Add voice switcher for TTS

src/Translumo.Processing/TranslationProcessingService.cs

@@ -52,10 +52,10 @@ public class TranslationProcessingService : IProcessingService, IDisposable
         private long _lastTranslatedTextTicks;
 
         private const float MIN_SCORE_THRESHOLD = 2.1f;
-        
+
         public TranslationProcessingService(ICapturerFactory capturerFactory, IChatTextMediator chatTextMediator, OcrEnginesFactory ocrEnginesFactory,
             TranslatorFactory translationFactory, TtsFactory ttsFactory, TtsConfiguration ttsConfiguration,
-            TextDetectionProvider textProvider, TranslationConfiguration translationConfiguration, OcrGeneralConfiguration ocrConfiguration, 
+            TextDetectionProvider textProvider, TranslationConfiguration translationConfiguration, OcrGeneralConfiguration ocrConfiguration,
             TextResultCacheService textResultCacheService, TextProcessingConfiguration textConfiguration, ILogger<TranslationProcessingService> logger)
         {
             _logger = logger;
@@ -236,7 +236,7 @@ void CapturerEnsureInitialized()
                             continue;
                         }
 
-                        if (_textResultCacheService.IsCached(bestDetected.Text, bestDetected.ValidityScore, sequentialText, 
+                        if (_textResultCacheService.IsCached(bestDetected.Text, bestDetected.ValidityScore, sequentialText,
                                 bestDetected.Language.Asian, out iterationId))
                         {
                             sequentialText = false;
@@ -257,7 +257,7 @@ void CapturerEnsureInitialized()
                     }
 
                     _logger.LogError(ex, $"Screen capture failed (code: {ex.ErrorCode})");
-                    
+
                     _capturer.Dispose();
                     _capturer = null;
                     CapturerEnsureInitialized();
@@ -407,9 +407,15 @@ private void TtsConfigurationOnPropertyChanged(object sender, PropertyChangedEve
             if (e.PropertyName == nameof(_ttsConfiguration.TtsLanguage)
                 || e.PropertyName == nameof(_ttsConfiguration.TtsSystem))
             {
-                _ttsEngine.Dispose();
+                _ttsEngine?.Dispose();
+                _ttsEngine = null;
                 _ttsEngine = _ttsFactory.CreateTtsEngine(_ttsConfiguration);
             }
+            else if (e.PropertyName == nameof(_ttsConfiguration.CurrentVoice)
+                && _ttsEngine != null && _ttsConfiguration.CurrentVoice != null)
+            {
+                _ttsEngine.SetVoice(_ttsConfiguration.CurrentVoice);
+            }
         }
 
         private void OcrGeneralConfigurationOnPropertyChanged(object sender, PropertyChangedEventArgs e)

src/Translumo.TTS/Engines/ITTSEngine.cs

@@ -1,6 +1,10 @@
 namespace Translumo.TTS.Engines;
 
-public interface ITTSEngine: IDisposable
+public interface ITTSEngine : IDisposable
 {
     void SpeechText(string text);
+
+    string[] GetVoices();
+
+    void SetVoice(string voice);
 }

src/Translumo.TTS/Engines/NoneTTSEngine.cs

@@ -10,6 +10,12 @@ public void Dispose()
     {
     }
 
+    public string[] GetVoices() => new[] { "None" };
+
+    public void SetVoice(string voice)
+    {
+    }
+
     public void SpeechText(string text)
     {
     }

src/Translumo.TTS/Engines/SileroTTSEngine.cs

@@ -8,6 +8,7 @@ public class SileroTTSEngine : ITTSEngine
 {
     private dynamic _ipython;
     private dynamic _model;
+    private string[] _voices;
     private string _voice;
     private readonly string _modelPath;
     private readonly PythonEngineWrapper _pythonEngine;
@@ -51,7 +52,8 @@ private void Init()
             _pyObjects.Add(_ipython);
         });
 
-        _voice = ((string[])_model.speakers).First();
+        _voices = (string[])_model.speakers;
+        _voice = _voices.First();
     }
 
     public void SpeechText(string text)
@@ -146,5 +148,9 @@ private string GetModelFullPath(string langCode)
             _ => null
         };
 
