使用Python检测和录制音频

问题:使用Python检测和录制音频

我需要将音频片段捕获为WAV文件,然后可以将其传递给另一位python进行处理。问题是我需要确定何时有音频存在,然后记录下来,在静音时停止,然后将该文件传递给处理模块。

我认为wave模块应该可以检测到何时存在纯静音,并在检测到非静音时立即将其丢弃,然后在线路变为静音时再次停止录制。

只是不能完全理解我的想法,没有人可以让我开始做一个基本的例子。

I need to capture audio clips as WAV files that I can then pass to another bit of python for processing. The problem is that I need to determine when there is audio present and then record it, stop when it goes silent and then pass that file to the processing module.

I’m thinking it should be possible with the wave module to detect when there is pure silence and discard it then as soon as something other than silence is detected start recording, then when the line goes silent again stop the recording.

Just can’t quite get my head around it, can anyone get me started with a basic example.


回答 0

作为Nick Fortescue回答的后续措施,下面是一个更完整的示例,说明如何从麦克风进行录音并处理结果数据:

from sys import byteorder
from array import array
from struct import pack

import pyaudio
import wave

THRESHOLD = 500
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)"
    silence = [0] * int(seconds * RATE)
    r = array('h', silence)
    r.extend(snd_data)
    r.extend(silence)
    return r

def record():
    """
    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 > 30:
            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

def record_to_file(path):
    "Records from the microphone and outputs the resulting data to 'path'"
    sample_width, data = record()
    data = pack('<' + ('h'*len(data)), *data)

    wf = wave.open(path, 'wb')
    wf.setnchannels(1)
    wf.setsampwidth(sample_width)
    wf.setframerate(RATE)
    wf.writeframes(data)
    wf.close()

if __name__ == '__main__':
    print("please speak a word into the microphone")
    record_to_file('demo.wav')
    print("done - result written to demo.wav")

As a follow up to Nick Fortescue’s answer, here’s a more complete example of how to record from the microphone and process the resulting data:

from sys import byteorder
from array import array
from struct import pack

import pyaudio
import wave

THRESHOLD = 500
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)"
    silence = [0] * int(seconds * RATE)
    r = array('h', silence)
    r.extend(snd_data)
    r.extend(silence)
    return r

def record():
    """
    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 > 30:
            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

def record_to_file(path):
    "Records from the microphone and outputs the resulting data to 'path'"
    sample_width, data = record()
    data = pack('<' + ('h'*len(data)), *data)

    wf = wave.open(path, 'wb')
    wf.setnchannels(1)
    wf.setsampwidth(sample_width)
    wf.setframerate(RATE)
    wf.writeframes(data)
    wf.close()

if __name__ == '__main__':
    print("please speak a word into the microphone")
    record_to_file('demo.wav')
    print("done - result written to demo.wav")

回答 1

我相信WAVE模块不支持记录,仅处理现有文件。您可能需要查看PyAudio进行实际录制。WAV是关于世界上最简单的文件格式。在paInt16中,您仅获得一个表示电平的有符号整数,而接近0则更安静。我不记得WAV文件是高字节开头还是低字节开头,但是类似这样的东西应该起作用(对不起,我并不是真正的python程序员:

from array import array

# you'll probably want to experiment on threshold
# depends how noisy the signal
threshold = 10 
max_value = 0

as_ints = array('h', data)
max_value = max(as_ints)
if max_value > threshold:
    # not silence

保留用于记录的PyAudio代码以供参考:

import pyaudio
import sys

chunk = 1024
FORMAT = pyaudio.paInt16
CHANNELS = 1
RATE = 44100
RECORD_SECONDS = 5

p = pyaudio.PyAudio()

stream = p.open(format=FORMAT,
                channels=CHANNELS, 
                rate=RATE, 
                input=True,
                output=True,
                frames_per_buffer=chunk)

print "* recording"
for i in range(0, 44100 / chunk * RECORD_SECONDS):
    data = stream.read(chunk)
    # check for silence here by comparing the level with 0 (or some threshold) for 
    # the contents of data.
    # then write data or not to a file

print "* done"

stream.stop_stream()
stream.close()
p.terminate()

