Update embedded speech samples (#2201)

Azure-Samples · Jan 2, 2024 · fdc0644 · fdc0644
1 parent e12b0cc
commit fdc0644
Show file tree

Hide file tree

Showing 10 changed files with 55 additions and 22 deletions.
diff --git a/samples/cpp/embedded-speech/samples/intent_recognition_samples.cpp b/samples/cpp/embedded-speech/samples/intent_recognition_samples.cpp
@@ -181,7 +181,7 @@ void RecognizeIntent(bool useKeyword)
         {
             // NoMatch occurs when no speech was recognized.
             auto reason = NoMatchDetails::FromResult(e.Result)->Reason;
-            cout << "NOMATCH: Reason=";
+            cout << "NO MATCH: Reason=";
             switch (reason)
             {
             case NoMatchReason::NotRecognized:

diff --git a/samples/cpp/embedded-speech/samples/speech_recognition_samples.cpp b/samples/cpp/embedded-speech/samples/speech_recognition_samples.cpp
@@ -153,7 +153,7 @@ void RecognizeSpeech(shared_ptr<SpeechRecognizer> recognizer, bool useKeyword, b
         {
             // NoMatch occurs when no speech phrase was recognized.
             auto reason = NoMatchDetails::FromResult(e.Result)->Reason;
-            cout << "NOMATCH: Reason=";
+            cout << "NO MATCH: Reason=";
             switch (reason)
             {
             case NoMatchReason::NotRecognized:

diff --git a/samples/cpp/embedded-speech/samples/speech_translation_samples.cpp b/samples/cpp/embedded-speech/samples/speech_translation_samples.cpp
@@ -71,32 +71,42 @@ void TranslateSpeech(shared_ptr<TranslationRecognizer> recognizer)
         // the source language.
         if (e.Result->Reason == ResultReason::TranslatingSpeech)
         {
+            // Source (input) language identification is enabled when TranslationRecognizer
+            // is created with an AutoDetectSourceLanguageConfig argument.
+            // In case the model does not support this functionality or the language cannot
+            // be identified, the result Language is "Unknown".
+            auto sourceLangResult = AutoDetectSourceLanguageResult::FromResult(e.Result);
+            const auto& sourceLang = sourceLangResult->Language;
+
             for (const auto& translation : e.Result->Translations)
             {
                 auto targetLang = translation.first;
                 auto outputText = translation.second;
-                cout << "Translating [" << targetLang << "]: " << outputText << endl;
+                cout << "Translating [" << sourceLang << " -> " << targetLang << "]: " << outputText << endl;
             }
         }
     };
 
     recognizer->Recognized += [](const TranslationRecognitionEventArgs& e)
     {
+        // Final result. May differ from the last intermediate result.
         if (e.Result->Reason == ResultReason::TranslatedSpeech)
         {
-            // Final result. May differ from the last intermediate result.
+            auto sourceLangResult = AutoDetectSourceLanguageResult::FromResult(e.Result);
+            const auto& sourceLang = sourceLangResult->Language;
+
             for (const auto& translation : e.Result->Translations)
             {
                 auto targetLang = translation.first;
                 auto outputText = translation.second;
-                cout << "TRANSLATED [" << targetLang << "]: " << outputText << endl;
+                cout << "TRANSLATED [" << sourceLang << " -> " << targetLang << "]: " << outputText << endl;
             }
         }
         else if (e.Result->Reason == ResultReason::NoMatch)
         {
             // NoMatch occurs when no speech phrase was recognized.
             auto reason = NoMatchDetails::FromResult(e.Result)->Reason;
-            cout << "NOMATCH: Reason=";
+            cout << "NO MATCH: Reason=";
             switch (reason)
             {
             case NoMatchReason::NotRecognized:
@@ -182,8 +192,9 @@ void TranslateSpeech(shared_ptr<TranslationRecognizer> recognizer)
 void EmbeddedSpeechTranslationFromMicrophone()
 {
     auto speechConfig = CreateEmbeddedSpeechConfig();
+    auto sourceLangConfig = AutoDetectSourceLanguageConfig::FromOpenRange(); // optional, for input language identification
     auto audioConfig = AudioConfig::FromDefaultMicrophoneInput();
 
