I'm using Microsoft.CongnitiveServices.Speech SDK to convert text to speech by creating a custom API in C#. The return datatype is 'AudioDataStream'. However, I need to convert it to Stream or bytes so that I can use it in my ui. How can I convert it to a Stream or bytes or is there any other way I can use a stream output in my ui?

Edit: Here is the code I'm using. It gives me the error saying 'Cannot convert Microsoft.CongnitiveServices.Speech.AudioDataStream to System.IO.Stream. I also tried using AudioDataStream.ReadData() but it gives the return type as 'uint' not as bytes.

    [HttpPost]
    public async Task<IActionResult> ConversationIOStreamAsync(IFormFile audioFile)
    {
        if (audioFile == null || audioFile.Length == 0)
        {
            return BadRequest("No audio file received.");
        }

        // Save the audio file to a temporary location 
        var filePath = Path.Combine(Path.GetTempPath(), audioFile.FileName);
        using (var stream = new FileStream(filePath, FileMode.Create))
        {
            await audioFile.CopyToAsync(stream);
        }

        var memoryStream = new MemoryStream();
        using (var stream = new FileStream(filePath, FileMode.Open, FileAccess.Read))
        {
            await stream.CopyToAsync(memoryStream);
        }
        memoryStream.Position = 0;

        try
        {
            var audioStream = await ConversationService.ConversationIOStreamAsync(memoryStream);

            using var stream = AudioDataStream.FromResult(audioStream);


            return File(stream, "audio/wav", audioFile.FileName);
        }
        catch (System.Exception ex)
        {
            return StatusCode(500, $"Internal server error: {ex.Message}");
        }

    }

I'm using Microsoft.CongnitiveServices.Speech SDK to convert text to speech by creating a custom API in C#. The return datatype is 'AudioDataStream'. However, I need to convert it to Stream or bytes so that I can use it in my ui. How can I convert it to a Stream or bytes or is there any other way I can use a stream output in my ui?

Edit: Here is the code I'm using. It gives me the error saying 'Cannot convert Microsoft.CongnitiveServices.Speech.AudioDataStream to System.IO.Stream. I also tried using AudioDataStream.ReadData() but it gives the return type as 'uint' not as bytes.

    [HttpPost]
    public async Task<IActionResult> ConversationIOStreamAsync(IFormFile audioFile)
    {
        if (audioFile == null || audioFile.Length == 0)
        {
            return BadRequest("No audio file received.");
        }

        // Save the audio file to a temporary location 
        var filePath = Path.Combine(Path.GetTempPath(), audioFile.FileName);
        using (var stream = new FileStream(filePath, FileMode.Create))
        {
            await audioFile.CopyToAsync(stream);
        }

        var memoryStream = new MemoryStream();
        using (var stream = new FileStream(filePath, FileMode.Open, FileAccess.Read))
        {
            await stream.CopyToAsync(memoryStream);
        }
        memoryStream.Position = 0;

        try
        {
            var audioStream = await ConversationService.ConversationIOStreamAsync(memoryStream);

            using var stream = AudioDataStream.FromResult(audioStream);


            return File(stream, "audio/wav", audioFile.FileName);
        }
        catch (System.Exception ex)
        {
            return StatusCode(500, $"Internal server error: {ex.Message}");
        }

    }

• c#
• azure
• text-to-speech
• audio-streaming

• Can you share your code in the question? – Dasari Kamali Commented Mar 19 at 7:42
• Have you tried ReadData? i.e. create a loop that reads a buffer and writes it to the stream until the return value from ReadData is zero. – JonasH Commented Mar 19 at 7:45
• 1 @Maryam Mirza Use AudioDataStream.ReadData() to read the stream into a MemoryStream and convert it to byte[] or Stream. – Dasari Kamali Commented Mar 19 at 7:48
• Please provide enough code so others can better understand or reproduce the problem. – Community Bot Commented Mar 19 at 14:47

1 Answer 1

I have created an ASP.NET Web API project that converts AudioDataStream to a MemoryStream by reading the audio data in chunks using ReadData(), then wraps it in a WAV format before returning it as a Stream. This allows the API to send the audio as a downloadable file in the response.

