IBM Watson 文本到语音音频文件在合成后无法播放

Question

我正在做的是写入音频输出文件，等到文件存在并且大小不是 0，然后播放它（我尝试了许多不同的库，例如 subprocess、playsound、pygame、vlc 等. 我也尝试了许多不同的文件类型 mp3、wav 等），但由于某种原因，我收到一个错误，说它没有关闭或已损坏。偶尔它会播放一次，但一旦播放另一个 watson 制作的 mp3，它就会再次出错。有谁知道解决办法吗？

...
from ibm_watson import TextToSpeechV1
from ibm_cloud_sdk_core.authenticators import IAMAuthenticator
...
authenticator = IAMAuthenticator(ibmApiKey);
textToSpeech = TextToSpeechV1(authenticator = authenticator);
textToSpeech.set_service_url(ibmServiceUrl);
...
file = str(int(random.random() * 100000)) + ".mp3";
    with open(file, "wb") as audioFile:
        audioFile.write(textToSpeech.synthesize(text, voice = "en-GB_JamesV3Voice", accept = "audio/mp3").get_result().content);

    fileExists = False;

    while (fileExists == False):
        if (os.path.isfile(file)):
            fileExists = os.stat(file).st_size != 0;
            playsound(file);
            os.remove(file);

Error 263 for command:
        open temp/77451.mp3
    The specified device is not open or is not recognized by MCI.

    Error 263 for command:
        close temp/77451.mp3
    The specified device is not open or is not recognized by MCI.
Failed to close the file: temp/77451.mp3
Traceback (most recent call last):
  File "main.py", line 457, in <module>
    runMain(name, config.get("main", "callName"), voice);
  File "main.py", line 156, in runMain
    speak("The time is: " + datetime.now().strptime(datetime.now().time().strftime("%H:%M"), "%H:%M").strftime("%I:%M %p"), voice);
  File "main.py", line 123, in speak
    playsound(file);
  File "C:\Users\turtsis\AppData\Local\Programs\Python\Python35-32\lib\site-packages\playsound.py", line 72, in _playsoundWin
    winCommand(u'open {}'.format(sound))
  File "C:\Users\turtsis\AppData\Local\Programs\Python\Python35-32\lib\site-packages\playsound.py", line 64, in winCommand
    raise PlaysoundException(exceptionMessage)
playsound.PlaysoundException:
    Error 263 for command:
        open temp/77451.mp3
    The specified device is not open or is not recognized by MCI.

Answer 1

错误可能存在于不同的地方。

首先我会试试这个：

from ibm_watson import ApiException
try:
    file = str(int(random.random() * 100000)) + ".mp3";
        with open(file, "wb") as audioFile:
            audioFile.write(textToSpeech.synthesize(text, voice = "en-GB_JamesV3Voice", accept = "audio/mp3").get_result().content);
except ApiException as ex:
    print ("Method failed with status code " + str(ex.code) + ": " + ex.message)

如果对 Watson 的调用返回错误，它可能会将您从运行时中弹出。

但是，如果问题出在 playsound 上，我会建议这条路线：

import pyttsx3
from ibm_watson import ApiException

engine = pyttsx3.init()
try:
    file = str(int(random.random() * 100000)) + ".mp3";
        with open(file, "wb") as audioFile:
            audioFile.write(textToSpeech.synthesize(text, voice = "en-GB_JamesV3Voice", accept = "audio/mp3").get_result().content);

        fileExists = False;

        while (fileExists == False):
            if (os.path.isfile(file)):
                fileExists = os.stat(file).st_size != 0;
                engine.say(file);
                os.remove(file); 
                engine.runAndWait()          

except ApiException as ex:
    print ("Method failed with status code " + str(ex.code) + ": " + ex.message)

如果这些都不起作用，我会尝试使用 curl 看看你是否可以复制你的场景：

Replace {apikey} and {url} with your API key and URL.


curl -X POST -u "apikey:{apikey}" --header "Content-Type: application/json" --data "{\"text\":\"hello world\"}" --output hello_world.ogg "{url}/v1/synthesize?voice=en-US_AllisonV3Voice"

祝你好运。