CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutPoliciesSign UpSign In
y33-j3T

Real-time collaboration for Jupyter Notebooks, Linux Terminals, LaTeX, VS Code, R IDE, and more,
all in one place. Commercial Alternative to JupyterHub.

GitHub Repository: y33-j3T/Coursera-Deep-Learning
 Path: blob/master/Sequence Models/Week 3/Trigger word detection/td_utils.py
Views: 13377
1
import matplotlib.pyplot as plt

2
from scipy.io import wavfile

3
import os

4
from pydub import AudioSegment

5


6
# Calculate and plot spectrogram for a wav audio file

7
def graph_spectrogram(wav_file):

8
    rate, data = get_wav_info(wav_file)

9
    nfft = 200 # Length of each window segment

10
    fs = 8000 # Sampling frequencies

11
    noverlap = 120 # Overlap between windows

12
    nchannels = data.ndim

13
    if nchannels == 1:

14
        pxx, freqs, bins, im = plt.specgram(data, nfft, fs, noverlap = noverlap)

15
    elif nchannels == 2:

16
        pxx, freqs, bins, im = plt.specgram(data[:,0], nfft, fs, noverlap = noverlap)

17
    return pxx

18


19
# Load a wav file

20
def get_wav_info(wav_file):

21
    rate, data = wavfile.read(wav_file)

22
    return rate, data

23


24
# Used to standardize volume of audio clip

25
def match_target_amplitude(sound, target_dBFS):

26
    change_in_dBFS = target_dBFS - sound.dBFS

27
    return sound.apply_gain(change_in_dBFS)

28


29
# Load raw audio files for speech synthesis

30
def load_raw_audio():

31
    activates = []

32
    backgrounds = []

33
    negatives = []

34
    for filename in os.listdir("./raw_data/activates"):

35
        if filename.endswith("wav"):

36
            activate = AudioSegment.from_wav("./raw_data/activates/"+filename)

37
            activates.append(activate)

38
    for filename in os.listdir("./raw_data/backgrounds"):

39
        if filename.endswith("wav"):

40
            background = AudioSegment.from_wav("./raw_data/backgrounds/"+filename)

41
            backgrounds.append(background)

42
    for filename in os.listdir("./raw_data/negatives"):

43
        if filename.endswith("wav"):

44
            negative = AudioSegment.from_wav("./raw_data/negatives/"+filename)

45
            negatives.append(negative)

46
    return activates, negatives, backgrounds

47