def get_wav_info(wav_file):
wav = wave.open(wav_file, 'r')
frames = wav.readframes(-1)
sound_info = pylab.frombuffer(frames, 'int16')
frame_rate = wav.getframerate()
wav.close()
return sound_info, frame_rate
def loadMNISTImages(filename):
f = open(filename, 'rb')
# Verify Magic Number
s = f.read(4)
magic = int(s.encode('hex'),16)
assert(magic == 2051)
# Get Number of Images
s = f.read(4)
numImages = int(s.encode('hex'),16)
s = f.read(4)
numRows = int(s.encode('hex'),16)
s = f.read(4)
numCols = int(s.encode('hex'),16)
# Get Data
s = f.read()
a = frombuffer(s, uint8)
# Use 'F' to ensure that we read by column
a = reshape(a, (numCols , numRows, numImages), order='F');
images = transpose(a, (1, 0, 2))
f.close()
# Reshape to #pixels * #examples
images = reshape(a, (shape(images)[0] * shape(images)[1], numImages),
order='F');
images = double(images)/255
return images
def __init__(self, audioName):
self.wf = wave.open(audioName, "r")
buffer = self.wf.readframes(self.wf.getnframes())
self.data = plt.frombuffer(buffer, dtype="int16")
self.Y = None
self.freqs = None
self.bins = None
print ("MyWav_init")
def __init__(self, audioName):
self.wf = wave.open(audioName, "r")
buffer = self.wf.readframes(self.wf.getnframes())
self.data = plt.frombuffer(buffer, dtype="int16")
self.Y = None
self.freqs = None
self.bins = None
print("MyWav_init")
def get_raw_specgram_for_frame(self, frame, pad_seconds=DEFAULT_PADDING):
if pad_seconds:
pad_frames = int(self.audio_params.framerate * pad_seconds)
(nchannels, sampwidth, framerate, nframes, comptype, compname) = self.audio_params
raw_data = self.extract_specgram_data_for_frame(frame=frame, pad_frames=pad_frames)
data = pylab.frombuffer(raw_data, 'Int' + str(8 * self.audio_params.sampwidth))
min_frame = max(0, frame - pad_frames)
max_frame = min(frame + pad_frames, nframes)
audio_pos = frame * 1.0 / framerate
audio_min = max(0, audio_pos - pad_seconds)
audio_max = min(audio_pos + pad_seconds, nframes * 1.0 / framerate)
pylab.figure(num=None, figsize=(8, 6))
pylab.subplot(111)
pylab.title('Spectrogram of {0}: {1:04.5f} - {2:04.5f} s'.format(self.filename, audio_min, audio_max))
pylab.suptitle('Frames {0:09d} - {1:09d}'.format(min_frame, max_frame))
# results = pyfigaxes.specgram(data, Fs=self.audio_params.framerate)
return pylab.specgram(data, Fs=framerate, NFFT=1024, noverlap=512)
def loadMNISTLabels(filename):
f = open(filename, 'rb')
# Verify Magic Number
s = f.read(4)
magic = int(s.encode('hex'), 16)
assert(magic == 2049)
# Read Number Labels
s = f.read(4)
numLabels = int(s.encode('hex'), 16)
# Get Data
s = f.read()
f.close()
labels = frombuffer(s, uint8)
assert(len(labels) == numLabels)
return labels
