librosa & soundfile & ffmpeg
===
Google Colab:
https://colab.research.google.com/drive/1yL9xmNR0iurtKI86UMIkMllhpjrUkxrg?usp=sharing

Install library (ex. .mp3 support):
https://github.com/librosa/librosa#hints-for-the-installation

Soundfile:
---
https://github.com/librosa/librosa#hints-for-the-installation

- installation:
- pip install soundfile
- or (sudo apt-get install libsndfile1)
- import:
- import soundfile as sf
- Load audio:
- data, samplerate = sf.read('name.wav')
- Write audio:
- sf.write('new.flac', data, samplerate)
- Information:
- sf.info('name.wav')
- Read short blocks:
> rms = [np.sqrt(np.mean(block**2)) for block in sf.blocks('name.wav', blocksize=1024, overlap=512)]

- load mp3?
- ffmpeg below:

ffmpeg:
---
- sudo apt-get install ffmpeg
- mp3 -> wav
> ffmpeg -i Heros.mp3 -vn -acodec pcm_s16le -ac 1 -ar 44100 -f wav Heros.wav
- wav -> mp3
> mpg123 -w foo.wav foo.mp3

librosa:
---
Installation: https://github.com/librosa/librosa#hints-for-the-installation
**Document**: https://librosa.org/doc/latest/core.html

- import:
- impoer librosa
- import librosa.display
- Load audio:
- y, sr = librosa.load(filename)
- y: waveform
- sr: sampling rate
- Output audio:
with soundfile:
- sf.write('test01.wav', y, samplerate=sr)
- Beat Tracker:
- tempo, beat_frames = librosa.beat.beat_track(y=y, sr=sr)
- tempo: beat per minute
- beat_frames: type 'beat_frames'
- Frame to time
- librosa.frames_to_time(beat_frames, sr=sr)
- Separate harmonics and percussives

    # Separate harmonics and percussives into two waveforms
    y_harmonic, y_percussive = librosa.effects.hpss(y)

- Compute Feature

    # Compute MFCC features from the raw signal
    mfcc = librosa.feature.mfcc(y=y, sr=sr, hop_length=512, n_mfcc=13)

    # And the first-order differences (delta features)
    mfcc_delta = librosa.feature.delta(mfcc)

- Stack and synchronize between beat events

    # Stack and synchronize between beat events
    ## This time, we'll use the mean value (default) instead of median
    beat_mfcc_delta = librosa.util.sync(np.vstack([mfcc, mfcc_delta]),
                                        beat_frames)

- Compute chroma features from the harmonic signal

    # Compute chroma features from the harmonic signal
    chromagram = librosa.feature.chroma_cqt(y=y_harmonic,
                                        sr=sr)

- Aggregate chroma features between beat events

    # Aggregate chroma features between beat events
    # We'll use the median value of each feature between beat frames
    beat_chroma = librosa.util.sync(chromagram,
                                    beat_frames,
                                    aggregate=np.median)

- Stack

    # Finally, stack all beat-synchronous features together
    beat_features = np.vstack([beat_chroma, beat_mfcc_delta])

- ----
- librosa.core
- librosa
- beat 偵測節奏
- decompose 消除背景(HPSS)(Scikit Learn)
- display 畫圖( matplotlib )
- effects 編輯效果啊
- feature 低階特徵偵測
> chromagrams, Mel, MFCC, and other spectral and rhythmic features.
- filters
- onset 偵測onset(開始？)
- segment 切割(segmentation)
- sequence // Funcitons for sequential modeling
- util 工具：normalization, padding, centering, etc.
- ----
- Examples

# Beat tracking example
import librosa

# 1. Get the file path to an included audio example
filename = librosa.example('nutcracker')


# 2. Load the audio as a waveform `y`
#    Store the sampling rate as `sr`
y, sr = librosa.load(filename)

# 3. Run the default beat tracker
tempo, beat_frames = librosa.beat.beat_track(y=y, sr=sr)

print('Estimated tempo: {:.2f} beats per minute'.format(tempo))

# 4. Convert the frame indices of beat events into timestamps
beat_times = librosa.frames_to_time(beat_frames, sr=sr)