def _miniaudio_mp3read(path: str, start=0, end=0) -> np.ndarray:
"""
Reads a mp3 files completely into an array
start, end: if given, in seconds
"""
import miniaudio
decoded = miniaudio.mp3_read_file_f32(path)
npsamples = np.frombuffer(decoded.samples, dtype='float32')
sr = decoded.sample_rate
if decoded.nchannels > 1:
npsamples.shape = (decoded.num_frames, decoded.nchannels)
if start > 0 or end != 0:
startframe = int(start * sr)
if end == 0:
endframe = decoded.num_frames - 1
elif end > 0:
endframe = int(end * sr)
else:
endframe = int((decoded.duration + end) * sr)
if endframe < startframe:
raise ValueError(
f"startframe ({startframe}) > endframe ({endframe})")
npsamples = npsamples[startframe:endframe + 1]
npsamples = npsamples.astype(float)
return npsamples, sr
def read_mp3(path, resample_rate=16000):
if isinstance(path, bytes):
# If path is a tf.string tensor, it will be in bytes
path = path.decode("utf-8")
f = miniaudio.mp3_read_file_f32(path)
# Downsample to target rate, 16 kHz is commonly used for speech data
new_len = round(len(f.samples) * float(resample_rate) / f.sample_rate)
signal = scipy.signal.resample(f.samples, new_len)
# Normalize to [-1, 1]
signal /= np.abs(signal).max()
return signal, resample_rate
import scipy.signal
url = "https://file-examples-com.github.io/uploads/2017/11/file_example_MP3_700KB.mp3"
signal, resample_rate = librosa.load("file_example_MP3_700KB.mp3")
u = urlopen(url)
import miniaudio
import librosa
from src.features import plot_signal
import matplotlib.pyplot as plt
plot_signal(signal)
plt.show()
signal.max()
f = miniaudio.mp3_read_file_f32("file_example_MP3_700KB.mp3")
f.sample_rate
f.nchannels
f.samples
play(signal, 16000)
sd.stop()
len(signal)
873216/32000
pyaud = pyaudio.PyAudio()
srate=32000
stream = pyaud.open(format = pyaud.get_format_from_width(1),
channels = 2,
rate = srate,
def miniaudio_read_mp3(path):
audio = miniaudio.mp3_read_file_f32(path.decode("utf-8"))
return np.array(audio.samples, np.float32).reshape(
(-1, audio.nchannels)), audio.sample_rate