I believe the WAVE module does not support recording, just processing existing files. You might want to look at PyAudio for actually recording. WAV is about the world’s simplest file format. In paInt16 you just get a signed integer representing a level, and closer to 0 is quieter. I can’t remember if WAV files are high byte first or low byte, but something like this ought to work (sorry, I’m not really a python programmer:

from array import array

# you'll probably want to experiment on threshold
# depends how noisy the signal
threshold = 10 
max_value = 0

as_ints = array('h', data)
max_value = max(as_ints)
if max_value > threshold:
    # not silence

PyAudio code for recording kept for reference:

import pyaudio
import sys

chunk = 1024
FORMAT = pyaudio.paInt16
CHANNELS = 1
RATE = 44100
RECORD_SECONDS = 5

p = pyaudio.PyAudio()

stream = p.open(format=FORMAT,
                channels=CHANNELS, 
                rate=RATE, 
                input=True,
                output=True,
                frames_per_buffer=chunk)

print "* recording"
for i in range(0, 44100 / chunk * RECORD_SECONDS):
    data = stream.read(chunk)
    # check for silence here by comparing the level with 0 (or some threshold) for 
    # the contents of data.
    # then write data or not to a file

print "* done"

stream.stop_stream()
stream.close()
p.terminate()

回答 2

感谢cryo的改进版本,我基于下面的测试代码:

#Instead of adding silence at start and end of recording (values=0) I add the original audio . This makes audio sound more natural as volume is >0. See trim()
#I also fixed issue with the previous code - accumulated silence counter needs to be cleared once recording is resumed.

from array import array
from struct import pack
from sys import byteorder
import copy
import pyaudio
import wave

THRESHOLD = 500  # audio levels not normalised.
CHUNK_SIZE = 1024
SILENT_CHUNKS = 3 * 44100 / 1024  # about 3sec
FORMAT = pyaudio.paInt16
FRAME_MAX_VALUE = 2 ** 15 - 1
NORMALIZE_MINUS_ONE_dB = 10 ** (-1.0 / 20)
RATE = 44100
CHANNELS = 1
TRIM_APPEND = RATE / 4

def is_silent(data_chunk):
    """Returns 'True' if below the 'silent' threshold"""
    return max(data_chunk) < THRESHOLD

def normalize(data_all):
    """Amplify the volume out to max -1dB"""
    # MAXIMUM = 16384
    normalize_factor = (float(NORMALIZE_MINUS_ONE_dB * FRAME_MAX_VALUE)
                        / max(abs(i) for i in data_all))

    r = array('h')
    for i in data_all:
        r.append(int(i * normalize_factor))
    return r

def trim(data_all):
    _from = 0
    _to = len(data_all) - 1
    for i, b in enumerate(data_all):
        if abs(b) > THRESHOLD:
            _from = max(0, i - TRIM_APPEND)
            break

    for i, b in enumerate(reversed(data_all)):
        if abs(b) > THRESHOLD:
            _to = min(len(data_all) - 1, len(data_all) - 1 - i + TRIM_APPEND)
            break

    return copy.deepcopy(data_all[_from:(_to + 1)])

def record():
    """Record a word or words from the microphone and 
    return the data as an array of signed shorts."""

    p = pyaudio.PyAudio()
    stream = p.open(format=FORMAT, channels=CHANNELS, rate=RATE, input=True, output=True, frames_per_buffer=CHUNK_SIZE)

    silent_chunks = 0
    audio_started = False
    data_all = array('h')

    while True:
        # little endian, signed short
        data_chunk = array('h', stream.read(CHUNK_SIZE))
        if byteorder == 'big':
            data_chunk.byteswap()
        data_all.extend(data_chunk)

        silent = is_silent(data_chunk)

        if audio_started:
            if silent:
                silent_chunks += 1
                if silent_chunks > SILENT_CHUNKS:
                    break
            else: 
                silent_chunks = 0
        elif not silent:
            audio_started = True              

    sample_width = p.get_sample_size(FORMAT)
    stream.stop_stream()
    stream.close()
    p.terminate()

    data_all = trim(data_all)  # we trim before normalize as threshhold applies to un-normalized wave (as well as is_silent() function)
    data_all = normalize(data_all)
    return sample_width, data_all

def record_to_file(path):
    "Records from the microphone and outputs the resulting data to 'path'"
    sample_width, data = record()
    data = pack('<' + ('h' * len(data)), *data)

    wave_file = wave.open(path, 'wb')
    wave_file.setnchannels(CHANNELS)
    wave_file.setsampwidth(sample_width)
    wave_file.setframerate(RATE)
    wave_file.writeframes(data)
    wave_file.close()

if __name__ == '__main__':
    print("Wait in silence to begin recording; wait in silence to terminate")
    record_to_file('demo.wav')
    print("done - result written to demo.wav")

