ASP.NET Core 2.0文本转MP3音频：替代System.Speech方案求助

Hey there, I’ve been in your shoes before—losing access to System.Speech when moving to .NET Core was a frustrating roadblock for text-to-speech (TTS) tasks. Let me break down the most reliable alternatives that work seamlessly with ASP.NET Core 2.0:

1. Azure Cognitive Services Speech SDK (推荐，跨平台、稳定) #

This is the most robust option for production use, as it’s fully supported for .NET Core and handles MP3 generation natively. Here’s how to set it up:

• Install the NuGet package:

  Install-Package Microsoft.CognitiveServices.Speech
  

• Sample code in an ASP.NET Core controller:

  using Microsoft.CognitiveServices.Speech;
  using Microsoft.CognitiveServices.Speech.Audio;
  using Microsoft.AspNetCore.Mvc;
  using System.IO;
  using System.Threading.Tasks;
  
  [ApiController]
  [Route("api/tts")]
  public class TextToSpeechController : ControllerBase
  {
      [HttpPost]
      public async Task<IActionResult> GenerateMp3([FromBody] string text)
      {
          // Replace with your Azure Speech resource key and region
          var config = SpeechConfig.FromSubscription("YOUR_SPEECH_KEY", "YOUR_REGION");
          config.SetSpeechSynthesisOutputFormat(SpeechSynthesisOutputFormat.Audio16Khz32KBitRateMonoMp3);
  
          using var audioOutput = AudioConfig.FromWavFileOutput(Path.GetTempFileName() + ".mp3");
          using var synthesizer = new SpeechSynthesizer(config, audioOutput);
  
          var result = await synthesizer.SpeakTextAsync(text);
  
          if (result.Reason == ResultReason.SynthesizingAudioCompleted)
          {
              var mp3Bytes = System.IO.File.ReadAllBytes(audioOutput.OutputFilePath);
              System.IO.File.Delete(audioOutput.OutputFilePath); // Clean up temp file
              return File(mp3Bytes, "audio/mpeg", "output.mp3");
          }
          else
          {
              return BadRequest($"TTS failed: {result.Reason}");
          }
      }
  }
  

  Note: You’ll need an Azure account and a Speech resource to get the key/region—this offers a free tier for testing.

2. NAudio + eSpeak (开源免费，本地运行) #

If you don’t want to rely on cloud services, use NAudio with the open-source eSpeak TTS engine. It runs locally but requires a bit more setup:

• Install NuGet packages:

  Install-Package NAudio
  Install-Package NAudio.Vorbis
  

• Setup steps:
  1. Download the eSpeak binaries from its official repo and extract them to your project’s wwwroot or a dedicated folder.
  2. Use NAudio to interface with eSpeak and convert its output to MP3:

  using NAudio.Wave;
  using System.Diagnostics;
  using System.IO;
  
  public byte[] GenerateMp3Local(string text)
  {
      var tempWavPath = Path.GetTempFileName() + ".wav";
      var tempMp3Path = Path.GetTempFileName() + ".mp3";
  
      // Call eSpeak to generate WAV file
      var process = new ProcessStartInfo
      {
          FileName = "path/to/espeak.exe",
          Arguments = $"\"{text}\" -w {tempWavPath}",
          UseShellExecute = false,
          CreateNoWindow = true
      };
      Process.Start(process).WaitForExit();
  
      // Convert WAV to MP3 using NAudio
      using var reader = new WaveFileReader(tempWavPath);
      using var writer = new LameMP3FileWriter(tempMp3Path, reader.WaveFormat, 128);
      reader.CopyTo(writer);
  
      var mp3Bytes = File.ReadAllBytes(tempMp3Path);
      // Clean up temp files
      File.Delete(tempWavPath);
      File.Delete(tempMp3Path);
  
      return mp3Bytes;
  }
  

  Keep in mind: eSpeak’s voice quality is more basic compared to cloud services, but it’s great for offline scenarios.

Why ScriptCs didn’t work #

ScriptCs is built around .NET Framework components, so it often relies on APIs like System.Speech that aren’t available in .NET Core. It’s not a viable solution for Core-based projects.

内容的提问来源于stack exchange，提问作者Vikram V