TimeSignatureEstimator/mfcc_data_extraction.py at main · pianistprogrammer/TimeSignatureEstimator · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
from operator import index
import os
import librosa
import math
import json

DATASET_PATH = "Data/genres_original"
JSON_PATH = "data.json"
sample_rate = 22050
duration = 30  # mesaured in seconds
samples_per_track = sample_rate * duration


def extract_mfcc(dataset_path, json_path, n_mfcc=13, hop_length=512, n_fft=2048, num_segments=5):
    # dictionary to store the data
    data = {
        "mapping": [],
        "mfcc": [],
        "labels": []
    }
    num_samples_per_segment = int(samples_per_track / num_segments)
    expected_num_mfcc_vectors_per_segment = math.ceil(
        num_samples_per_segment / hop_length)

    # loop through all the genres
    for i, (dirpath, dirnames, filenames) in enumerate(os.walk(dataset_path)):

        if dirpath is not dataset_path:
            dirpath_components = dirpath.split("/")
            semantic_label = dirpath_components[-1]
            data["mapping"].append(semantic_label)
            print("\nProcessing {}".format(semantic_label))

            # looop through the audio files
            for f in filenames:
                file_path = os.path.join(dirpath, f)
                signal, sr = librosa.load(file_path, sr=sample_rate)

                # process mfcc for each window segment
                for s in range(num_segments):
                    start_sample = num_samples_per_segment * s
                    finish_sample = start_sample + num_samples_per_segment

                    mfcc = librosa.feature.mfcc(signal[start_sample:finish_sample],
                                                sr=sr,
                                                n_mfcc=n_mfcc,
                                                hop_length=hop_length,
                                                n_fft=n_fft)
                    mfcc = mfcc.T
                    if len(mfcc) == expected_num_mfcc_vectors_per_segment:
                        data["mfcc"].append(mfcc.tolist())
                        data["labels"].append(i-1)
                        print("{}, segment:{}".format(file_path, s))

    with open(json_path, "w") as fp:
        json.dump(data, fp, indent=4)


if __name__ == "__main__":
    extract_mfcc(DATASET_PATH, JSON_PATH, num_segments=10)