개강 8,9주차 진척 상황

1. 저번에 템포 분석을 했는데, 암만 생각해도 너무 부정확한 것 같아서 뭔가 정확도를 높이려고 좋은 이론 베이스 자료를 찾아서 좀 공부해봄.

Challenges in beat tracking

Pulse level often unclear  Local/sudden tempo changes (e.g. rubato)  Vague information (e.g., soft onsets, extracted onsets corrupt)  Sparse information (often only note onsets are used)

 

Onset is the time position where a note is played

Onset typically goes along with a change of the signal’s properties: – energy or loudness – pitch or harmony – timbre

와!!! Onset Detection (Energy-Based) Example: C4 played by piano 이부분 보니까 그냥 완전 개념 바로 이해갔음!!!!28쪽!~!!

 

No energy increase may be observable in complex sound mixtures  More refined methods needed that capture – changes of spectral content – changes of pitch – changes of harmony

 

Beat and Tempo  Steady pulse that drives music forward and provides the temporal framework of a piece of music  Sequence of perceived pulses that are equally spaced in time  The pulse a human taps along when listening to the music [Parncutt 1994] [Sethares 2007] [Large/Palmer 2002] [Lerdahl/ Jackendoff 1983] [Fitch/ Rosenfeld 2007] What is a beat? The term tempo then refers to the speed of the pulse

 

Beat and Tempo 분석 전략  Analyze the novelty curve with respect to reoccurring or quasiperiodic patterns  Avoid the explicit determination of note onsets (no peak picking)

Methods: Comb-filter methods  Autocorrelation  Fourier transfrom

 

A tempogram is a time-tempo representation that encodes the local tempo of a music signal over time.

- Fourier-based method:  Compute a spectrogram (STFT) of the novelty curve  Convert frequency axis (given in Hertz) into tempo axis (given in BPM)  Magnitude spectrogram indicates local tempo

- Autocorrelation-based method: Compare novelty curve with time-lagged local sections of itself  Convert lag-axis (given in seconds) into tempo axis (given in BPM)  Autocorrelogram indicates local tempo

 

 

2. 그리고 나는 특징별로 이제 데이터는 다 추출을 했는데, 도대체가 이걸 무슨 퉁계지표로 분석을 해야하냐??? 뭐 평균이야 최빈이야 뭐야???

이 자료를 못찾으면 그냥 내 맘대로 임의로 그런 평균, 최빈 이런 대표젹 기술통계값을 도출해내는 수밖에 없고,

그게 아니라면 데이터를 찾아야만 함.