name | Dimensions | Extraction tool |
---|---|---|---|
Energy | Low energy frame ratio | 3 | Mir toolbox |
RMS energy, intensity, loudness | 6 | OpenSMILE | |
Pitch | Fundamental frequency, fundamental frequency | 6 | OpenSMILE |
By smoothing | |||
Timbre | MFCC, spectral flatness measure, spectral crest factor | 244 | Marsyas |
Linear prediction coefficient, line-spectral pair, spectral slope, spectral entropy, zero-crossing rate, spectral centroid, spectral flux, spectral roll-off, sharpness, spectral harmony, spectral spread, mid-low frequency energy ratio | 93 | OpenSMILE | |
Rhythm | Fluctuation, beatspectrum, events, eventdensity, tempo, pulseclarity | 18 | MirToolbox |
Mode | Key strength | 3 | Mirtoolbox |
Chroma | 56 | Marsyas | |
Octave-based subband energy | 15 | OpenSMILE | |
Total | Â | 447 | Â |