Thanks to cryo for improved version that I based my tested code below:

#Instead of adding silence at start and end of recording (values=0) I add the original audio . This makes audio sound more natural as volume is >0. See trim()
#I also fixed issue with the previous code - accumulated silence counter needs to be cleared once recording is resumed.

from array import array
from struct import pack
from sys import byteorder
import copy
import pyaudio
import wave

THRESHOLD = 500  # audio levels not normalised.
CHUNK_SIZE = 1024
SILENT_CHUNKS = 3 * 44100 / 1024  # about 3sec
FORMAT = pyaudio.paInt16
FRAME_MAX_VALUE = 2 ** 15 - 1
NORMALIZE_MINUS_ONE_dB = 10 ** (-1.0 / 20)
RATE = 44100
CHANNELS = 1
TRIM_APPEND = RATE / 4

def is_silent(data_chunk):
    """Returns 'True' if below the 'silent' threshold"""
    return max(data_chunk) < THRESHOLD

def normalize(data_all):
    """Amplify the volume out to max -1dB"""
    # MAXIMUM = 16384
    normalize_factor = (float(NORMALIZE_MINUS_ONE_dB * FRAME_MAX_VALUE)
                        / max(abs(i) for i in data_all))

    r = array('h')
    for i in data_all:
        r.append(int(i * normalize_factor))
    return r

def trim(data_all):
    _from = 0
    _to = len(data_all) - 1
    for i, b in enumerate(data_all):
        if abs(b) > THRESHOLD:
            _from = max(0, i - TRIM_APPEND)
            break

    for i, b in enumerate(reversed(data_all)):
        if abs(b) > THRESHOLD:
            _to = min(len(data_all) - 1, len(data_all) - 1 - i + TRIM_APPEND)
            break

    return copy.deepcopy(data_all[_from:(_to + 1)])

def record():
    """Record a word or words from the microphone and 
    return the data as an array of signed shorts."""

    p = pyaudio.PyAudio()
    stream = p.open(format=FORMAT, channels=CHANNELS, rate=RATE, input=True, output=True, frames_per_buffer=CHUNK_SIZE)

    silent_chunks = 0
    audio_started = False
    data_all = array('h')

    while True:
        # little endian, signed short
        data_chunk = array('h', stream.read(CHUNK_SIZE))
        if byteorder == 'big':
            data_chunk.byteswap()
        data_all.extend(data_chunk)

        silent = is_silent(data_chunk)

        if audio_started:
            if silent:
                silent_chunks += 1
                if silent_chunks > SILENT_CHUNKS:
                    break
            else: 
                silent_chunks = 0
        elif not silent:
            audio_started = True              

    sample_width = p.get_sample_size(FORMAT)
    stream.stop_stream()
    stream.close()
    p.terminate()

    data_all = trim(data_all)  # we trim before normalize as threshhold applies to un-normalized wave (as well as is_silent() function)
    data_all = normalize(data_all)
    return sample_width, data_all

def record_to_file(path):
    "Records from the microphone and outputs the resulting data to 'path'"
    sample_width, data = record()
    data = pack('<' + ('h' * len(data)), *data)

    wave_file = wave.open(path, 'wb')
    wave_file.setnchannels(CHANNELS)
    wave_file.setsampwidth(sample_width)
    wave_file.setframerate(RATE)
    wave_file.writeframes(data)
    wave_file.close()

if __name__ == '__main__':
    print("Wait in silence to begin recording; wait in silence to terminate")
    record_to_file('demo.wav')
    print("done - result written to demo.wav")

回答 3

import pyaudio
import wave
from array import array

FORMAT=pyaudio.paInt16
CHANNELS=2
RATE=44100
CHUNK=1024
RECORD_SECONDS=15
FILE_NAME="RECORDING.wav"

audio=pyaudio.PyAudio() #instantiate the pyaudio

#recording prerequisites
stream=audio.open(format=FORMAT,channels=CHANNELS, 
                  rate=RATE,
                  input=True,
                  frames_per_buffer=CHUNK)