-    auto recognizer = TranslationRecognizer::FromConfig(speechConfig, audioConfig);
+    auto recognizer = TranslationRecognizer::FromConfig(speechConfig, sourceLangConfig, audioConfig);
     TranslateSpeech(recognizer);
 }
diff --git a/samples/csharp/dotnetcore/embedded-speech/samples/IntentRecognitionSamples.cs b/samples/csharp/dotnetcore/embedded-speech/samples/IntentRecognitionSamples.cs
@@ -175,7 +175,7 @@ private static async Task RecognizeIntentAsync(bool useKeyword)
                 {
                     // NoMatch occurs when no speech was recognized.
                     var reason = NoMatchDetails.FromResult(e.Result).Reason;
-                    Console.WriteLine($"NOMATCH: Reason={reason}");
+                    Console.WriteLine($"NO MATCH: Reason={reason}");
                 }
             };
 

diff --git a/samples/csharp/dotnetcore/embedded-speech/samples/SpeechRecognitionSamples.cs b/samples/csharp/dotnetcore/embedded-speech/samples/SpeechRecognitionSamples.cs
@@ -143,7 +143,7 @@ private static async Task RecognizeSpeechAsync(SpeechRecognizer recognizer, bool
                 {
                     // NoMatch occurs when no speech was recognized.
                     var reason = NoMatchDetails.FromResult(e.Result).Reason;
-                    Console.WriteLine($"NOMATCH: Reason={reason}");
+                    Console.WriteLine($"NO MATCH: Reason={reason}");
                 }
             };
 

diff --git a/samples/csharp/dotnetcore/embedded-speech/samples/SpeechTranslationSamples.cs b/samples/csharp/dotnetcore/embedded-speech/samples/SpeechTranslationSamples.cs
@@ -71,32 +71,42 @@ private static async Task TranslateSpeechAsync(TranslationRecognizer recognizer)
                 // the source language.
                 if (e.Result.Reason == ResultReason.TranslatingSpeech)
                 {
+                    // Source (input) language identification is enabled when TranslationRecognizer
+                    // is created with an AutoDetectSourceLanguageConfig argument.
+                    // In case the model does not support this functionality or the language cannot
+                    // be identified, the result Language is "Unknown".
+                    var sourceLangResult = AutoDetectSourceLanguageResult.FromResult(e.Result);
+                    var sourceLang = sourceLangResult.Language;
+
                     foreach (var translation in e.Result.Translations)
                     {
                         var targetLang = translation.Key;
                         var outputText = translation.Value;
-                        Console.WriteLine($"Translating [{targetLang}]: {outputText}");
+                        Console.WriteLine($"Translating [{sourceLang} -> {targetLang}]: {outputText}");
                     }
                 }
             };
 
             recognizer.Recognized += (s, e) =>
             {
+                // Final result. May differ from the last intermediate result.
                 if (e.Result.Reason == ResultReason.TranslatedSpeech)
                 {
-                    // Final result. May differ from the last intermediate result.
+                    var sourceLangResult = AutoDetectSourceLanguageResult.FromResult(e.Result);
+                    var sourceLang = sourceLangResult.Language;
+
                     foreach (var translation in e.Result.Translations)
                     {
                         var targetLang = translation.Key;
                         var outputText = translation.Value;
-                        Console.WriteLine($"TRANSLATED [{targetLang}]: {outputText}");
+                        Console.WriteLine($"TRANSLATED [{sourceLang} -> {targetLang}]: {outputText}");
                     }
                 }
                 else if (e.Result.Reason == ResultReason.NoMatch)
                 {
                     // NoMatch occurs when no speech was recognized.
                     var reason = NoMatchDetails.FromResult(e.Result).Reason;
-                    Console.WriteLine($"NOMATCH: Reason={reason}");
+                    Console.WriteLine($"NO MATCH: Reason={reason}");
                 }
             };
 
@@ -146,9 +156,10 @@ private static async Task TranslateSpeechAsync(TranslationRecognizer recognizer)
         public static void EmbeddedTranslationFromMicrophone()
         {
             var speechConfig = Settings.CreateEmbeddedSpeechConfig();
+            var sourceLangConfig = AutoDetectSourceLanguageConfig.FromOpenRange(); // optional, for input language identification
             using var audioConfig = AudioConfig.FromDefaultMicrophoneInput();
 
-            using var recognizer = new TranslationRecognizer(speechConfig, audioConfig);
+            using var recognizer = new TranslationRecognizer(speechConfig, sourceLangConfig, audioConfig);
             TranslateSpeechAsync(recognizer).Wait();
         }
     }

diff --git a/samples/java/android/embedded-speech/app/build.gradle b/samples/java/android/embedded-speech/app/build.gradle
@@ -25,7 +25,7 @@ dependencies {
     implementation fileTree(include: ['*.jar'], dir: 'libs')
 
