我正在使用PyAudio来检测音频源的活动。我正在通常死亡的流上从每个音频事件创建WAV文件。使用memory_profiler我注意到record_to_file（）方法中的pack方法通常使用的内存量是删除数据对象所能恢复的内存量的4倍。此代码取自Detect & Record Audio in Python
from sys import byteorder
import sys
from array import array
import struct
import gc
import pyaudio
import wave
import subprocess
import objgraph
from memory_profiler import profile
from guppy import hpy
THRESHOLD = 5000
CHUNK_SIZE = 1024
FORMAT = pyaudio.paInt16
RATE = 44100
def is_silent(snd_data):
"Returns 'True' if below the 'silent' threshold"
return max(snd_data) < THRESHOLD
def normalize(snd_data):
"Average the volume out"
MAXIMUM = 16384
times = float(MAXIMUM)/max(abs(i) for i in snd_data)
r = array('h')
for i in snd_data:
r.append(int(i*times))
return r
def trim(snd_data):
"Trim the blank spots at the start and end"
def _trim(snd_data):
snd_started = False
r = array('h')
for i in snd_data:
if not snd_started and abs(i)>THRESHOLD:
snd_started = True
r.append(i)
elif snd_started:
r.append(i)
return r
# Trim to the left
snd_data = _trim(snd_data)
# Trim to the right
snd_data.reverse()
snd_data = _trim(snd_data)
snd_data.reverse()
return snd_data
def add_silence(snd_data, seconds):
"Add silence to the start and end of 'snd_data' of length 'seconds' (float)"
r = array('h', [0 for i in xrange(int(seconds*RATE))])
r.extend(snd_data)
r.extend([0 for i in xrange(int(seconds*RATE))])
return r
def record():
"""g
Record a word or words from the microphone and
return the data as an array of signed shorts.
Normalizes the audio, trims silence from the
start and end, and pads with 0.5 seconds of
blank sound to make sure VLC et al can play
it without getting chopped off.
"""
p = pyaudio.PyAudio()
stream = p.open(format=FORMAT, channels=1, rate=RATE,
input=True, output=True,
frames_per_buffer=CHUNK_SIZE)
num_silent = 0
snd_started = False
r = array('h')
while 1:
# little endian, signed short
snd_data = array('h', stream.read(CHUNK_SIZE))
if byteorder == 'big':
snd_data.byteswap()
r.extend(snd_data)
silent = is_silent(snd_data)
if silent and snd_started:
num_silent += 1
elif not silent and not snd_started:
snd_started = True
if snd_started and num_silent > 400:
break
sample_width = p.get_sample_size(FORMAT)
stream.stop_stream()
stream.close()
p.terminate()
#r = normalize(r)
r = trim(r)
r = add_silence(r, 0.5)
return sample_width, r
@profile
def record_to_file(path):
"Records from the microphone and outputs the resulting data to 'path'"
sample_width, data = record()
data = struct.pack('<' + ('h'*len(data)), *data)
print(sys.getsizeof(data))
wf = wave.open(path, 'wb')
wf.setnchannels(1)
wf.setsampwidth(sample_width)
wf.setframerate(RATE)
wf.writeframes(data)
wf.close()
del data
gc.collect()
if __name__ == '__main__':
count = 1
h = hpy()
f = open('heap.txt','w')
objgraph.show_growth(limit=3)
while(1):
filename = 'demo' + str(count) + '.wav'
print("please speak a word into the microphone")
record_to_file(filename)
print("done - result written to {0}".format(filename))
cmd = 'cd "C:\\Users\\user\\Desktop\\Raudio" & ffmpeg\\bin\\ffmpeg -i {0} -acodec libmp3lame {1}.mp3'.format(filename, filename)
"subprocess.call(cmd, shell=True)"
count += 1
objgraph.show_growth()
print h.heap()
以下是内存分析器模块的一次输出迭代：
Line # Mem usage Increment Line Contents
================================================
117 19.6 MiB 0.0 MiB @profile
118 def record_to_file(path):
119 "Records from the microphone and outputs the resulting data to 'path'"
120 22.4 MiB 2.8 MiB sample_width, data = record()
125 31.3 MiB 8.9 MiB data = struct.pack('<' + ('h'*len(data)), *data)
126 31.3 MiB 0.0 MiB print(sys.getsizeof(data))
127 31.3 MiB 0.0 MiB wf = wave.open(path, 'wb')
128 31.3 MiB 0.0 MiB wf.setnchannels(1)
129 31.3 MiB 0.0 MiB wf.setsampwidth(sample_width)
130 31.4 MiB 0.0 MiB wf.setframerate(RATE)
131 31.4 MiB 0.0 MiB wf.writeframes(data)
132 31.4 MiB 0.0 MiB wf.close()
133 30.4 MiB -0.9 MiB del data
134 27.9 MiB -2.5 MiB gc.collect()
使用VS调试进程，我在系统内存中看到了非常大量的“h”字符，这可能就是泄漏发生的原因。任何帮助将不胜感激