语音转文本 Ibm Watson C#，长音频超过 100 MB

Question

我正在实现 C# 代码，我试图在其中转录超过 100 mb 的音频，但它不允许我开发需要可以在 C# 中发送超过 100 mb 音频的程序
在这段代码中，我使用的是网络套接字，但我如何像流式传输音频一样发送

public static void CallWatson()
{
    using (var nf = new Notifier())
    using (var ws = new WebSocket("wss://stream.watsonplatform.net/speech-to-text/api/v1/recognize?watson-token=""))
    {
        string startActionjson = "{\"action\": \"start\", \"content-type\": \"audio/wav\", \"continuous\" : true, \"interim_results\": true}";
        ws.OnOpen += (sender, e) => ws.Send(startActionjson);

        // Set the WebSocket events.
        string result = string.Empty;

        ws.OnMessage += Ws_OnMessage;


        ws.OnError += (sender, e) =>
          nf.Notify(
            new NotificationMessage
            {
                Summary = "WebSocket Error",
                Body = e.Message,
                Icon = "notification-message-im"
            });

        ws.OnClose += (sender, e) =>
          nf.Notify(
            new NotificationMessage
            {
                Summary = String.Format("WebSocket Close ({0})", e.Code),
                Body = e.Reason,
                Icon = "notification-message-im"
            });

        ws.Connect();

        //ws.SendAsync(startActionjson, b =>
        //{
        //    if (b == true)
        //    {
        //        //send the audio as binary
        //        string filePath = "E:\\test33.wav";
        //        byte[] bytes = System.IO.File.ReadAllBytes(filePath);

        //        ws.SendAsync(bytes, b1 =>
        //        {
        //            if (b1)
        //                ws.Close();
        //        });

        //        // result+=result+ws.
        //    }
        //});
        // Connect to the server asynchronously.
        //ws.ConnectAsync ();

        //Console.WriteLine("\nType 'exit' to exit.\n");
        string filePath = "E:\\Test3.wav";
        byte[] bytes = System.IO.File.ReadAllBytes(filePath);
        while (true)
        {
            Thread.Sleep(3000);

            ws.SendAsync(bytes, b1 =>
            {
                if (b1)
                    ws.Close();
            });

        }

        //for (int i = 0; i < bytes.Length; i += 1000000)
        //{
        //    Thread.Sleep(1000);
        //    byte[] buffer = new byte[1000000];
        //    Buffer.BlockCopy(bytes, i, buffer, 0, 128);
        //  //  ws.Send(buffer);
        //    ws.SendAsync(buffer, b1 =>
        //    {
        //        if (b1)
        //            ws.Close();
        //    });
        //}
    }
}

private static void Ws_OnMessage(object sender, MessageEventArgs e)
{
    string s = e.Data;
}

Answer 1

根据文档，无论输入法如何，都有一个100mb limit。因此，您必须将音频文件拆分为

要流式传输音频，而不是调用 System.IO.File.ReadAllBytes(filePath); 并迭代结果，我认为您需要创建一个 FileStream。

此外，您不应在到达文件末尾后立即关闭 websocket - 这可能会阻止您接收所有结果。相反，发送字符串{"action": "stop"} 并等待，直到您收到{"state": "listening"} 的响应，其中indicates that it has completed processing your audio 并将所有文本发回。

更新：我拿到了一台 Windows 机器，安装了 Visual Studio，并整理了一个工作示例。我从来没有弄清楚您使用的是什么 WebSocket API/Library，但这只是使用了我可以在 microsoft.com 上找到文档的内置内容，因此希望它对您有用。

我用几个不同的 .ogg 和 .wav 文件对其进行了测试，并确认我得到了预期的多个中期和最终结果。

using System;
using System.Net.WebSockets;
using System.Net;
using System.Runtime.Serialization.Json;
using System.Threading;
using System.Threading.Tasks;
using System.Text;
using System.IO;
using System.Runtime.Serialization;


