def emotions(file1):
#read the audio and create the time array (timeline)
audio, sfreq = lr.load(
file1
) #.load an audio file and decodes it into a 1-dimensional which is a time series AUDIO, and sfreq is a sampling
#rate fo AUDIO , by default is sfreq = 22kHz.
time = np.arange(0, len(audio)) / sfreq
#plot audio over time
#determine the duration of the signal
dur_sec = time.shape[0] / sfreq
#calculate information extra related with the audio
out = biosppy.signals.ecg.ecg(audio, sfreq, show=True)
fig, ax = plt.subplots()
ax.plot(time, out[1])
ax.set(xlabel='Time(s)', ylabel='Sound Amplitude')
plt.show()
#get melspectogram of the audio
spec = lr.feature.melspectrogram(y=audio, sr=sfreq)
# #Separate the initial audio series into harmonic and percussive components
data_h, data_p = lr.effects.hpss(audio)
spec_h = lr.feature.melspectrogram(data_h, sr=sfreq)
spec_p = lr.feature.melspectrogram(data_p, sr=sfreq)
chroma = lr.feature.chroma_cqt(y=data_h, sr=sfreq)
plt.figure(figsize=(18, 5))
lr.display.specshow(chroma,
sr=sfreq,
x_axis='time',
y_axis='chroma',
vmin=0,
vmax=1)
plt.figure(figsize=(20, 8))
plt.title('Audio: Chroma Spectrogram')
lr.display.specshow(chroma, sr=sfreq, x_axis='s', y_axis='chroma')
#how fast is the song, estima el tempo de la muetra del audioe n latidos por minuto
tempo = lr.beat.tempo(audio, sfreq)
#resample the signal
from gtda.time_series import Resampler
periodic_sampler = Resampler(period=1000)
y_k = periodic_sampler.fit_transform(audio)
fig, ax = plt.subplots()
ax.plot(time[0:len(y_k)], y_k)
datDelayInformation = get_parameters(Signal)
# embedded in two dimensions
embedder = takensEmbedding(y_k, datDelayInformation[0],
datDelayInformation[1])
return embedder, dur_sec, tempo
def test_resampler_resample(period, expected):
resampler = Resampler(period=period)
resampler.fit(signal_array)
assert_almost_equal(resampler.resample(signal_array), expected)
def test_resampler_transform(period, expected):
resampler = Resampler(period=period)
assert_almost_equal(resampler.fit_transform(signal_array), expected)
def test_resampler_errors():
resampler = Resampler(period='blah')
with pytest.raises(TypeError):
resampler.fit(signal_array)