如何直接从麦克风播放输入数据

Question

I want to play many input data from a microphone without buffering. I tried, but there is buffering. Here is my code.

import pyaudio
import wave
import urllib.request
import struct
import numpy as np
import sounddevice as sd
import matplotlib.pyplot as plt

# Callback function---------------------------------
def callback(indata, outdata, frames, time, status):
#    if status:
#        print(status)
    outdata[:] = indata
#---------------------------------------------------

# Parameters ----------------------------------------------
Window_Size = 22050 # Point
FORMAT_D = pyaudio.paFloat32; FORMAT_W = pyaudio.paInt32
CHANNELS = 1 # Mono
Sample_Rate = 22050 # Hz
dT = 1/Sample_Rate
RECORD_SECONDS = 20 # s
NOFFRAMES = int(Sample_Rate/Window_Size * RECORD_SECONDS)
WAVE_OUTPUT_FILENAME = "output.wav"
#-----------------------------------------------------------

p = pyaudio.PyAudio()

stream_D = p.open(format=FORMAT_D,
                channels=CHANNELS,
                rate=Sample_Rate,
                input=True,
                frames_per_buffer=Window_Size)

stream_W = p.open(format=FORMAT_W,
            channels=CHANNELS,
            rate=Sample_Rate,
            input=True,
            frames_per_buffer=Window_Size)

print("* recording")

frames = []

# "I think the problem appears from here"------------------------------
for i in range(0, int(Sample_Rate/Window_Size * RECORD_SECONDS)):
    data_D = stream_D.read(Window_Size)
#    data_W = stream_W.read(Window_Size)
    decoded = np.fromstring(data_D, 'Float32')
#    np.savetxt(str(i)+'ttt.txt',transform)
    sd.play(decoded,22050)
#    frames.append(data_W)
#-------------------------------------------------------

print("* done recording")

stream_D.stop_stream()
stream_D.close()
p.terminate()

#plt.plot(transform)
#plt.show()

# Save as a wave file---------------------------
#wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
#wf.setnchannels(CHANNELS)
#wf.setsampwidth(p.get_sample_size(FORMAT_W))
#wf.setframerate(Sample_Rate)
#wf.writeframes(b''.join(frames))
#wf.close()
#-------------------------------------------

This code performs to save input data from a microphone at 1s intervals, transform byte data to nparray data (np.transform()), and play the data with a speaker (sd.play()). This code works, but there is buffering when for loop start again. I want to play the sound from a microphone smoothly. When I asked first, someone recommended to use callback function, so I added it, But, I don't know how to use it. How do I get rid of the buffering? Is there some examples? Should I use Threads or multiprocessing?

Answer 1

The delay is due to buffer size ... you will get a negligible delay using a 1k buffer as per

# Window_Size = 22050 # Point
Window_Size = 1024 # Point