如何播放.mp3文件从Amazon Polly文本到语音，而无需保存文件

Question

我正在使用Amazon Polly作为文本到语音，但是我不想将转换后的文件保存到PC检索。相反，我想播放它作为流。

我已经使用了playsound，但它不适用于返回的变量。

client = boto3.client('polly')
output = client.synthesize_speech (
  Text = "Some random text I want to convert", OutputFormat = "mp3", VoiceId = 'Aditi'
)
print(output['AudioStream'])
file = open('speech.mp3', 'wb')
file.write(output['AudioStream'].read())
file.close()

Answer 1

您可以使用pydub直接将MP3流加载到视觉中。

from pydub import AudioSegment
sound = AudioSegment.from_mp3(output['AudioStream'])

Answer 2

如果您不想在当地保存文件，则可以使用pyaudio直接播放流。根据您的用例，您可以阅读具有定义的缓冲区大小的流块，或者只是阅读整个流并播放：

import sys
from contextlib import closing
import boto3
import pyaudio
from botocore.exceptions import BotoCoreError, ClientError

session = boto3.Session(aws_access_key_id=ACCESS_KEY, aws_secret_access_key=SECRET_KEY)
polly = session.client("polly")
SAMPLE_RATE = 16000
READ_CHUNK = 4096
CHANNELS = 1
BYTES_PER_SAMPLE = 2
try:
    response = polly.synthesize_speech(Text="Hello world!",
                                       VoiceId="Joanna",
                                       LanguageCode="en-US",
                                       OutputFormat="pcm",
                                       SampleRate=str(SAMPLE_RATE))
except (BotoCoreError, ClientError) as error:
    print(error)
    sys.exit(-1)
p = pyaudio.PyAudio()
stream = p.open(format=p.get_format_from_width(BYTES_PER_SAMPLE),
                channels=CHANNELS,
                rate=SAMPLE_RATE,
                output=True)
with closing(response["AudioStream"]) as polly_stream:
    while True:
        data = polly_stream.read(READ_CHUNK)
        if data is None:
            break
        stream.write(data)
stream.stop_stream()
stream.close()
p.terminate()