Question

我正在尝试使用Microsoft语音对象库在C＃中将文本转换为音频。我将音频直接保存到wav文件时已成功完成此操作但是我的主要目标是将音频保存到字节数组，然后我可以将其写入asp.net中的响应（因此最终用户可以将其下载到他们的机器）。

当我尝试打开写入下载的响应的wav文件时，没有播放任何内容，并且说明了Windows媒体播放器无法打开文件的错误。

下面的代码显示了我的工作内容和不起作用的内容。

在尝试将字节数组作为wav写入响应时，任何人都对第二部分中可能缺少的内容有任何想法？

        ////////////////////////////////////////////////
        // THIS WORKS
        //SpVoice my_Voice = new SpVoice();                   //declaring and initializing SpVoice Class
        //SpeechVoiceSpeakFlags my_Spflag = SpeechVoiceSpeakFlags.SVSFlagsAsync; // declaring and initializing Speech Voice Flags

        //SpFileStream spFileStream = new SpFileStream();     //declaring and Initializing fileStream obj
        //SpeechStreamFileMode spFileMode = SpeechStreamFileMode.SSFMCreateForWrite;  //declaring fileStreamMode as to Create or Write
        //spFileStream.Open("C:\\temp\\hellosample.wav", spFileMode, false);
        //my_Voice.AudioOutputStream = spFileStream;
        //my_Voice.Speak("test text to audio in asp.net", my_Spflag);
        //my_Voice.WaitUntilDone(-1);
        //spFileStream.Close();
        ////////////////////////////////////////////////

        ////////////////////////////////////////////////
        // THIS DOES NOT WORK
        SpVoice my_Voice = new SpVoice();                   //declaring and initializing SpVoice Class
        SpeechVoiceSpeakFlags my_Spflag = SpeechVoiceSpeakFlags.SVSFlagsAsync; // declaring and initializing Speech Voice Flags

        SpMemoryStream spMemStream = new SpMemoryStream();
        spMemStream.Format.Type = SpeechAudioFormatType.SAFT11kHz8BitMono;
        object buf = new object();
        my_Voice.AudioOutputStream = spMemStream;
        my_Voice.Speak("test text to audio!", my_Spflag);
        my_Voice.WaitUntilDone(-1);
        spMemStream.Seek(0, SpeechStreamSeekPositionType.SSSPTRelativeToStart);
        buf = spMemStream.GetData();
        byte[] byteArray = (byte[])buf;
        Response.Clear();
        Response.ContentType = "audio/wav";
        Response.AppendHeader("Content-Disposition", "attachment; filename=mergedoutput.wav");
        Response.BinaryWrite(byteArray);
        Response.Flush();
        ////////////////////////////////////////////////

Answer 1

我建议您使用SpeechSynthesizer程序集中的System.Speech类而不是Microsoft Speech Object Library，因为程序集包含在.NET库中。

我在下面发布了一个示例，解释如何使用在ASP.NET MVC上创建的SpeechSynthesizer类来解决您的问题。我希望这可以解决你的问题。

public class HomeController : Controller
{
    public async Task<ActionResult> Index()
    {
        Task<FileContentResult> task = Task.Run(() =>
        {
            using (var synth = new SpeechSynthesizer())
            using (var stream = new MemoryStream())
            {
                synth.SetOutputToWaveStream(stream);
                synth.Speak("test text to audio!");
                byte[] bytes = stream.GetBuffer();
                return File(bytes, "audio/x-wav");
            }
        });

        return await task;
    }
}

Answer 2

我最终使用了微软翻译服务和这个.net lib（http://translatorservice.codeplex.com/）来连接它。效果很好。代码如下：

        // Connect to translator service
        SpeechSynthesizer speech = new SpeechSynthesizer("clientID", "secretKey");
        speech.AudioFormat = SpeakStreamFormat.MP3;
        speechStream = speech.GetSpeakStream(text, language);

        // Write it out to the stream
        Response.Clear();
        Response.ContentType = "audio/mp3";
        Response.AppendHeader("Content-Disposition", "attachment; filename=" + fileName + "");
        speechStream.CopyTo(Response.OutputStream);
        Response.Flush();

Answer 3

我使用ISpStream完成了它。使用ispstream的Setbasestream函数将其绑定到istream，然后将ispvoice的输出设置为ispstream。

如果有人想要的话，这是我的解决方案：

https://github.com/itsyash/MS-SAPI-demo

如何使用Microsoft语音对象库将wav字节数组写入响应？

3 个答案: