Real-time speech recognition from microphone for .Net

CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.13 FATAL_ERROR)
 project(sherpa-onnx)
 
-set(SHERPA_ONNX_VERSION "1.4.6")
+set(SHERPA_ONNX_VERSION "1.4.7")
 
 # Disable warning about
 #

dotnet-examples/sherpa-onnx.sln

@@ -7,6 +7,8 @@ Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "online-decode-files", "onli
 EndProject
 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "offline-decode-files", "offline-decode-files\offline-decode-files.csproj", "{2DAB152C-9E24-47A0-9DB0-781297ECE458}"
 EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "speech-recognition-from-microphone", "speech-recognition-from-microphone\speech-recognition-from-microphone.csproj", "{FE4EA1FF-062A-46B3-B78D-C828FED7B82E}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -24,5 +26,9 @@ Global
 		{2DAB152C-9E24-47A0-9DB0-781297ECE458}.Debug|Any CPU.Build.0 = Debug|Any CPU
 		{2DAB152C-9E24-47A0-9DB0-781297ECE458}.Release|Any CPU.ActiveCfg = Release|Any CPU
 		{2DAB152C-9E24-47A0-9DB0-781297ECE458}.Release|Any CPU.Build.0 = Release|Any CPU
+		{FE4EA1FF-062A-46B3-B78D-C828FED7B82E}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{FE4EA1FF-062A-46B3-B78D-C828FED7B82E}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{FE4EA1FF-062A-46B3-B78D-C828FED7B82E}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{FE4EA1FF-062A-46B3-B78D-C828FED7B82E}.Release|Any CPU.Build.0 = Release|Any CPU
 	EndGlobalSection
 EndGlobal