Below is my complete code.

SpeechService.cs :

using Microsoft.CognitiveServices.Speech;

namespace WebApplication15.Services
{
    public class SpeechService
    {
        private readonly string _subscriptionKey;
        private readonly string _region;

        public SpeechService(IConfiguration configuration)
        {
            _subscriptionKey = "<speechKey>";
            _region = "<speechRegion>";
        }

        public async Task<Stream> ConvertTextToSpeechAsync(string text)
        {
            var config = SpeechConfig.FromSubscription(_subscriptionKey, _region);
            using var synthesizer = new SpeechSynthesizer(config, null);
            var result = await synthesizer.SpeakTextAsync(text);
            if (result.Reason != ResultReason.SynthesizingAudioCompleted)
            {
                throw new Exception($"Speech synthesis failed: {result.Reason}");
            }

            using var audioStream = AudioDataStream.FromResult(result);
            var memoryStream = new MemoryStream();

            byte[] buffer = new byte[4096];
            uint bytesRead;
            while ((bytesRead = audioStream.ReadData(buffer)) > 0)
            {
                memoryStream.Write(buffer, 0, (int)bytesRead);
            }
            var wavStream = ConvertToWaveFormat(memoryStream);
            return wavStream;
        }

        private Stream ConvertToWaveFormat(MemoryStream rawAudioStream)
        {
            rawAudioStream.Position = 0;
            using var reader = new BinaryReader(rawAudioStream);
            using var wavStream = new MemoryStream();

            using var writer = new BinaryWriter(wavStream);
            int sampleRate = 16000;
            int bitsPerSample = 16;
            int channels = 1;

            int byteRate = sampleRate * channels * (bitsPerSample / 8);
            int dataSize = (int)rawAudioStream.Length;

            writer.Write("RIFF".ToCharArray());
            writer.Write(36 + dataSize);
            writer.Write("WAVE".ToCharArray());
            writer.Write("fmt ".ToCharArray());
            writer.Write(16);
            writer.Write((short)1);
            writer.Write((short)channels);
            writer.Write(sampleRate);
            writer.Write(byteRate);
            writer.Write((short)(channels * (bitsPerSample / 8)));
            writer.Write((short)bitsPerSample);
            writer.Write("data".ToCharArray());
            writer.Write(dataSize);
            rawAudioStream.Position = 0;
            rawAudioStream.CopyTo(wavStream);
            wavStream.Position = 0;
            return new MemoryStream(wavStream.ToArray());
        }
    }
}

SpeechController.cs :

using Microsoft.AspNetCore.Mvc;
using WebApplication15.Services;

namespace WebApplication15.Controllers
{
    [ApiController]
    [Route("api/speech")]
    public class SpeechController : ControllerBase
    {
        private readonly SpeechService _speechService;
        public SpeechController(SpeechService speechService)
        {
            _speechService = speechService;
        }

        [HttpPost("convert")]
        public async Task<IActionResult> ConvertTextToSpeech([FromBody] SpeechRequest request)
        {
            if (string.IsNullOrEmpty(request.Text))
            {
                return BadRequest("Text cannot be empty.");
            }
            try
            {
                var audioStream = await _speechService.ConvertTextToSpeechAsync(request.Text);
                return File(audioStream, "audio/wav", "speech.wav");
            }
            catch (System.Exception ex)
            {
                return StatusCode(500, $"Internal server error: {ex.Message}");
            }
        }
    }
    public class SpeechRequest
    {
        public string Text { get; set; }
    }
}

Add builder.Services.AddSingleton<SpeechService>(); line in the Program.cs class.

Output :

I successfully converted the text to speech output in the browser as shwon below.