Pyaudio 将 mp3 文件与麦克风混合答案

【问题标题】：Pyaudio mix mp3 files with microphonePyaudio 将 mp3 文件与麦克风混合
【发布时间】：2021-02-24 14:46:40
【问题描述】：

我正在尝试使用 pydub 和 pyaudio 将 2 个 mp3 文件与麦克风输入混合。这是我所做的：

from pydub import AudioSegment
from pydub.playback import play

sound1 = AudioSegment.from_file(r"ΑΓΙΑ ΣΚΕΠΗ.mp3")
sound1_channels = sound1.split_to_mono()
sound1 = sound1_channels[0].overlay(sound1_channels[1])
sound1 = sound1 - 48 # make sound1 quiter 48dB

sound2 = AudioSegment.from_file(r"ΑΓΙΑ ΚΥΡΙΑΚΗ.mp3")
sound2_channels = sound2.split_to_mono()
sound2 = sound2_channels[0].overlay(sound2_channels[1])
sound2 = sound2 - 48 # make sound2 quiter 48dB

import pyaudio

FORMAT = pyaudio.paInt16
CHANNELS = 1
RATE = 44100
CHUNK = 1024

p = pyaudio.PyAudio()
player = p.open(format=FORMAT, channels=CHANNELS, rate=RATE, output=True,frames_per_buffer=CHUNK)
mic_stream = p.open(format=FORMAT, channels=CHANNELS,rate=RATE, input=True,frames_per_buffer=CHUNK)

chunk_time_in_seconds = int(RATE/CHUNK)
chunk_number = 0

while(True):
    mic_data = mic_stream.read(CHUNK)
    mic_sound = AudioSegment(mic_data, sample_width=2, channels=1, frame_rate=RATE)
    
    sound1_part = sound1[chunk_number*chunk_time_in_seconds:(chunk_number+1)*chunk_time_in_seconds]
    sound2_part = sound2[chunk_number*chunk_time_in_seconds:(chunk_number+1)*chunk_time_in_seconds]
    
    
    
    #player.write(mic_sound.raw_data) works well
    mix_sound = sound1_part.overlay(sound2_part).overlay(mic_sound)
    player.write(mix_sound.raw_data) # low microphone quality
    
    chunk_number = chunk_number+1

上面的代码似乎可以工作，但它输出的麦克风输入质量很低。

你能帮我一下吗？

【问题讨论】：

三个mp3混用有用吗？这可能有助于确定问题是特定于麦克风输入还是可能是混音方法...

标签： python mp3 microphone pyaudio pydub

【解决方案1】：

我不知道这种方法是否适合实时音频处理，但下面的代码效果很好！

from pydub import AudioSegment
from pydub.playback import play

sound1 = AudioSegment.from_file(r"ΑΓΙΑ ΣΚΕΠΗ.mp3")
sound1_channels = sound1.split_to_mono()
sound1 = sound1_channels[0].overlay(sound1_channels[1])
sound1 = sound1 - 30 # make sound1 quiter 30dB

sound2 = AudioSegment.from_file(r"ΑΓΙΑ ΚΥΡΙΑΚΗ.mp3")
sound2_channels = sound2.split_to_mono()
sound2 = sound2_channels[0].overlay(sound2_channels[1])
sound2 = sound2 - 30 # make sound2 quiter 30dB

import pyaudio

FORMAT = pyaudio.paInt16
CHANNELS = 1
RATE = 44100
CHUNK = 1024

p = pyaudio.PyAudio()
player = p.open(format=FORMAT, channels=CHANNELS, rate=RATE, output=True,frames_per_buffer=CHUNK)
mic_stream = p.open(format=FORMAT, channels=CHANNELS,rate=RATE, input=True,frames_per_buffer=CHUNK)

#chunk_time_in_seconds = int(RATE/CHUNK)
chunk_number = 0

while(True):
    mic_data = mic_stream.read(CHUNK)
    mic_sound = AudioSegment(mic_data, sample_width=2, channels=1, frame_rate=RATE)
    mic_sound_duration = len(mic_sound)
    
    sound1_part = sound1[chunk_number*mic_sound_duration:(chunk_number+1)*mic_sound_duration]
    sound2_part = sound2[chunk_number*mic_sound_duration:(chunk_number+1)*mic_sound_duration]
    
    
    
    #player.write(mic_sound.raw_data) works well
    mix_sound = sound1_part.overlay(sound2_part).overlay(mic_sound)
    player.write(mix_sound.raw_data) # low microphone quality
    
    chunk_number = chunk_number+1

【讨论】：