dotnet-examples/speech-recognition-from-microphone/Program.cs

@@ -0,0 +1,234 @@
+// Copyright (c)  2023  Xiaomi Corporation
+//
+// This file shows how to use a streaming model for real-time speech
+// recognition from a microphone.
+// Please refer to
+// https://k2-fsa.github.io/sherpa/onnx/pretrained_models/online-transducer/zipformer-transducer-models.html
+// to download streaming models
+
+using CommandLine.Text;
+using CommandLine;
+using PortAudioSharp;
+using System.Threading;
+using SherpaOnnx;
+using System.Collections.Generic;
+using System.Runtime.InteropServices;
+using System;
+
+
+class OnlineDecodeFiles
+{
+  class Options
+  {
+    [Option(Required = true, HelpText = "Path to tokens.txt")]
+    public string Tokens { get; set; }
+
+    [Option(Required = false, Default = "cpu", HelpText = "Provider, e.g., cpu, coreml")]
+    public string Provider { get; set; }
+
+    [Option(Required = true, HelpText = "Path to encoder.onnx")]
+    public string Encoder { get; set; }
+
+    [Option(Required = true, HelpText = "Path to decoder.onnx")]
+    public string Decoder { get; set; }
+
+    [Option(Required = true, HelpText = "Path to joiner.onnx")]
+    public string Joiner { get; set; }
+
+    [Option("num-threads", Required = false, Default = 1, HelpText = "Number of threads for computation")]
+    public int NumThreads { get; set; }
+
+    [Option("decoding-method", Required = false, Default = "greedy_search",
+            HelpText = "Valid decoding methods are: greedy_search, modified_beam_search")]
+    public string DecodingMethod { get; set; }
+
+    [Option(Required = false, Default = false, HelpText = "True to show model info during loading")]
+    public bool Debug { get; set; }
+
+    [Option("sample-rate", Required = false, Default = 16000, HelpText = "Sample rate of the data used to train the model")]
+    public int SampleRate { get; set; }
+
+    [Option("max-active-paths", Required = false, Default = 4,
+        HelpText = @"Used only when --decoding--method is modified_beam_search.
+It specifies number of active paths to keep during the search")]
+    public int MaxActivePaths { get; set; }
+
+    [Option("enable-endpoint", Required = false, Default = true,
+        HelpText = "True to enable endpoint detection.")]
+    public bool EnableEndpoint { get; set; }
+
+    [Option("rule1-min-trailing-silence", Required = false, Default = 2.4F,
+        HelpText = @"An endpoint is detected if trailing silence in seconds is
+larger than this value even if nothing has been decoded. Used only when --enable-endpoint is true.")]
+    public float Rule1MinTrailingSilence { get; set; }
+
+    [Option("rule2-min-trailing-silence", Required = false, Default = 0.8F,
+        HelpText = @"An endpoint is detected if trailing silence in seconds is
+larger than this value after something that is not blank has been decoded. Used
+only when --enable-endpoint is true.")]
+    public float Rule2MinTrailingSilence { get; set; }
+
+    [Option("rule3-min-utterance-length", Required = false, Default = 20.0F,
+        HelpText = @"An endpoint is detected if the utterance in seconds is
+larger than this value. Used only when --enable-endpoint is true.")]
+    public float Rule3MinUtteranceLength { get; set; }
+  }
+
+  static void Main(string[] args)
+  {
+    var parser = new CommandLine.Parser(with => with.HelpWriter = null);
+    var parserResult = parser.ParseArguments<Options>(args);
+
+    parserResult
+      .WithParsed<Options>(options => Run(options))
+      .WithNotParsed(errs => DisplayHelp(parserResult, errs));
+  }
+
+  private static void DisplayHelp<T>(ParserResult<T> result, IEnumerable<Error> errs)
+  {
+    string usage = @"
+dotnet run -c Release \
+  --tokens ./icefall-asr-zipformer-streaming-wenetspeech-20230615/data/lang_char/tokens.txt \
+  --encoder ./icefall-asr-zipformer-streaming-wenetspeech-20230615/exp/encoder-epoch-12-avg-4-chunk-16-left-128.onnx \
+  --decoder ./icefall-asr-zipformer-streaming-wenetspeech-20230615/exp/decoder-epoch-12-avg-4-chunk-16-left-128.onnx \
+  --joiner ./icefall-asr-zipformer-streaming-wenetspeech-20230615/exp/joiner-epoch-12-avg-4-chunk-16-left-128.onnx \
+
+Please refer to
+https://k2-fsa.github.io/sherpa/onnx/pretrained_models/online-transducer/index.html
+to download pre-trained streaming models.
+";
+
+    var helpText = HelpText.AutoBuild(result, h =>
+    {
+      h.AdditionalNewLineAfterOption = false;
+      h.Heading = usage;
+      h.Copyright = "Copyright (c) 2023 Xiaomi Corporation";
+      return HelpText.DefaultParsingErrorsHandler(result, h);
+    }, e => e);
+    Console.WriteLine(helpText);
+  }
+
+  private static void Run(Options options)
+  {
+    OnlineRecognizerConfig config = new OnlineRecognizerConfig();
+    config.FeatConfig.SampleRate = options.SampleRate;
+
+    // All models from icefall using feature dim 80.
+    // You can change it if your model has a different feature dim.
+    config.FeatConfig.FeatureDim = 80;
+
+    config.TransducerModelConfig.Encoder = options.Encoder;
+    config.TransducerModelConfig.Decoder = options.Decoder;
+    config.TransducerModelConfig.Joiner = options.Joiner;
+    config.TransducerModelConfig.Tokens = options.Tokens;
+    config.TransducerModelConfig.Provider = options.Provider;
+    config.TransducerModelConfig.NumThreads = options.NumThreads;
+    config.TransducerModelConfig.Debug = options.Debug ? 1 : 0;
+
+    config.DecodingMethod = options.DecodingMethod;
+    config.MaxActivePaths = options.MaxActivePaths;
+    config.EnableEndpoint = options.EnableEndpoint ? 1 : 0;
+
+    config.Rule1MinTrailingSilence = options.Rule1MinTrailingSilence;
+    config.Rule2MinTrailingSilence = options.Rule2MinTrailingSilence;
+    config.Rule3MinUtteranceLength = options.Rule3MinUtteranceLength;
+
+    OnlineRecognizer recognizer = new OnlineRecognizer(config);
+
+
+    OnlineStream s = recognizer.CreateStream();
+
+    Console.WriteLine(PortAudio.VersionInfo.versionText);
+    PortAudio.Initialize();
+
+    Console.WriteLine($"Number of devices: {PortAudio.DeviceCount}");
+    for (int i = 0; i != PortAudio.DeviceCount; ++i)
+    {
+      Console.WriteLine($" Device {i}");
+      DeviceInfo deviceInfo = PortAudio.GetDeviceInfo(i);
+      Console.WriteLine($"   Name: {deviceInfo.name}");
+      Console.WriteLine($"   Max input channels: {deviceInfo.maxInputChannels}");
+      Console.WriteLine($"   Default sample rate: {deviceInfo.defaultSampleRate}");
+    }
+    int deviceIndex = PortAudio.DefaultInputDevice;
+    if (deviceIndex == PortAudio.NoDevice)
+    {
+      Console.WriteLine("No default input device found");
+      Environment.Exit(1);
+    }
+
+    DeviceInfo info = PortAudio.GetDeviceInfo(deviceIndex);
+
+    Console.WriteLine();
+    Console.WriteLine($"Use default device {deviceIndex} ({info.name})");
+
+    StreamParameters param = new StreamParameters();
+    param.device = deviceIndex;
+    param.channelCount = 1;
+    param.sampleFormat = SampleFormat.Float32;
+    param.suggestedLatency = info.defaultLowInputLatency;
+    param.hostApiSpecificStreamInfo = IntPtr.Zero;
+
+    PortAudioSharp.Stream.Callback callback = (IntPtr input, IntPtr output,
+        UInt32 frameCount,
+        ref StreamCallbackTimeInfo timeInfo,
+        StreamCallbackFlags statusFlags,
+        IntPtr userData
+        ) =>
+    {
+      float[] samples = new float[frameCount];
+      Marshal.Copy(input, samples, 0, (Int32)frameCount);
+
+      s.AcceptWaveform(options.SampleRate, samples);
+
+      return StreamCallbackResult.Continue;
+    };
+
+    PortAudioSharp.Stream stream = new PortAudioSharp.Stream(inParams: param, outParams: null, sampleRate: options.SampleRate,
+        framesPerBuffer: 0,
+        streamFlags: StreamFlags.ClipOff,
+        callback: callback,
+        userData: IntPtr.Zero
+        );
+
+    Console.WriteLine(param);
+
+    stream.Start();
+
+    int segment_index = 0;
+    String lastText = "";
+    int segmentIndex = 0;
+
+    while (true)
+    {
+      while (recognizer.IsReady(s))
+      {
+        recognizer.Decode(s);
+      }
+
+      var text = recognizer.GetResult(s).Text;
+      bool isEndpoint = recognizer.IsEndpoint(s);
+      if (!string.IsNullOrWhiteSpace(text) && lastText != text)
+      {
+        lastText = text;
+        Console.Write($"\r{segmentIndex}: {lastText}");
+      }
+
+      if (isEndpoint)
+      {
+        if (!string.IsNullOrWhiteSpace(text))
+        {
+          ++segmentIndex;
+          Console.WriteLine();
+        }
+        recognizer.Reset(s);
+      }
+
+      Thread.Sleep(200); // ms
+    }
+
+    PortAudio.Terminate();
+
+
+  }
+}

