Question

我刚开始使用pyaudio，我写了一个简单的函数来播放音符。然而，根据我正在使用的Python版本，音符听起来有所不同：

from __future__ import division
import math
import pyaudio


BITS_PER_BYTE = 8  # for clarity
SAMPLE_BIT_DEPTH = 8  # i.e. each sample is 1 byte
SAMPLES_PER_SECOND = 16000
NOTE_TIME_SECONDS = 1
MIDDLE_C_HZ = 523.3

CYCLES_PER_SECOND = SAMPLES_PER_SECOND / MIDDLE_C_HZ
NUM_SAMPLES = SAMPLES_PER_SECOND * NOTE_TIME_SECONDS


def play_note():
    audio = pyaudio.PyAudio()

    stream = audio.open(
        format=audio.get_format_from_width(SAMPLE_BIT_DEPTH / BITS_PER_BYTE),
        channels=1,
        rate=SAMPLES_PER_SECOND,
        output=True,
    )

    byte_string = str()

    for i in range(NUM_SAMPLES):
        # calculate the amplitude for this frame as a float between -1 and 1
        frame_amplitude = math.sin(i / (CYCLES_PER_SECOND / math.pi))
        # scale the amplitude to an integer between 0 and 255 (inclusive)
        scaled_amplitude = int(frame_amplitude * 127 + 128)
        # convert amplitude to byte string (ascii value)
        byte_string += chr(scaled_amplitude)

    stream.write(byte_string)
    stream.close()
    audio.terminate()


if __name__ == '__main__':
    play_note()

在Python 2.7.13中，我听到了正确，清晰的语调。在3.6.2中，它听起来很粗糙，就像方波一样。

为什么会这样，我将如何解决这个问题（或者至少开始调试）？

我使用portaudio v10.11.6在OSX v19.6.0上。

Answer 1

这是因为您应该使用str时使用byte。

这对我有用：

byte_array = bytearray()  # bytearray instead of str

for i in range(NUM_SAMPLES):
    frame_amplitude = math.sin(i / (CYCLES_PER_SECOND / math.pi))
    scaled_amplitude = int(frame_amplitude * 127 + 128)
    # Note the append here, not +=
    byte_array.append(scaled_amplitude)

stream.write(bytes(byte_array))

pythonio在python2和python3中播放不同的音调

1 个答案: