musicassessr melodic features
melodic_features.RmdThe melodic features used in musicassessr mainly come
from the FANTASTIC library (Müllensiefen, 2009), except where
noted.
| Feature | Description | Equation | Reference |
|---|---|---|---|
| N | The length of the target melody. | - | - |
| freq | The count of its occurence in the corpus. | freq | - |
| rel_freq | The relative count of a frequency in the corpus | rel_freq | - |
| log_freq | The log of the relative count of a frequency in the corpus | - | - |
| IDF | The inverse document frequency | IDF | - |
| i.entropy | The average level of “information” or “surprise” in intervallic representations. Specifically, a variant of Shannon entropy on interval representations (Shannon, 1948) | \(- \frac{ \sum_{i} f_{i} \cdot \log_{2} f_{i}}{\log_{2} 139}\) | Müllensiefen, 2009 |
| span | The difference between the highest MIDI pitch and the lowest MIDI pitch in the melody. | span | span |
| tonalness | The magnitude of the highest correlation value in a vector of tonality correlation values computed with the Krumhansl-Schmuckler algorithm (Krumhansl, 1990). It expresses how strongly a melody correlates to a single key. | - | Müllensiefen, 2009 |
| tonal.clarity | Inspired by Temperley’s notion of tonal clarity (Temperley, 2007), the ratio between the magnitude of the highest correlation in a tonality.vector (see tonalness) A0 and the second highest correlation A1. | tonal.clarity | Müllensiefen, 2009 |
| tonal.spike | Similar to tonal.clarity, tonal.spike depends on the magnitude of the highest correlation but in contrast to the previous feature is divided by the sum of all correlation values > 0. | tonal.spike | Müllensiefen, 2009 |
| mode | The mode of the tonality with the highest correlation in the tonality.vector. It can assume the values major and minor. | mode | Müllensiefen, 2009 |
| step.cont.glob.var | step.cont.glob.var | step.cont.glob.var | Müllensiefen, 2009. |
| step.cont.glob.dir | step.cont.glob.dir | step.cont.glob.dir | Müllensiefen, 2009 |
| step.cont.loc.var | The mean absolute difference between adjacent values in the vector representing of step contour. | \(\frac{ \sum_{i=1}^{N-1} \lvert x_{i+1} - x_{i} \rvert }{N-1}\) | Müllensiefen, 2009 |
| mean_int_size | The average level of “information” or “surprise” in rhythm values. Specifically, a variant of Shannon entropy on rhythmic representations (Shannon, 1948) | mean_int_size | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| int_range | d.eq.trans | int_range | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| dir_change | The average interval size | dir_change | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| mean_dir_change | int_range. | mean_dir_change | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| int_variety | dir_change | int_variety | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| pitch_variety | mean_dir_change | pitch_variety | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| mean_run_length | int_variety | mean_run_length | Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021) |
| d.entropy | The average level of “information” or “surprise” in rhythm values. Specifically, a variant of Shannon entropy on rhythmic representations (Shannon, 1948) | \(- \frac{ \sum_{i} f_{i} \cdot \log_{2} f_{i}}{\log_{2} 140}\) | Müllensiefen, 2009 |
| d.eq.trans | mean_run_length | d.eq.trans | Müllensiefen, 2009 |
| mean_duration | mean_duration | mean_duration | mean_duration |
| mean_information_content | The average information content contained in the pitch values of a melody. Can be thought of as quantifying a melody’s self-similarity. | - | Harrison, Bianco, Chait & Pearce (2020) |
References
Beaty, R. E., Frieler, K., Norgaard, M., Merseal, H. M., MacDonald, M. C., & Weiss, D. J. (2021). Expert musical improvisations contain sequencing biases seen in language production. Journal of Experimental Psychology. https://doi.org/10.1037/xge0001107
Harrison, P. M. C., Bianco, R., Chait, M., & Pearce, M. T. (2020). PPM-Decay: A computational model of auditory prediction with memory decay. PLOS Computational Biology, 16(11), e1008304. https://doi.org/10.1371/journal.pcbi.1008304
Müllensiefen, D. (2009). FANTASTIC: Feature ANalysis Technology Accessing STatistics (In a Corpus; Technical report). 37.