Python で音響信号処理：オリジナルの Python スペクトルアナライザーに入力デバイス選択の処理、ケプストラム分析の実験を追加。

readme.md

Python で音響信号処理 spectrum.py

2016-07-16 Takuya Nishimoto (@24motz)

広島版IoT縛りの勉強会! IoTLT広島 vol.2

• http://iotlt.connpass.com/event/33441/

$ sudo apt-get install portaudio19-dev python-dev
$ curl https://bootstrap.pypa.io/get-pip.py | sudo python
$ sudo pip install pyaudio
$ sudo pip install numpy

• Raspberry Pi 3 (Raspbian) で USB Audio Interface を接続、マイクをつなぐ
• python spectrum.py を実行
• オーディオミキサーで入力ゲインを調節

DragonBoard 410c (Debian) でも動作確認している。

スペクトルの表示

• 濃淡をそれっぽく変更

FFTケプストラムの表示

• callback() の self.fft(data) を self.cepstrum(data) に変更。

PIP

• https://github.com/pypa/get-pip

ケプストラム分析

• ケプストラム (ウィキペディア）
• Cepstrum (Wikipedia)
• ケプストラムって要するに何ですか? from 音響学入門ペディア

PyAudio

• https://people.csail.mit.edu/hubert/pyaudio/

numpy fromstring

• http://docs.scipy.org/doc/numpy/reference/generated/numpy.fromstring.html

numpy fft

• http://docs.scipy.org/doc/numpy/reference/routines.fft.html

spectrum.py

#!/usr/bin/env python
# encoding: utf-8

from __future__ import unicode_literals, print_function

import sys
import os
import atexit
import time

import numpy as np
import pyaudio


class Spectrum(object):

    FORMAT = pyaudio.paFloat32
    CHANNELS = 1
    RATE = 16000
    FRAME_LEN = 512
    DEVICE_NAME = 'USB Audio'
    PATTERNS = ' .,:;|#@89'
    OUTPUT_WIDTH = 200

    def __init__(self):
        self.pa = pyaudio.PyAudio()
        self.last_samples = None
        atexit.register(self.pa.terminate)
        self.input_device_index = None
        for i in xrange(self.pa.get_device_count()):
            dev = self.pa.get_device_info_by_index(i)
            if self.DEVICE_NAME in dev.get('name','') and \
               dev.get('maxInputChannels',0) > 0:
                self.input_device_index = i
                print('using %r' % dev)
        self.win = np.hamming(self.FRAME_LEN)

    def fft(self, samples):
        res = np.fft.fft(self.win * samples)
        res = res[:len(res)/2]
        res = np.log(np.abs(res)**2)
        res += 10
        return res[:self.OUTPUT_WIDTH]

    def cepstrum(self, samples):
        res = np.fft.fft(self.win * samples)
        res = np.log(np.abs(res)**2)
        res = np.fft.ifft(res)
        res = res[:len(res)/2]
        res = np.real(res)
        p = 70
        res = res[p:p+self.OUTPUT_WIDTH]
        res *= 30.0
        return res

    def valueToChar(self, value):
        #return "%.1f " % value
        try:
            return self.PATTERNS[min(9, max(0, int(value)))]
        except ValueError, OverflowError:
            return ' '

    def callback(self, in_data, frame_count, time_info, status):
        data = np.fromstring(in_data, np.float32)
        if len(data) != self.FRAME_LEN:
            return
        d = self.fft(data)
        print(''.join([self.valueToChar(v) for v in d]))
        return (in_data, self.recording)

    def record(self):
        self.recording = pyaudio.paContinue
        stream = self.pa.open(format = self.FORMAT,
                        channels = self.CHANNELS, 
                        rate = self.RATE, 
                        input = True,
                        input_device_index = self.input_device_index,
                        output = False,
                        frames_per_buffer = self.FRAME_LEN,
                        stream_callback = self.callback)
        stream.start_stream()

        while stream.is_active():
            try:
                time.sleep(1)
            except KeyboardInterrupt:
                self.recording = pyaudio.paAbort

        stream.start_stream()
        stream.close()

if __name__ == '__main__':
    spe = Spectrum()
    spe.record()