Repository navigation

Core Module Interactions

Analyzer

Input:

list of songs:
- directory containing audio files
- parses directory and gets all existing metadata from songs

Output:

unordered list of song objects:
- song_analysis_object:
  - file_path (relative : string)
  - [prop_1, prop_2, ... , prop_N]
    - tempo (BPM : float)
    - key (Camelot : string)
    - mood (Scale : float)
    - energy (Scale : float)
  - [sect_1, sect_2, ... , sect_M]
    - section:
      - start_sample (range : long)
      - end_sample (range : long)
      - array_of_beats: [sample_num_1, sample_num_2, ... , sample_num_N] (array : long)
      - key (Camelot : string)
      - tempo (BPM : float)
      - 4_beat_avg_amplitude_start (0 - 1 : float)
      - 4_beat_avg_amplitude_end (0 - 1 : float)
      - avg_amplitude (0 - 1 : float)
      - section_classification (chorus, verse, build-up, drop, etc... : string)
        
        determined by grouping average frequency of 4 beat sections

Optimizer

Input:

output from analyzer

Output:

ordered list of song objects (size : N):
- song_analysis_object
ordered list of transition dicts (size : N - 1) [{transition from A to B}, {transition from B to C}, ... ,{}]
- song_a Track a of transition, analyzed Song object.
- song_b Track b of transition, analyzed Song object.
- start_a index of track A start beat (index : integer)
- start_b index of track B start beat (index : integer)
- sections [{section 0}, {section 1}, ... , {}]
  - section:
    - offset offset of the section start in beats from start_a/star_b intersection (offset : integer)
    - length number of beats in the section (length : integer)
    - type transitions applied to that section (name of transition : string) --> [transition_name_1, transition_name_2, ... , transition_name_N]

Composer

Input:

output from analyzer

Output:

MIXTAPE (an audio file)