def read_single_wav(self, path):
clip = AudioFileClip(str(path))
subsampled_audio = clip.set_fps(16000)
chunk_size = 640
audio = np.array(list(subsampled_audio.iter_frames())).mean(1)
audio = np.pad(audio, (0, chunk_size - audio.shape[0] % chunk_size),
'constant')
audio = audio.reshape(-1, chunk_size)
return audio.astype(np.float32)
def get_samples(subject_id):
arousal_label_path = root_dir / 'Ratings_affective_behaviour_CCC_centred/arousal/{}.csv'.format(subject_id)
valence_label_path = root_dir / 'Ratings_affective_behaviour_CCC_centred/valence/{}.csv'.format(subject_id)
clip = AudioFileClip(str(root_dir / "Audio_recordings_WAV/{}.wav".format(subject_id)))
subsampled_audio = clip.set_fps(16000)
audio_frames = []
for i in range(1, 7501):
time = 0.04 * i
audio = np.array(list(subsampled_audio.subclip(time - 0.04, time).iter_frames())) #count the number of audio frames
audio = audio.mean(1)[:640] # takes the mean along axis=1 (along the row)
audio_frames.append(audio.astype(np.float32))
arousal = np.loadtxt(str(arousal_label_path), delimiter=';', skiprows=1)[:, 1][1:]
valence = np.loadtxt(str(valence_label_path), delimiter=';', skiprows=1)[:, 1][1:]
return audio_frames, np.dstack([arousal, valence])[0].astype(np.float32)
