我是python中的新手，我尝试构建一个能够解码和编码用于拨打电话的双音多频（DTMF）信号的程序。 目前编码部分工作正常，但由于某种原因编码不起作用，我得到了以下异常
Traceback (most recent call last):
File "C:\Users\matant\workspace\dialer2\dialer.py", line 239, in <module>
x = d.decoder()
File "C:\Users\matant\workspace\dialer2\dialer.py", line 218, in decoder
data = self.read_wav()
File "C:\Users\matant\workspace\dialer2\dialer.py", line 201, in read_wav
n = fin.getnframes()
AttributeError: 'file' object has no attribute 'getnframes'
你可以看到我在文件中写入帧，所以我不明白为什么会发生这种情况： 这是我的代码：
'''
Created on Jan 10, 2016
@author: matant
'''
import json
from math import pi, sin
import wave
import logging
import struct
import os
ROW_FREQ = (697, 770, 852, 941)
COL_FREQ = (1209, 1336, 1477, 1633)
SAMPLE_RATE = 44100
SAMPLE_WIDTH = 2
NUMBER_OF_CHANNELS = 1
COMPRESSION_TYPE = "NONE"
COMPRESSION_NAME = "Uncompressed"
PI2 = 6.283185306
scale = 32767 #16-bit unsigned short
keys= '1','2','3','A',\
'4','5','6','B',\
'7','8','9','C',\
'*','0','#','D'
FREQUENCY_MAP = dict()
FREQUENCY_MAP['1'] = (697, 1209)
FREQUENCY_MAP['2'] = (697, 1336)
FREQUENCY_MAP['3'] = (697, 1477)
FREQUENCY_MAP['A'] = (697, 1633)
FREQUENCY_MAP['4'] = (770, 1209)
FREQUENCY_MAP['5'] = (770, 1336)
FREQUENCY_MAP['6'] = (770, 1477)
FREQUENCY_MAP['B'] = (770, 1633)
FREQUENCY_MAP['7'] = (852, 1209)
FREQUENCY_MAP['8'] = (852, 1336)
FREQUENCY_MAP['9'] = (852, 1477)
FREQUENCY_MAP['C'] = (852, 1633)
FREQUENCY_MAP['*'] = (941, 1209)
FREQUENCY_MAP['0'] = (941, 1336)
FREQUENCY_MAP['#'] = (941, 1477)
FREQUENCY_MAP['D'] = (941, 1633)
FREQUENCY_MAP['S'] = (0, 0)
logging.basicConfig(level=logging.DEBUG,
format='%(asctime)s <%(levelname)s> %(module)s.%(funcName)s() %(message)s',
datefmt='%Y-%m-%d %H:%M:%S')
log = logging.getLogger(__name__)
class DTMF:
VALID_SEQUENCE_TYPES = [list, tuple, set]
def __init__(self, input_string=None, input_list=None):
"""
Initializes a DTMF instance with an option DTMF sequence. This can be a list of lists or a json string.
If both are supplied, it tries to parse the json_string. If it does, it uses that. If there are errors, it
validates the list and tries to use that. Basically input_string takes precedence.
General workflow would be setting dtmf_sequence and calling generate_raw_data. This data can then be saved to a
.wav file or compressed and saved as other, smaller, file formats.
:param input_list: list of lists or tuples of the form [['A', 100], ['S', 50], ['2', 100], ['S', 50]]
:param input_string: json_string of the form '[["A", 100], ["S", 50], ["2", 100], ["S", 50]]'
"""
log.debug("Creating instance of DTMF")
log.debug("input_string = {}".format(input_string))
log.debug("input_list = {}".format(input_list))
self._dtmf_sequence = None
self._raw_data = None
if input_string is not None:
converted_json_sequence = self.parse_json_string(input_string)
self._dtmf_sequence = converted_json_sequence
elif input_list is not None:
self._dtmf_sequence = input_list
@property
def dtmf_sequence(self):
return self._dtmf_sequence
@dtmf_sequence.setter
def dtmf_sequence(self, input_sequence):
if type(input_sequence) == str:
input_sequence = self.parse_json_string(input_sequence)
if type(input_sequence) == list:
if self._dtmf_sequence_is_valid(input_sequence):
self._dtmf_sequence = input_sequence
log.debug("Set _dtmf_sequence to {}".format(self._dtmf_sequence))
def parse_json_string(self, input_string):
return json.loads(input_string)
def generate_raw_data(self):
"""
Generates raw data that can be saved into a .wav file. This can take some time to generate.