#starting recording
frames=[]

for i in range(0,int(RATE/CHUNK*RECORD_SECONDS)):
    data=stream.read(CHUNK)
    data_chunk=array('h',data)
    vol=max(data_chunk)
    if(vol>=500):
        print("something said")
        frames.append(data)
    else:
        print("nothing")
    print("\n")


#end of recording
stream.stop_stream()
stream.close()
audio.terminate()
#writing to file
wavfile=wave.open(FILE_NAME,'wb')
wavfile.setnchannels(CHANNELS)
wavfile.setsampwidth(audio.get_sample_size(FORMAT))
wavfile.setframerate(RATE)
wavfile.writeframes(b''.join(frames))#append frames recorded to file
wavfile.close()

我认为这会有所帮助,这是一个简单的脚本,它将检查是否存在静音,如果检测到静音,则不会记录,否则会记录。

import pyaudio
import wave
from array import array

FORMAT=pyaudio.paInt16
CHANNELS=2
RATE=44100
CHUNK=1024
RECORD_SECONDS=15
FILE_NAME="RECORDING.wav"

audio=pyaudio.PyAudio() #instantiate the pyaudio

#recording prerequisites
stream=audio.open(format=FORMAT,channels=CHANNELS, 
                  rate=RATE,
                  input=True,
                  frames_per_buffer=CHUNK)

#starting recording
frames=[]

for i in range(0,int(RATE/CHUNK*RECORD_SECONDS)):
    data=stream.read(CHUNK)
    data_chunk=array('h',data)
    vol=max(data_chunk)
    if(vol>=500):
        print("something said")
        frames.append(data)
    else:
        print("nothing")
    print("\n")


#end of recording
stream.stop_stream()
stream.close()
audio.terminate()
#writing to file
wavfile=wave.open(FILE_NAME,'wb')
wavfile.setnchannels(CHANNELS)
wavfile.setsampwidth(audio.get_sample_size(FORMAT))
wavfile.setframerate(RATE)
wavfile.writeframes(b''.join(frames))#append frames recorded to file
wavfile.close()

I think this will help.It is a simple script which will check if there is a silence or not.If silence is detected it will not record otherwise it will record.


回答 4

pyaudio网站上有许多简短明了的示例:http ://people.csail.mit.edu/hubert/pyaudio/

2019年12月14日更新-上述链接网站自2017年以来的主要示例:


"""PyAudio Example: Play a WAVE file."""

import pyaudio
import wave
import sys

CHUNK = 1024

if len(sys.argv) < 2:
    print("Plays a wave file.\n\nUsage: %s filename.wav" % sys.argv[0])
    sys.exit(-1)

wf = wave.open(sys.argv[1], 'rb')

p = pyaudio.PyAudio()

stream = p.open(format=p.get_format_from_width(wf.getsampwidth()),
                channels=wf.getnchannels(),
                rate=wf.getframerate(),
                output=True)

data = wf.readframes(CHUNK)

while data != '':
    stream.write(data)
    data = wf.readframes(CHUNK)

stream.stop_stream()
stream.close()

p.terminate()

The pyaudio website has many examples that are pretty short and clear: http://people.csail.mit.edu/hubert/pyaudio/

Update 14th of December 2019 – Main example from the above linked website from 2017:


"""PyAudio Example: Play a WAVE file."""

import pyaudio
import wave
import sys

CHUNK = 1024

if len(sys.argv) < 2:
    print("Plays a wave file.\n\nUsage: %s filename.wav" % sys.argv[0])
    sys.exit(-1)

wf = wave.open(sys.argv[1], 'rb')

p = pyaudio.PyAudio()

stream = p.open(format=p.get_format_from_width(wf.getsampwidth()),
                channels=wf.getnchannels(),
                rate=wf.getframerate(),
                output=True)

data = wf.readframes(CHUNK)

while data != '':
    stream.write(data)
    data = wf.readframes(CHUNK)

stream.stop_stream()
stream.close()

p.terminate()

回答 5

您可能还想看看csounds。它具有多个API,包括Python。它可能能够与AD界面进行交互并收集声音样本。

You might want to look at csounds, also. It has several API’s, including Python. It might be able to interact with an A-D interface and gather sound samples.