使用 alsaaudio 计算现场麦克风音频的分贝值

Question

我有一个在 Ubuntu 笔记本电脑上运行的 Python 脚本，用于从笔记本电脑上的麦克风获取实时环境噪声级别。我认为这个脚本效果不佳，而且我不明白为什么。我怀疑这可能只是一个较长的周期大小，但我不确定是否就是这样。有人可以帮我理解这个脚本有什么问题吗？谢谢！

import alsaaudio
import audioop
import time
import math

# Set the audio parameters
device = 'default'
channels = 1  # mono
sample_rate = 44100  # 44.1 kHz
format = alsaaudio.PCM_FORMAT_S16_LE

# Open the audio input stream
inp = alsaaudio.PCM(alsaaudio.PCM_CAPTURE, alsaaudio.PCM_NORMAL, device)
inp.setchannels(channels)
inp.setrate(sample_rate)
inp.setformat(format)
inp.setperiodsize(256)

# Main loop to continuously print audio levels
try:
    while True:
        # Read audio data from the microphone
        _, data = inp.read()

        # Calculate the peak audio level in decibels
        rms = audioop.rms(data, 2)  # 2 for sample width 16-bit

        # Print the decibel level without dynamic range adjustment
        decibels = 20 * math.log10(rms)
        print(f"Decibels: {decibels:.2f} dB")

        time.sleep(0.01)  # Adjust the sleep duration as needed

except KeyboardInterrupt:
    print("Stopping the script.")

我希望这个脚本很敏感，如果我发出声音，分贝值会立即跳跃。不幸的是，现在情况并非如此，我不知道为什么。我使用了 sounddevice 库，它似乎运行良好，但我无法使用 pyalsaaudio 获得相同的结果。

编辑： 根据 @fdcpp 的评论，我修改了我的脚本，但得到了负值，但不知道为什么

import alsaaudio
import audioop
import math
import struct

# Set the audio parameters
device = 'default'
channels = 1  # mono
sample_rate = 44100  # 44.1 kHz
format = alsaaudio.PCM_FORMAT_S16_LE

# Open the audio input stream
inp = alsaaudio.PCM(alsaaudio.PCM_CAPTURE, alsaaudio.PCM_NORMAL, device, channels=channels, rate=sample_rate, format=format, periodsize=1024)

# Main loop to continuously print audio levels
try:
    while True:
        # Read audio data from the microphone
        _, data = inp.read()

        # Convert byte data to 16-bit integer format. Have tried with and without this and get the same result.
        samples = struct.unpack('<' + ('h' * (len(data) // 2)), data)

        # Calculate the peak audio level in decibels
        peak_amplitude = max(map(abs, samples))

        # Convert amplitude to decibels
        decibels = 20 * math.log10(peak_amplitude / 32767)  # Assuming 16-bit audio
        print(f"Decibels: {decibels:.2f} dB")

except KeyboardInterrupt:
    print("Stopping the script.")

Answer 1

根据@fdcpp 的评论，这似乎效果很好。

import alsaaudio
import audioop
import math

# Set the audio parameters
device = 'default'
sample_rate = 44100  # 44.1 kHz

# Open the audio input stream
inp = alsaaudio.PCM(alsaaudio.PCM_CAPTURE, alsaaudio.PCM_NORMAL, device, channels=1, rate=sample_rate, format=alsaaudio.PCM_FORMAT_S16_LE, periodsize=1024)


# Main loop to continuously print audio levels
try:
    while True:
        # Read audio data from the microphone
        _, data = inp.read()

        # Calculate the peak audio level in decibels
        peak_amplitude = audioop.max(data, 2)

        # Convert amplitude to decibels
        decibels = 20 * math.log10(peak_amplitude)  # Assuming 16-bit audio
        print(f"Decibels: {decibels:.2f} dB")

except KeyboardInterrupt:
    print("Stopping the script.")