dotnet-examples/speech-recognition-from-microphone/run.sh

@@ -0,0 +1,21 @@
+#!/usr/bin/env bash
+
+# Please refer to
+# https://k2-fsa.github.io/sherpa/onnx/pretrained_models/online-transducer/zipformer-transducer-models.html#csukuangfj-sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20-bilingual-chinese-english
+# to download the model files
+#
+export LD_LIBRARY_PATH=$PWD:$LD_LIBRARY_PATH
+export DYLD_LIBRARY_PATH=$PWD:$DYLD_LIBRARY_PATH
+
+if [ ! -d ./icefall-asr-zipformer-streaming-wenetspeech-20230615 ]; then
+  GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/pkufool/icefall-asr-zipformer-streaming-wenetspeech-20230615
+  cd icefall-asr-zipformer-streaming-wenetspeech-20230615
+  git lfs pull --include "*.onnx"
+  cd ..
+fi
+
+dotnet run -c Release \
+  --tokens ./icefall-asr-zipformer-streaming-wenetspeech-20230615/data/lang_char/tokens.txt \
+  --encoder ./icefall-asr-zipformer-streaming-wenetspeech-20230615/exp/encoder-epoch-12-avg-4-chunk-16-left-128.onnx \
+  --decoder ./icefall-asr-zipformer-streaming-wenetspeech-20230615/exp/decoder-epoch-12-avg-4-chunk-16-left-128.onnx \
+  --joiner ./icefall-asr-zipformer-streaming-wenetspeech-20230615/exp/joiner-epoch-12-avg-4-chunk-16-left-128.onnx

dotnet-examples/speech-recognition-from-microphone/speech-recognition-from-microphone.csproj

@@ -0,0 +1,17 @@
+<Project Sdk="Microsoft.NET.Sdk">
+
+  <PropertyGroup>
+    <OutputType>Exe</OutputType>
+    <TargetFramework>net6.0</TargetFramework>
+    <RootNamespace>speech_recognition_from_microphone</RootNamespace>
+    <ImplicitUsings>enable</ImplicitUsings>
+    <Nullable>enable</Nullable>
+  </PropertyGroup>
+
+  <ItemGroup>
+    <PackageReference Include="CommandLineParser" Version="2.9.1" />
+    <PackageReference Include="org.k2fsa.sherpa.onnx" Version="*" />
+    <PackageReference Include="PortAudioSharp2" Version="*" />
+  </ItemGroup>
+
+</Project>