+    public string[] GetVoices() => _voices;
+
+    public void SetVoice(string voice) => _voice = _voices.First(x => x.Equals(voice, StringComparison.OrdinalIgnoreCase));
+
     private sealed record ModelDescription(string FileUrl, string WarmUpText);
 }

src/Translumo.TTS/Engines/WindowsTTSEngine.cs

@@ -1,20 +1,25 @@
-using System.Globalization;
+using System.Collections;
+using System.Collections.ObjectModel;
+using System.Globalization;
+using System.Reflection;
 using System.Speech.Synthesis;
 
 namespace Translumo.TTS.Engines;
 
 public class WindowsTTSEngine : ITTSEngine
 {
-    private readonly VoiceInfo _voiceInfo;
+    private VoiceInfo _voiceInfo;
     private readonly SpeechSynthesizer _synthesizer;
+    private readonly ReadOnlyDictionary<string, VoiceInfo> _voices;
 
     public WindowsTTSEngine(string languageCode)
     {
         _synthesizer = new SpeechSynthesizer();
         _synthesizer.SetOutputToDefaultAudioDevice();
         _synthesizer.Rate = 1;
-
-        _voiceInfo = _synthesizer.GetInstalledVoices(new CultureInfo(languageCode)).FirstOrDefault()?.VoiceInfo;
+        SpeechApiReflectionHelper.InjectOneCoreVoices(_synthesizer);
+        _voices = _synthesizer.GetInstalledVoices(new CultureInfo(languageCode)).ToDictionary(x => x.VoiceInfo.Name, x => x.VoiceInfo).AsReadOnly();
+        _voiceInfo = _voices.First().Value;
     }
 
     public void SpeechText(string text)
@@ -36,4 +41,88 @@ public void Dispose()
     {
         _synthesizer.Dispose();
     }
+
+    public string[] GetVoices() => _voices.Keys.ToArray();
+
+    public void SetVoice(string voice) => _voiceInfo = _voices.First(x => x.Key.Equals(voice, StringComparison.OrdinalIgnoreCase)).Value;
+
+    // by default SpeechSynthesizer show not all available voices
+    // https://stackoverflow.com/a/71198211
+    private static class SpeechApiReflectionHelper
+    {
+        private const string PROP_VOICE_SYNTHESIZER = "VoiceSynthesizer";
+        private const string FIELD_INSTALLED_VOICES = "_installedVoices";
+
+        private const string ONE_CORE_VOICES_REGISTRY = @"HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech_OneCore\Voices";
+
+        private static readonly Type _objectTokenCategoryType = typeof(SpeechSynthesizer).Assembly
+            .GetType("System.Speech.Internal.ObjectTokens.ObjectTokenCategory")!;
+
+        private static readonly Type _voiceInfoType = typeof(SpeechSynthesizer).Assembly
+            .GetType("System.Speech.Synthesis.VoiceInfo")!;
+
+        private static readonly Type _installedVoiceType = typeof(SpeechSynthesizer).Assembly
+            .GetType("System.Speech.Synthesis.InstalledVoice")!;
+
+
+        public static void InjectOneCoreVoices(SpeechSynthesizer synthesizer)
+        {
+            var voiceSynthesizer = GetProperty(synthesizer, PROP_VOICE_SYNTHESIZER);
+            if (voiceSynthesizer == null)
+                throw new NotSupportedException($"Property not found: {PROP_VOICE_SYNTHESIZER}");
+
+            var installedVoices = GetField(voiceSynthesizer, FIELD_INSTALLED_VOICES) as IList;
+            if (installedVoices == null)
+                throw new NotSupportedException($"Field not found or null: {FIELD_INSTALLED_VOICES}");
+
+            if (_objectTokenCategoryType
+                    .GetMethod("Create", BindingFlags.Static | BindingFlags.NonPublic)?
+                    .Invoke(null, new object?[] { ONE_CORE_VOICES_REGISTRY }) is not IDisposable otc)
+                throw new NotSupportedException($"Failed to call Create on {_objectTokenCategoryType} instance");
+
+            using (otc)
+            {
+                if (_objectTokenCategoryType
+                        .GetMethod("FindMatchingTokens", BindingFlags.Instance | BindingFlags.NonPublic)?
+                        .Invoke(otc, new object?[] { null, null }) is not IList tokens)
+                    throw new NotSupportedException($"Failed to list matching tokens");
+
+                foreach (var token in tokens)
+                {
+                    if (token == null || GetProperty(token, "Attributes") == null)
+                        continue;
+
+                    var voiceInfo =
+                        typeof(SpeechSynthesizer).Assembly
+                            .CreateInstance(_voiceInfoType.FullName!, true,
+                                BindingFlags.Instance | BindingFlags.NonPublic, null,
+                                new object[] { token }, null, null);
+
+                    if (voiceInfo == null)
+                        throw new NotSupportedException($"Failed to instantiate {_voiceInfoType}");
+
+                    var installedVoice =
+                        typeof(SpeechSynthesizer).Assembly
+                            .CreateInstance(_installedVoiceType.FullName!, true,
+                                BindingFlags.Instance | BindingFlags.NonPublic, null,
+                                new object[] { voiceSynthesizer, voiceInfo }, null, null);
+
+                    if (installedVoice == null)
+                        throw new NotSupportedException($"Failed to instantiate {_installedVoiceType}");
+
+                    installedVoices.Add(installedVoice);
+                }
+            }
+        }
+
+        private static object? GetProperty(object target, string propName)
+        {
+            return target.GetType().GetProperty(propName, BindingFlags.Instance | BindingFlags.NonPublic)?.GetValue(target);
+        }
+
+        private static object? GetField(object target, string propName)
+        {
+            return target.GetType().GetField(propName, BindingFlags.Instance | BindingFlags.NonPublic)?.GetValue(target);
+        }
+    }
 }