:raise AttributeError: If no dtmf sequence has been set
"""
_data = list()
if self._dtmf_sequence is None:
raise AttributeError("No dtmf sequence set")
for tone_tuple in self._dtmf_sequence:
key = tone_tuple[0]
tone_duration = tone_tuple[1]
f1 = FREQUENCY_MAP[key][0]
f2 = FREQUENCY_MAP[key][1]
_data += (self.generate_tone(f1, f2, tone_duration))
self._raw_data = _data
def save_wave_file(self, file_path):
if self._raw_data is None or len(self._raw_data) < 1:
self.generate_raw_data()
f = wave.open(file_path, 'w')
f.setnchannels(NUMBER_OF_CHANNELS)
f.setsampwidth(SAMPLE_WIDTH)
f.setframerate(SAMPLE_RATE)
f.setnframes(len(self._raw_data))
f.setcomptype(COMPRESSION_TYPE, COMPRESSION_NAME)
log.info("Saving wav file {} THIS MAY TAKE A WHILE".format(file_path))
for i in self._raw_data:
f.writeframes(struct.pack('i', i))
log.info("Saved file to {0}".format(file_path))
f.close()
@staticmethod
def dtmf_sequence_is_valid(input_list):
"""
Validates an input sequence for proper structure and contents.
:param input_list:
:return:
"""
if type(input_list) is not list:
log.warning('input_list must be a list instance')
return False
if [(type(item) in DTMF.VALID_SEQUENCE_TYPES) for item in input_list].count(False) != 0:
log.warning('input_list contains invalid sequence type')
return False
for item in input_list:
if type(item[0]) != str or type(item[1]) != int:
log.debug("Type list[0]: {}".format(type(item[0])))
log.debug("Type list[1]: {}".format(type(item[1])))
log.warning('input_list must contain a list of sequences of [str, int]')
return False
return True
@staticmethod
def generate_tone(f1, f2, _duration_in_ms):
"""
Generates a single value representing a sample of two combined frequencies.
:param f1:
:param f2:
:param _duration_in_ms:
:return:
"""
assert f1 in ROW_FREQ or f1 == 0
assert f2 in COL_FREQ or f2 == 0
number_of_samples = int(SAMPLE_RATE * _duration_in_ms / 1000)
scale = 32767 # signed int / 2
result = list()
for i in range(number_of_samples):
p = i * 1.0 / SAMPLE_RATE
result.append(int((sin(p * f1 * pi * 2) + sin(p * f2 * pi * 2)) / 2 * scale))
log.info(
"Generated {0}ms tone of {1} samples with F1: {2} F2: {3}".format(_duration_in_ms, number_of_samples, f1,
f2))
return result
def create_dtmf_wave_file(self, input_sequence, file_path, dump_to_csv=False):
"""
A convenience method. Validates and assigns a dtmf_sequence, then generates data and saves to a .wav
:param input_sequence: list of lists or tuples of the form [['A', 100], ['S', 50], ['2', 100], ['S', 50]] or json_string of the form '[["A", 100], ["S", 50], ["2", 100], ["S", 50]]'
:param file_path: the full path of the wav file that will be saved
"""
self._dtmf_sequence = input_sequence
self.generate_raw_data()
try:
os.remove('dtmf_dump.csv')
except:
pass # file doesn't exist
if dump_to_csv:
with open('dtmf_dump.csv', 'w') as f:
for d in self._raw_data:
f.write(str(d))
f.write(",")
self.save_wave_file(file_path)
def read_wav(self):
fin = open('testNum.wav','r')
n = fin.getnframes()
d = fin.readframes(n)
fin.close()
data = []
for i in range(n):
#LS8bit = inv_endian(ord(d[2*i]))
#MS8bit = inv_endian(ord(d[2*i+1]))
LS8bit, MS8bit = ord(d[2*i]),ord(d[2*i+1])
data.append((MS8bit<<8)+LS8bit)
return data
# Decoder takes a DTMF signal file (.wav), sampled at 44,000
# 16-bit samples per second, and decode the corresponding symbol X.
def decoder(self):
data = self.read_wav()
temp = []
for f1 in ROW_FREQ:
for f2 in COL_FREQ:
diff = 0
for i in range(SAMPLE_RATE): #assume phase has not shifted dramatically
p = i*1.0/SAMPLE_RATE
S=int(scale+scale*(sin(p*f1*PI2)+sin(p*f2*PI2))/2)
diff += abs(S-data[i])
temp.append((diff,f1,f2))
f1,f2 = min(temp)[1:] #retrieve the frequency of minimum signal distortion
i, j = ROW_FREQ.index(f1), COL_FREQ.index(f2)
X = keys[4*i+j]
print 'Decoded key is: ', X
return X
if __name__ == '__main__':
d = 100
sample_input = [('0', d), ('5', d), ('0', d), ('8', d), ('6', d), ('9', d), ('0',d), ('1',d) , ('8',d),('6',d)]
d = DTMF()
d.create_dtmf_wave_file(sample_input, file_path='testNum.wav', dump_to_csv=True)
x = d.decoder()
答案 0 :(得分：1)
fin = open('testNum.wav','r')
看起来您使用的是内置open功能，而不是wave模块中的功能。尝试：
fin = wave.open('testNum.wav','r')
答案 1 :(得分：0)
您必须使用可以从wave read object返回的wave.open进行操作，该方法将返回包含您的代码正在尝试访问的属性的文件。
此外，您添加了新的from wave import open语句，因此您将以这种方式覆盖默认的open方法，但最好将wave open方法思维点存在视为wave.open