     // Speech SDK
-    implementation 'com.microsoft.cognitiveservices.speech:client-sdk-embedded:1.34.0'
+    implementation 'com.microsoft.cognitiveservices.speech:client-sdk-embedded:1.34.0@aar'
 
     implementation 'androidx.appcompat:appcompat:1.3.1'
     implementation 'androidx.constraintlayout:constraintlayout:2.1.0'

diff --git a/...src/com/microsoft/cognitiveservices/speech/samples/embedded/IntentRecognitionSamples.java b/...src/com/microsoft/cognitiveservices/speech/samples/embedded/IntentRecognitionSamples.java
@@ -174,7 +174,7 @@ else if (e.getResult().getReason() == ResultReason.NoMatch)
             {
                 // NoMatch occurs when no speech was recognized.
                 NoMatchReason reason = NoMatchDetails.fromResult(e.getResult()).getReason();
-                System.out.println("NOMATCH: Reason=" + reason);
+                System.out.println("NO MATCH: Reason=" + reason);
             }
         });
 

diff --git a/...src/com/microsoft/cognitiveservices/speech/samples/embedded/SpeechRecognitionSamples.java b/...src/com/microsoft/cognitiveservices/speech/samples/embedded/SpeechRecognitionSamples.java
@@ -160,7 +160,7 @@ else if (e.getResult().getReason() == ResultReason.NoMatch)
             {
                 // NoMatch occurs when no speech was recognized.
                 NoMatchReason reason = NoMatchDetails.fromResult(e.getResult()).getReason();
-                System.out.println("NOMATCH: Reason=" + reason);
+                System.out.println("NO MATCH: Reason=" + reason);
             }
         });
 

diff --git a/...src/com/microsoft/cognitiveservices/speech/samples/embedded/SpeechTranslationSamples.java b/...src/com/microsoft/cognitiveservices/speech/samples/embedded/SpeechTranslationSamples.java
@@ -78,32 +78,42 @@ private static void translateSpeechAsync(TranslationRecognizer recognizer) throw
             // the source language.
             if (e.getResult().getReason() == ResultReason.TranslatingSpeech)
             {
+                // Source (input) language identification is enabled when TranslationRecognizer
+                // is created with an AutoDetectSourceLanguageConfig argument.
+                // In case the model does not support this functionality or the language cannot
+                // be identified, the result Language is "Unknown".
+                AutoDetectSourceLanguageResult sourceLangResult = AutoDetectSourceLanguageResult.fromResult(e.getResult());
+                String sourceLang = sourceLangResult.getLanguage();
+
                 for (Map.Entry<String, String> translation : e.getResult().getTranslations().entrySet())
                 {
                     String targetLang = translation.getKey();
                     String outputText = translation.getValue();
-                    System.out.println("Translating [" + targetLang + "]: " + outputText);
+                    System.out.println("Translating [" + sourceLang + " -> " + targetLang + "]: " + outputText);
                 }
             }
         });
 
         recognizer.recognized.addEventListener((s, e) ->
         {
+            // Final result. May differ from the last intermediate result.
             if (e.getResult().getReason() == ResultReason.TranslatedSpeech)
             {
-                // Final result. May differ from the last intermediate result.
+                AutoDetectSourceLanguageResult sourceLangResult = AutoDetectSourceLanguageResult.fromResult(e.getResult());
+                String sourceLang = sourceLangResult.getLanguage();
+
                 for (Map.Entry<String, String> translation : e.getResult().getTranslations().entrySet())
                 {
                     String targetLang = translation.getKey();
                     String outputText = translation.getValue();
-                    System.out.println("TRANSLATED [" + targetLang + "]: " + outputText);
+                    System.out.println("TRANSLATED [" + sourceLang + " -> " + targetLang + "]: " + outputText);
                 }
             }
             else if (e.getResult().getReason() == ResultReason.NoMatch)
             {
                 // NoMatch occurs when no speech was recognized.
                 NoMatchReason reason = NoMatchDetails.fromResult(e.getResult()).getReason();
-                System.out.println("NOMATCH: Reason=" + reason);
+                System.out.println("NO MATCH: Reason=" + reason);
             }
         });
 
@@ -154,9 +164,10 @@ else if (e.getResult().getReason() == ResultReason.NoMatch)
     public static void embeddedTranslationFromMicrophone() throws InterruptedException, ExecutionException
     {
         EmbeddedSpeechConfig speechConfig = Settings.createEmbeddedSpeechConfig();
+        AutoDetectSourceLanguageConfig sourceLangConfig = AutoDetectSourceLanguageConfig.fromOpenRange(); // optional, for input language identification
         AudioConfig audioConfig = AudioConfig.fromDefaultMicrophoneInput();
 
-        TranslationRecognizer recognizer = new TranslationRecognizer(speechConfig, audioConfig);
+        TranslationRecognizer recognizer = new TranslationRecognizer(speechConfig, sourceLangConfig, audioConfig);
         translateSpeechAsync(recognizer);
 
         recognizer.close();