src/Translumo.TTS/IObserverAvailableVoices.cs

@@ -0,0 +1,6 @@
+namespace Translumo.TTS;
+
+public interface IObserverAvailableVoices
+{
+    void UpdateVoice(IList<string> currentVoices);
+}

src/Translumo.TTS/TtsConfiguration.cs

@@ -1,6 +1,6 @@
-using Translumo.Infrastructure.Language;
+using System.Collections.ObjectModel;
+using Translumo.Infrastructure.Language;
 using Translumo.Utils;
-using Windows.Security.EnterpriseData;
 
 namespace Translumo.TTS;
 
@@ -12,12 +12,14 @@ public class TtsConfiguration : BindableBase
         {
             TtsLanguage = Languages.English,
             TtsSystem = TTSEngines.None,
-            InstalledWinTtsLanguages = new List<Languages>()
+            InstalledWinTtsLanguages = new List<Languages>(),
+            _currentVoice = string.Empty,
         };
 
     private TTSEngines _ttsSystem;
     private Languages _ttsLanguage;
     private List<Languages> _installedWinTtsLanguages;
+    private string _currentVoice;
 
     public TTSEngines TtsSystem
     {
@@ -37,6 +39,12 @@ public Languages TtsLanguage
         }
     }
 
+    public string CurrentVoice
+    {
+        get => _currentVoice;
+        set => SetProperty(ref _currentVoice, value);
+    }
+
     public List<Languages> InstalledWinTtsLanguages
     {
         get => _installedWinTtsLanguages;

src/Translumo.TTS/TtsFactory.cs

@@ -9,24 +9,33 @@ public class TtsFactory
     {
         private readonly LanguageService _languageService;
         private readonly PythonEngineWrapper _pythonEngine;
+        private readonly IObserverAvailableVoices _observerAvailableVoices;
         private readonly ILogger _logger;
 
-        public TtsFactory(LanguageService languageService, PythonEngineWrapper pythonEngine, ILogger<TtsFactory> logger)
+        public TtsFactory(LanguageService languageService, PythonEngineWrapper pythonEngine, IObserverAvailableVoices observerAvailableVoices, ILogger<TtsFactory> logger)
         {
             _languageService = languageService;
             _pythonEngine = pythonEngine;
+            _observerAvailableVoices = observerAvailableVoices;
             _logger = logger;
         }
 
-        public ITTSEngine CreateTtsEngine(TtsConfiguration ttsConfiguration) =>
-            ttsConfiguration.TtsSystem switch
+        public ITTSEngine CreateTtsEngine(TtsConfiguration ttsConfiguration)
+        {
+            ITTSEngine ttsEngine = ttsConfiguration.TtsSystem switch
             {
                 TTSEngines.None => new NoneTTSEngine(),
                 TTSEngines.WindowsTTS => new WindowsTTSEngine(GetLangCode(ttsConfiguration)),
                 TTSEngines.SileroTTS => new SileroTTSEngine(_pythonEngine, GetLangCode(ttsConfiguration)),
                 _ => throw new NotSupportedException()
             };
 
+            var voices = ttsEngine.GetVoices();
+            _observerAvailableVoices.UpdateVoice(voices);
+
+            return ttsEngine;
+        }
+
         private string GetLangCode(TtsConfiguration ttsConfiguration) =>
             _languageService.GetLanguageDescriptor(ttsConfiguration.TtsLanguage).Code;
     }

src/Translumo/App.xaml.cs

@@ -97,6 +97,7 @@ private void ConfigureServices(ServiceCollection services)
             services.AddScoped<HotkeysSettingsViewModel>();
             services.AddScoped<LanguagesSettingsViewModel>();
             services.AddScoped<OcrSettingsViewModel>();
+            services.AddScoped<IObserverAvailableVoices, LanguagesSettingsViewModel>();
 
             var chatWindowConfiguration = ChatWindowConfiguration.Default;
             services.AddSingleton<OcrGeneralConfiguration>(OcrGeneralConfiguration.Default);