Filosax_Lite

The Filosax dataset was conceived, curated and compiled by Dave Foster (a PhD student on the AIM programme at QMUL) and his supervisor Simon Dixon (C4DM @ QMUL). The full dataset is a collection of 48 multitrack jazz recordings, where each piece has 8 corresponding audio files: 1) The original Aebersold backing track (stereo) - which needs to be purchased separately 2) Bass_Drums, a mono file of a mix of bass and drums 3) Piano_Drums, a mono file of a mix of piano and drums 4) Participant 1 Sax, a mono file of solo saxophone 5) Participant 2 Sax, a mono file of solo saxophone 6) Participant 3 Sax, a mono file of solo saxophone 7) Participant 4 Sax, a mono file of solo saxophone 8) Participant 5 Sax, a mono file of solo saxophone This "lite" version has only 2 tracks and 2 participant recordings for each. For each piece, there is a corresponding .jams file containing piece-level annotations: 1) Beat annotation for the start of each bar and any mid-bar chord change 2) Chord annotation for each bar, and mid-bar chord change 3) Section annotation for when the solo changes between the 3 categories: a) head (melody) b) written solo (interpretation of transcribed solo) c) improvised solo For each Sax recording, there is a corresponding .json file containing note annotations: 1) (float) a_start_time, a_end_time, a_duration: the time stamps of the start, end and duration of the note, in seconds 2) (int) piece_num and bar_num: the parent piece and bar number of the start of the note 3) (string) musician: the participant ID 4) (int) bar_type: the section annotation where 0 = head, 1 = written solo, 2 = improvised solo 5) (float) s_start_time and s_duration: the time stamps of the score representation, in seconds 6) (int) crochet_num: the number of sub-divisions which define a crochet (always 24) 7) (int) s_rhythmic_position and s_rhythmic_duration: the start and duration of the score note (compared to crochet_num) 8) (int) midi_pitch: the quantised midi pitch 9) (bool) is_grace: is the note a grace note, associated with the following note 10) (int) num_chord_changes: the number of chords which accompany the note (usually 1, sometimes >1 for long notes) 11) (dict{int, [int, int, int]}) chord changes: the chords, where the key is the rhythmic position of the chord (using crochet_num) and the value a list [a, b, c], where: a: chord_root (0 = C, 1 = C#... 11 = B) b: chord_bass (0 = C, 1 = C#... 11 = B) c: chord_type (0 = maj7, 1 = maj7#11, 2 = 6/9, 3 = 7, 4 = 7b9, 5 = 7#11, 6 = min7, 7 = min(maj7), 8 = dim, 9 = 1/2 dim) (An additional chord is added in the case of a quaver at the end of the bar, followed by a rest on the downbeat) 12) (int) main_chord_num: usually 0, sometimes 1 in the quaver case described above 13) ([int]) scale_changes: the degree of the chromatic scale when midi_pitch is compared to chord_root 14) (float) loudness_max_val and loudness_max_time: the value (db) and time (seconds) of the maximum loudness (could be used for onset value) 15) ([float]) loudness_curve: the inter-note loudness values, 1 per millisecond 16) (float) pitch_average_val and pitch_average_time: the value (midi) and time (seconds) of the average pitch 17) ([float]) pitch_curve: the inter-note pitch values, 1 per millisecond 18) (float) pitch_vib_freq and pitch_vib_ext: the vibrato frequency (Hz) and extent (midi), both 0.0 if no vibrato detected 19) (float) spec_cent and spec_flux: the spectral centroid and spectral flux values at the time of the maximum loudness 20) ([float]) spec_cent_curve and spec_flux_curve: the inter-note timbre values, 1 per millisecond The Participant folders also contain MIDI files of the transcriptions (frame level and score level) as well as a PDF and MusicXML of the typeset solo.