// Perform streaming transcription of an audio file using the IBM Watson Speech to Text service over a websocket
// http://www.ibm.com/smarterplanet/us/en/ibmwatson/developercloud/speech-to-text.html
// https://msdn.microsoft.com/en-us/library/system.net.websockets.clientwebsocket%28v=vs.110%29.aspx
namespace WatsonSTTWebsocketExample
{

    class Program
    {

        static void Main(string[] args)
        {
            Transcribe();
            Console.WriteLine("Press any key to exit");
            Console.ReadLine();
        }

        // http://www.ibm.com/smarterplanet/us/en/ibmwatson/developercloud/doc/getting_started/gs-credentials.shtml
        static String username = "<username>";
        static String password = "<password>";

        static String file = @"c:\audio.wav";

        static Uri url = new Uri("wss://stream.watsonplatform.net/speech-to-text/api/v1/recognize");
        static ArraySegment<byte> openingMessage = new ArraySegment<byte>( Encoding.UTF8.GetBytes(
            "{\"action\": \"start\", \"content-type\": \"audio/wav\", \"continuous\" : true, \"interim_results\": true}"
        ));
        static ArraySegment<byte> closingMessage = new ArraySegment<byte>(Encoding.UTF8.GetBytes(
            "{\"action\": \"stop\"}"
        ));


        static void Transcribe()
        {
            var ws = new ClientWebSocket();
            ws.Options.Credentials = new NetworkCredential(username, password);
            ws.ConnectAsync(url, CancellationToken.None).Wait();

            // send opening message and wait for initial delimeter 
            Task.WaitAll(ws.SendAsync(openingMessage, WebSocketMessageType.Text, true, CancellationToken.None), HandleResults(ws));

            // send all audio and then a closing message; simltaneously print all results until delimeter is recieved
            Task.WaitAll(SendAudio(ws), HandleResults(ws)); 

            // close down the websocket
            ws.CloseAsync(WebSocketCloseStatus.NormalClosure, "Close", CancellationToken.None).Wait();
        }

        static async Task SendAudio(ClientWebSocket ws)
        {

            using (FileStream fs = File.OpenRead(file))
            {
                byte[] b = new byte[1024];
                while (fs.Read(b, 0, b.Length) > 0)
                {
                    await ws.SendAsync(new ArraySegment<byte>(b), WebSocketMessageType.Binary, true, CancellationToken.None);
                }
                await ws.SendAsync(closingMessage, WebSocketMessageType.Text, true, CancellationToken.None);
            }
        }

        // prints results until the connection closes or a delimeterMessage is recieved
        static async Task HandleResults(ClientWebSocket ws)
        {
            var buffer = new byte[1024];
            while (true)
            {
                var segment = new ArraySegment<byte>(buffer);

                var result = await ws.ReceiveAsync(segment, CancellationToken.None);

                if (result.MessageType == WebSocketMessageType.Close)
                {
                    return;
                }

                int count = result.Count;
                while (!result.EndOfMessage)
                {
                    if (count >= buffer.Length)
                    {
                        await ws.CloseAsync(WebSocketCloseStatus.InvalidPayloadData, "That's too long", CancellationToken.None);
                        return;
                    }

                    segment = new ArraySegment<byte>(buffer, count, buffer.Length - count);
                    result = await ws.ReceiveAsync(segment, CancellationToken.None);
                    count += result.Count;
                }

                var message = Encoding.UTF8.GetString(buffer, 0, count);

                // you'll probably want to parse the JSON into a useful object here,
                // see ServiceState and IsDelimeter for a light-weight example of that.
                Console.WriteLine(message);

                if (IsDelimeter(message))
                {
                    return;
                }
            }
        }


        // the watson service sends a {"state": "listening"} message at both the beginning and the *end* of the results
        // this checks for that
        [DataContract]
        internal class ServiceState
        {
            [DataMember]
            public string state = "";
        }
        static bool IsDelimeter(String json)
        {
            MemoryStream stream = new MemoryStream(Encoding.UTF8.GetBytes(json));
            DataContractJsonSerializer ser = new DataContractJsonSerializer(typeof(ServiceState));
            ServiceState obj = (ServiceState)ser.ReadObject(stream);
            return obj.state == "listening";
        }

    }
}