Equalizer in Raspberry Pi and moc

As an example from which to start, you can write something like this:

import requests import matplotlib.pyplot as plt import numpy as np from io import BytesIO import wave import pyaudio test_wav_request = requests.get("http://download.wavetlan.com/SVV/Media/HTTP/WAV/Media-Convert/Media-Convert_test6_PCM_Stereo_VBR_16SS_8000Hz.wav") if test_wav_request.status_code != 200: print("Не могу скачать WAV. Код:", test_wav_request.status_code) exit(1) wav_file = BytesIO(initial_bytes=test_wav_request.content) wav = wave.open(wav_file) (nchannels, sampwidth, framerate, nframes, comptype, compname) = wav.getparams() # content = wav.readframes(nframes) width_to_bits = {1: np.int8, 2: np.int16, 4: np.int32} width_to_bits_ = {1: 8, 2: 16, 4: 32} width_pyaudio = {1: pyaudio.paInt8, 2: pyaudio.paInt16, 4: pyaudio.paInt32} audio = pyaudio.PyAudio() stream = audio.open(format=width_pyaudio[sampwidth], channels=nchannels, rate=framerate, input=False, output=True) print((nchannels, sampwidth, framerate, nframes, comptype, compname)) sample_size = 4 * 512 def normalize(samples, bits_per): return [s / 2**bits_per for s in samples] # Для примера возьмем последовательно 4 сэмпла размером 4*512 кадров. # Для данного файла частота 8000 кадров в секунду, то есть возьмем кадров за 0.25 секунд stream.start_stream() for i in range(1, 5): sample = wav.readframes(sample_size) # Для данного файла точно известно, что количество каналов 2 data = normalize(np.fromstring(sample, width_to_bits[sampwidth]), width_to_bits_[sampwidth]) stream.write(sample) left_channel = data[::2] right_channel = data[1::2] furier_left = np.abs(np.fft.rfft(left_channel)) furier_right = np.abs(np.fft.rfft(right_channel)) # Не знаю, что делают с разными каналами, но можно их объединить spectrum = np.hstack((furier_left[: int(len(furier_left) / 2)], furier_right[: int(len(furier_right) / 2)])) Hz = np.fft.fftfreq(len(left_channel), d=0.25) Hz *= framerate # По какой-то неизвестной причине в furier_left лишний элемент plt.subplot(2, 2, i) plt.plot(Hz[:len(Hz) // 2], 10 * np.log10(furier_left[:-1])) plt.ylabel("dB") plt.xlabel("Hz. Number: " + str(i)) stream.stop_stream() plt.show() 

This script will download the test wav file with the bleating of a sheep, and build 4 graphs - every 0.25 seconds, the sound is extracted, normalized, and the Fourier transform is performed. Then it calculates the decibel values ​​and displays the graph - along the X axis - Hertz, along the Y axis - Decibels. Here is one of the graphs - http://imgur.com/j1ApCEB . 4 graphics in one picture: http://imgur.com/vgQq0YA Honestly, I don’t know if this looks like an "equalizer" because Google says that the equalizer is a device.