排查Azure语音转文本代理或防火墙相关问题

我基于Azure开发了一款语音转文本的概念验证程序，Azure端配置已完成，代码如下：

using System;
using System.IO;
using System.Threading.Tasks;
using Microsoft.CognitiveServices.Speech;
using Microsoft.CognitiveServices.Speech.Audio;

class Program
{

    static SpeechRecognizer recognizer;

    async static Task FromMic(SpeechConfig speechConfig)
    {
        using var audioConfig = AudioConfig.FromDefaultMicrophoneInput();
        recognizer = new SpeechRecognizer(speechConfig, audioConfig);

        var stopRecognition = new TaskCompletionSource<int>();

        recognizer.SessionStarted += (s, e) =>
        {
            Console.WriteLine("\n    Session started event: " + e);
        };

        recognizer.Recognizing += (s, e) =>
        {
            Console.WriteLine($"RECOGNIZING: Text={e.Result.Text}");
        };

        recognizer.Recognized += (s, e) =>
        {
            if (e.Result.Reason == ResultReason.RecognizedSpeech)
            {
                Console.WriteLine($"RECOGNIZED: Text={e.Result.Text}");
            }
            else if (e.Result.Reason == ResultReason.NoMatch)
            {
                Console.WriteLine($"NOMATCH: Speech could not be recognized.");
            }
        };

        recognizer.Canceled += (s, e) =>
        {
            Console.WriteLine($"CANCELED: Reason={e.Reason}");

            if (e.Reason == CancellationReason.Error)
            {
                Console.WriteLine($"CANCELED: ErrorCode={e.ErrorCode}");
                Console.WriteLine($"CANCELED: ErrorDetails={e.ErrorDetails}");
                Console.WriteLine($"CANCELED: Did you set the speech resource key and region values?");
            }

            stopRecognition.TrySetResult(0);
        };

        recognizer.SessionStopped += (s, e) =>
        {
            Console.WriteLine("\n    Session stopped event.");
            stopRecognition.TrySetResult(0);
        };

        await recognizer.StartContinuousRecognitionAsync();

        // Waits for completion. Use Task.WaitAny to keep the task rooted.
        Task.WaitAny(new[] { stopRecognition.Task });

    }

    async static Task Main(string[] args)
    {
        var speechConfig = SpeechConfig.FromSubscription("xxxxxxxxxxxxxxxxxxxx", "xxxx");

        // Clearly I don't really know how to stop the recognition properly before exiting, but that's an issue for another day...
        AppDomain.CurrentDomain.ProcessExit += delegate
        {
            EndRecognition();
        };
        Console.CancelKeyPress += delegate
        {
            EndRecognition();
        };
        await FromMic(speechConfig);
        Console.WriteLine("Exiting");
    }

    static void EndRecognition()
    {
        Console.WriteLine("Ending recognition...");
        recognizer.StopContinuousRecognitionAsync();
        recognizer.Dispose();
        Console.WriteLine("Done.");
    }
}

程序在家用电脑运行正常，但在公司电脑上仅能收到会话启动消息，无法完成语音识别。公司网络为严格防火墙规则下的代理路由环境，已完成以下排查：

• 确认麦克风正常可用
• 设置HTTP_PROXY和HTTPS_PROXY环境变量指向公司代理
• 查看杀毒防火墙日志，无相关拦截记录
• 查看Azure端“总调用数”指标，无数据

预期无法连接Azure语音资源时会抛出异常，但实际无报错信息，问题静默发生。

下一步排查步骤 #

• 验证代理配置是否被Speech SDK正确读取：在代码中添加日志输出Environment.GetEnvironmentVariable("HTTP_PROXY")和HTTPS_PROXY的值，确认程序能获取到代理地址；同时需确保代理支持HTTPS，并允许访问对应区域的Azure语音服务端点（格式为https://<region>.stt.speech.microsoft.com）
• 手动测试端点连通性：在公司电脑上用PowerShell或curl测试访问Azure语音服务端点，比如执行Invoke-WebRequest -Uri https://<your-region>.stt.speech.microsoft.com/ -Proxy http://your-proxy-address:port，检查是否能正常获取响应，排查网络层连通性
• 启用Speech SDK详细日志：在初始化SpeechConfig后添加以下代码开启日志，日志会记录详细的网络交互和潜在错误，帮助定位问题：

  speechConfig.SetProperty(PropertyId.Speech_LogFilename, @"C:\temp\speech-sdk.log");
  speechConfig.SetProperty(PropertyId.Speech_LogLevel, "Detailed");
  

• 配置代理身份验证（若需）：如果公司代理需要用户名密码，仅设置环境变量可能无效，需在代码中显式配置：

  speechConfig.SetProperty(PropertyId.Proxy_Host, "your-proxy-host");
  speechConfig.SetProperty(PropertyId.Proxy_Port, "your-proxy-port");
  speechConfig.SetProperty(PropertyId.Proxy_Username, "proxy-username");
  speechConfig.SetProperty(PropertyId.Proxy_Password, "proxy-password");
  

• 检查WebSocket流量是否被拦截：Azure语音转文本使用WebSocket传输音频数据，部分企业防火墙可能拦截此类连接，可尝试用工具测试WebSocket连通性，比如用wscat执行wscat -w 0 -x wss://<your-region>.stt.speech.microsoft.com/speech/recognition/conversation/cognitiveservices/v1 -H "Ocp-Apim-Subscription-Key: your-azure-key"
• 统一SDK版本：确认公司电脑使用的Speech SDK版本与家用电脑一致，旧版本可能存在代理兼容性问题，尝试更新到最新稳定版
• 验证本地音频处理：临时使用Speech SDK的离线识别功能测试，确认问题是否出在网络层面而非本地音频采集或处理环节

内容的提问来源于stack exchange，提问作者user2686455