谷歌TTS调用synthesizeSpeech无响应及gRPC超时问题求助

问题描述 #

使用Google TextToSpeechClient进行文本转语音合成时，程序执行到SynthesizeSpeechResponse response = textToSpeechClient.synthesizeSpeech(input, voice, audioConfig);行时无响应停滞，此前还出现过gRPC deadline异常，现有解决方案多与当前版本不兼容。相关代码如下：

try (TextToSpeechClient textToSpeechClient = TextToSpeechClient.create(settings))
{
    System.out.println("In Try - Create TTS client");
    logger.info("In Try - Create TTS client");
    // Set the text input to be synthesized
    SynthesisInput input = SynthesisInput.newBuilder().setText(receivedtext).build();

    // Build the voice request, select the language code ("en-US") and the ssml
    VoiceSelectionParams voice = VoiceSelectionParams.newBuilder().setLanguageCode("en-US")
            .setSsmlGender(SsmlVoiceGender.FEMALE).build();
    logger.info("Set Voice Params");

    // Select the type of audio file you want returned
    AudioConfig audioConfig = AudioConfig.newBuilder().setAudioEncoding(AudioEncoding.MP3).build();
    logger.info("Audio Configured");
    
    // Perform the text-to-speech request
    SynthesizeSpeechResponse response = textToSpeechClient.synthesizeSpeech(input, voice, audioConfig);
    logger.info("Audio Synthesised");
    
    // Get the audio contents from the response
    ByteString audioContents = response.getAudioContent();

    // Write the response to the output file
    try (OutputStream out = new FileOutputStream(audioFileDirectory + audioFileName + audioFileFormat))
    {
        out.write(audioContents.toByteArray());
        logger.info("File Opened and Written");
    }
}

适配当前版本的解决方案 #

• 显式配置gRPC超时与重试策略
  当前Google Cloud客户端支持通过TextToSpeechSettings自定义超时和重试规则，避免无限等待或超时异常。示例代码：

// 构建包含超时与重试的Settings
TextToSpeechSettings settings = TextToSpeechSettings.newBuilder()
    .synthesizeSpeechSettings()
    .setRetrySettings(RetrySettings.newBuilder()
        .setTotalTimeout(Duration.ofMinutes(2)) // 总请求超时时间
        .setInitialRetryDelay(Duration.ofSeconds(1))
        .setMaxRetryDelay(Duration.ofSeconds(5))
        .setRetryDelayMultiplier(1.5)
        .setMaxAttempts(3)
        .build())
    .build();

// 使用配置后的Settings创建客户端
try (TextToSpeechClient textToSpeechClient = TextToSpeechClient.create(settings)) {
    // 原有业务逻辑代码...
}

• 检查并配置网络代理
  若运行环境存在代理，需为gRPC显式配置代理参数，避免请求被阻塞：

// 系统级代理配置
System.setProperty("http.proxyHost", "your-proxy-host");
System.setProperty("http.proxyPort", "your-proxy-port");
System.setProperty("https.proxyHost", "your-proxy-host");
System.setProperty("https.proxyPort", "your-proxy-port");

// 或通过Channel自定义代理
ManagedChannel channel = ManagedChannelBuilder.forTarget(TextToSpeechSettings.getDefaultEndpoint())
    .setProxyDetector(() -> Proxy.newProxyInstance(
        ClassLoader.getSystemClassLoader(),
        new Class[]{HttpURLConnection.class},
        new ProxySelector(null)
    ))
    .build();

TextToSpeechSettings settings = TextToSpeechSettings.newBuilder()
    .setTransportChannelProvider(TextToSpeechSettings.defaultHttpJsonTransportProviderBuilder()
        .setChannel(channel)
        .build())
    .build();

• 验证认证凭据有效性
  请求停滞可能源于凭据加载失败导致的静默等待，确保GOOGLE_APPLICATION_CREDENTIALS环境变量配置正确，或在代码中显式加载凭据：

GoogleCredentials credentials = GoogleCredentials.fromStream(new FileInputStream("path/to/your/service-account-key.json"));
TextToSpeechSettings settings = TextToSpeechSettings.newBuilder()
    .setCredentialsProvider(FixedCredentialsProvider.create(credentials))
    .build();

• 切换为异步调用避免主线程阻塞
  若同步调用持续停滞，可尝试异步调用方式，通过回调处理结果，避免主线程阻塞：

textToSpeechClient.synthesizeSpeechCallable().futureCall(input, voice, audioConfig)
    .addListener(() -> {
        try {
            SynthesizeSpeechResponse response = textToSpeechClient.synthesizeSpeechCallable().futureCall(input, voice, audioConfig).get();
            // 处理音频响应逻辑，如写入文件等...
            ByteString audioContents = response.getAudioContent();
            try (OutputStream out = new FileOutputStream(audioFileDirectory + audioFileName + audioFileFormat)) {
                out.write(audioContents.toByteArray());
                logger.info("音频文件写入完成");
            }
        } catch (Exception e) {
            logger.error("语音合成失败", e);
        }
    }, MoreExecutors.directExecutor());

内容的提问来源于stack exchange，提问作者Thomas A Mathew