def deserialize_eeg_wifi(self, eeg: np.ndarray, ids: np.ndarray,
context: Dict[str, Any]) -> np.ndarray:
"""From signed 24-bits integer to signed 32-bits integer by channels.
The `Cyton data format <https://docs.openbci.com/docs/02Cyton/CytonDataFormat>`_
says that only can send packages of 33 bits, when a Daisy board is
attached these same packages will be sent at double speed in favor to
keep the desired sample rate for 16 channels.
Parameters
----------
eeg
Numpy array in signed 24-bits integer (`8, LENGTH`)
ids
List of IDs for eeg data.
context
Information from the acquisition side useful for deserializing and
that will be packaged back in the stream.
Returns
-------
eeg_data
EEG data in microvolts, signed 32-bits integer, (`CHANNELS, LENGTH`),
if there is a Daisy board `CHANNELS` is 16, otherwise is 8.
"""
eeg_data = np.array([[
rawutil.unpack('>u', bytes(ch))[0]
for ch in row.reshape(-1, 3).tolist()
] for row in eeg])
eeg_data = eeg_data * self.scale_factor_eeg
if context['daisy']:
# # If offset, the pair index condition must change
if np.array(self.offset[0]).any():
eeg_data = np.concatenate([[self.offset[0]], eeg_data], axis=0)
ids = np.concatenate([[self.offset[1]], ids], axis=0)
# pair = not pair
if ids[0] != ids[1]:
eeg_data = np.delete(eeg_data, 0, axis=0)
ids = np.delete(ids, 0, axis=0)
# if not pair dataset, create an offeset
if eeg_data.shape[0] % 2:
self.offset = eeg_data[-1], ids[-1]
eeg_data = np.delete(eeg_data, -1, axis=0)
ids = np.delete(ids, -1, axis=0)
else:
self.offset = None, None
return eeg_data.reshape(-1, 16)
return eeg_data
# create a socket
sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
# bind the socket to the specified ip address and port
sock.bind((UDP_IP, UDP_PORT))
print ("F1 Telemetry ready")
print ("Listening on " + UDP_IP + ":" + str(UDP_PORT))
print("ses_t,lap_t,lap_d,ses_d,speed,steer,gear,rpm,fuel,fuel_cap,brake_temp,tyre_p_rl,tyre_p_rr,tyre_p_fl,tyre_p_fr,tyre_t_rl,tyre_t_rr,tyre_t_fl,tyre_t_fr")
while True:
data, address = sock.recvfrom(PACKET_SIZE)
ses_t, = rawutil.unpack('<f',data[0:4])
lap_t, = rawutil.unpack('<f',data[4:8])
lap_d, = rawutil.unpack('<f',data[8:12])
ses_d, = rawutil.unpack('<f',data[0:4])
speed, = rawutil.unpack('<f',data[28:32])
steer, = rawutil.unpack('<f',data[120:124])
gear, = rawutil.unpack('<f',data[132:136])
rpm, = rawutil.unpack('<f',data[148:152])
fuel, = rawutil.unpack('<f',data[180:184])
fuel_cap, = rawutil.unpack('<f',data[184:188])
brake_temp, = rawutil.unpack('<f',data[204:208])
tyre_p_rl, = rawutil.unpack('<f',data[220:224])
tyre_p_rr, = rawutil.unpack('<f',data[224:228])
tyre_p_fl, = rawutil.unpack('<f',data[228:232])
tyre_p_fr, = rawutil.unpack('<f',data[232:236])
tyre_t_rl, = rawutil.unpack('<b',data[304:305])
wf = wave.open(r'Ses01F_impro01_F000_S1.wav', 'rb')
fp=open(r'1.wav','rb')
nframes = wf.getnframes()
framerate = wf.getframerate()
str_data = wf.readframes(nframes)
sample_width = wf.getsampwidth()
channel = wf.getnchannels()
bytesPerSample = int(len(str_data) / nframes / channel)
# chunk=fp.read(bytesPerSample*nframes)
a=[]
for i in range(nframes):
j = i * channel
bytess = str_data[j * bytesPerSample:j * bytesPerSample + bytesPerSample]
# value = struct.unpack('<h',bytess)[0]
value = rawutil.unpack('<1u', bytess)[0]
a.append(value)
# print(value)
a2 = np.array(a, dtype='int32')
mean = np.mean(a2)
std = np.std(a2)
a2 = (a2 - mean) / std
a2 = a2.astype(np.float16)
print(a2.dtype)
wf.close()
time = np.arange(0, nframes)*(1.0/framerate)
plt.subplot(311)
plt.plot(time, a2, c='r')
plt.show()
#
# print("channel:",channel)
def deserialize_eeg_serial(self, eeg: np.ndarray, ids: np.ndarray,
context: Dict[str, Any]) -> np.ndarray:
"""From signed 24-bits integer to signed 32-bits integer by channels.
The `Cyton data format <https://docs.openbci.com/docs/02Cyton/CytonDataFormat>`_
says that only can send packages of 33 bits, over serial (RFduino) this
limit is absolute, when a Daisy board is attached these same amount of
packages will be sent, in this case, the data must be distributed and
interpolated in order to complete the sample rate.
Parameters
----------
eeg
Numpy array in signed 24-bits integer (`8, LENGTH`)
ids
List of IDs for eeg data.
context
Information from the acquisition side useful for deserializing and
that will be packaged back in the stream.
Returns
-------
eeg_data
EEG data in microvolts, signed 32-bits integer, (`CHANNELS, LENGTH`),
if there is a Daisy board `CHANNELS` is 16, otherwise is 8.
"""
eeg_data = np.array([[
rawutil.unpack('>u', bytes(ch))[0]
for ch in row.reshape(-1, 3).tolist()
] for row in eeg])
eeg_data = eeg_data * self.scale_factor_eeg
if context['daisy']:
even = not ids[0] % 2
# If offset, the even index condition must change
if np.array(self.offset[0]).any():
eeg_data = np.concatenate([[self.offset[0]], eeg_data], axis=0)
ids = np.concatenate([[self.offset[1]], ids], axis=0)
even = not even
# if not even dataset, create an offset
if eeg_data.shape[0] % 2:
self.offset = eeg_data[-1], ids[-1]
eeg_data = np.delete(eeg_data, -1, axis=0)
ids = np.delete(ids, -1, axis=0)
# Data can start with a even or odd id
if even:
board = eeg_data[::2]
daisy = eeg_data[1::2]
else:
daisy = eeg_data[::2]
board = eeg_data[1::2]
board = np.array([
np.interp(np.arange(0, p.shape[0], 0.5), np.arange(p.shape[0]),
p) for p in board.T
]).T
daisy = np.array([
np.interp(np.arange(0, p.shape[0], 0.5), np.arange(p.shape[0]),
p) for p in daisy.T
]).T
eeg = np.concatenate([np.stack(board), np.stack(daisy)], axis=1)
else:
eeg = eeg_data
return eeg