bitstream05.readBit(1)
bitstream05.closeFile()
# ---------------ADD NUMBER FEATURE (TO GET MORE EFFICIENCY)--------------
if test06:
bitstream06 = BitStream("./out/test06.txt", "wbs")
t_array01 = [1, 1, 1, 1, 1, 1, 1, 1]
t_array02 = [0, 0, 0, 0, 0, 0, 0, 0]
t_array03 = [1, 1, 1, 1]
t_array04 = [1, 1, 1, 1]
t_array05 = [1, 1, 1, 1, 1, 1, 1, 1, 1]
t_array06 = [0, 0, 0, 0, 0, 0, 0]
t_array07 = [1, 0, 1]
bitstream06.addNumber(t_array01)
bitstream06.addNumber(t_array02)
bitstream06.addNumber(t_array03)
bitstream06.addNumber(t_array04)
bitstream06.addNumber(t_array05)
bitstream06.addNumber(t_array06)
bitstream06.addNumber(t_array07)
print(bitstream06.write_array_final)
bitstream06.closeFile()
bitstream06 = BitStream("./out/test06.txt", "rb")
assert bitstream06.readBit(8) == [1, 1, 1, 1, 1, 1, 1, 1]
assert bitstream06.readBit(8) == [0, 0, 0, 0, 0, 0, 0, 0]
class IntraFrameEncoder():
"""
This class implements a lossless intra-frame encoder, using 7PEG linear predictors.
"""
def __init__(self, predictor):
"""
Default constructor.
@param matrix: initial matrix
@param predictor: linear predictor
"""
self.original_matrix = None
self.predictor = predictor
self.encoded_matrix = None
# Golomb encoder
self.golomb = Golomb(4)
self.bitstream = BitStream("../out/encoded_park_joy_444_720p50.bin",
"wbs")
self.written_bits = 0 # TODO: tira isto
self.codes = []
def write_code(self, code):
"""
This method writes a list of bits on file, using the Bitstream class.
@param code: list with bits to encode.
"""
# for bit in code:
# self.written_bits += 1
# self.bitstream.writeBit(bit,1)
self.written_bits += len(code)
# self.bitstream.writeArray(code)
self.bitstream.addNumber(code)
def setMatrix(self, new_matrix):
"""
This method sets current matrix of the encoder to 'new_matrix'.
@param new_matrix: new matrix of type Y,U or V.
"""
self.original_matrix = new_matrix
if self.encoded_matrix is None:
self.encoded_matrix = np.empty(
self.original_matrix.shape) # sighly faster
self.codes = []
def encode(self):
"""
This method encodes the original matrix in a new one, based on the current predictor.
It also uses golomb codification for the entropy encoding.
"""
if self.original_matrix is None:
logger.error("No matrix to encode was given!")
return False
# TODO: ver o que é aquele K do stor
# write header with bitstream
#self.bitstream.writeString("{}\t{}".format(self.original_matrix.shape[0],self.original_matrix.shape[1]))
# matrix size/shape is the same no mather which one
self.encoded_matrix[0, 0] = int(self.original_matrix[0, 0] -
self.predictor.predict(0, 0, 0))
for col in range(1, self.original_matrix.shape[1]):
self.encoded_matrix[0, col] = int(
self.original_matrix[0, col]) - self.predictor.predict(
self.original_matrix[0, col - 1], 0, 0)
for line in range(1, self.original_matrix.shape[0]):
self.encoded_matrix[line, 0] = int(
self.original_matrix[line, 0]) - self.predictor.predict(
0, self.original_matrix[line - 1, 0], 0)
for line in range(1, self.original_matrix.shape[0]):
for col in range(1, self.original_matrix.shape[1]):
self.encoded_matrix[line, col] = int(
self.original_matrix[line, col]) - self.predictor.predict(
self.original_matrix[line, col - 1],
self.original_matrix[line - 1, col],
self.original_matrix[line - 1, col - 1])
for line in range(self.encoded_matrix.shape[0]):
for col in range(self.encoded_matrix.shape[1]):
self.write_code(
self.golomb.encoded_values[self.encoded_matrix[line, col]])