python Aubio检测到频率超出范围

时间:2018-12-01 13:13:57

标签: python-3.x frequency-analysis pitch-tracking audio-analysis aubio

我目前正在使用称为Aubio的python模块进行实时音高跟踪,然后计算标准偏差以了解说话者的语调。但是,我得到的检测结果包括一个非常高的频率(如5067)和一个非常低的频率(42 Hz)。这让我感到困惑,并且我认为结果很奇怪,因为女性的正常基础频率(音高)范围是165〜255 Hz。有人知道怎么做吗?谢谢。

下面是我当前的代码。

import aubio
import numpy as np
import pyaudio
import sys
import time
from statistics import stdev

# Some constants for setting the PyAudio and the
# Aubio.
BUFFER_SIZE             = 2048
CHANNELS                = 1
FORMAT                  = pyaudio.paFloat32
METHOD                  = "default"
SAMPLE_RATE             = 44100
HOP_SIZE                = BUFFER_SIZE//2
PERIOD_SIZE_IN_FRAME    = HOP_SIZE

def main(args):

    # Initiating PyAudio object.
    pA = pyaudio.PyAudio()
    # Open the microphone stream.
    mic = pA.open(format=FORMAT, channels=CHANNELS,
        rate=SAMPLE_RATE, input=True,
        frames_per_buffer=PERIOD_SIZE_IN_FRAME)

    # Initiating Aubio's pitch detection object.
    pDetection = aubio.pitch(METHOD, BUFFER_SIZE,
        HOP_SIZE, SAMPLE_RATE)
    # Set unit.
    pDetection.set_unit("Hz")
    # Frequency under -40 dB will considered
    # as a silence.
    pDetection.set_silence(-40)

    sound_data = []
    end_time = time.time() + 5
    while time.time() < end_time:

        # Always listening to the microphone.
        data = mic.read(PERIOD_SIZE_IN_FRAME)
        # Convert into number that Aubio understand.
        samples = np.fromstring(data,
            dtype=aubio.float_type)
        # Finally get the pitch.
        pitch = pDetection(samples)[0]
        # Compute the energy (volume)
        # of the current frame.
        #volume = num.sum(samples**2)/len(samples)
        # Format the volume output so it only
        # displays at most six numbers behind 0.
        #volume = "{:6f}".format(volume)

        # Finally print the pitch and the volume.
        if pitch:
            sound_data.append(int(pitch))
    print(sound_data)
    data = np.array([sound_data])
    std = np.std(data)
    heikin = np.average(data)
    print('{0:.2f}'.format(std))
    print(std)
    print(heikin)

if __name__ == "__main__": main(sys.argv)

0 个答案:

没有答案