Refactor Kokoro TTS to use the new Generate API

c-api-examples/kokoro-tts-en-c-api.c

@@ -26,7 +26,7 @@ rm kokoro-en-v0_19.tar.bz2
 #include "sherpa-onnx/c-api/c-api.h"
 
 static int32_t ProgressCallback(const float *samples, int32_t num_samples,
-                                float progress) {
+                                float progress, void *arg) {
   fprintf(stderr, "Progress: %.3f%%\n", progress * 100);
   // return 1 to continue generating
   // return 0 to stop generating
@@ -60,15 +60,19 @@ int32_t main(int32_t argc, char *argv[]) {
   // 6->am_michael, 7->bf_emma, 8->bf_isabella, 9->bm_george, 10->bm_lewis
   int32_t sid = 0;
   float speed = 1.0;  // larger -> faster in speech speed
+  SherpaOnnxGenerationConfig cfg = {0};
+  cfg.silence_scale = 0.2f;
+  cfg.sid = sid;
+  cfg.speed = speed;
 
 #if 0
   // If you don't want to use a callback, then please enable this branch
   const SherpaOnnxGeneratedAudio *audio =
-      SherpaOnnxOfflineTtsGenerate(tts, text, sid, speed);
+      SherpaOnnxOfflineTtsGenerateWithConfig(tts, text, &cfg, NULL, NULL);
 #else
   const SherpaOnnxGeneratedAudio *audio =
-      SherpaOnnxOfflineTtsGenerateWithProgressCallback(tts, text, sid, speed,
-                                                       ProgressCallback);
+      SherpaOnnxOfflineTtsGenerateWithConfig(tts, text, &cfg, ProgressCallback,
+                                             NULL);
 #endif
 
   SherpaOnnxWriteWave(audio->samples, audio->n, audio->sample_rate, filename);

c-api-examples/kokoro-tts-zh-en-c-api.c

@@ -26,7 +26,7 @@ rm kokoro-multi-lang-v1_0.tar.bz2
 #include "sherpa-onnx/c-api/c-api.h"
 
 static int32_t ProgressCallback(const float *samples, int32_t num_samples,
-                                float progress) {
+                                float progress, void *arg) {
   fprintf(stderr, "Progress: %.3f%%\n", progress * 100);
   // return 1 to continue generating
   // return 0 to stop generating
@@ -58,15 +58,19 @@ int32_t main(int32_t argc, char *argv[]) {
   const SherpaOnnxOfflineTts *tts = SherpaOnnxCreateOfflineTts(&config);
   int32_t sid = 0;    // there are 53 speakers
   float speed = 1.0;  // larger -> faster in speech speed
+  SherpaOnnxGenerationConfig cfg = {0};
+  cfg.silence_scale = 0.2f;
+  cfg.sid = sid;
+  cfg.speed = speed;
 
 #if 0
   // If you don't want to use a callback, then please enable this branch
   const SherpaOnnxGeneratedAudio *audio =
-      SherpaOnnxOfflineTtsGenerate(tts, text, sid, speed);
+      SherpaOnnxOfflineTtsGenerateWithConfig(tts, text, &cfg, NULL, NULL);
 #else
   const SherpaOnnxGeneratedAudio *audio =
-      SherpaOnnxOfflineTtsGenerateWithProgressCallback(tts, text, sid, speed,
-                                                       ProgressCallback);
+      SherpaOnnxOfflineTtsGenerateWithConfig(tts, text, &cfg, ProgressCallback,
+                                             NULL);
 #endif
 
   SherpaOnnxWriteWave(audio->samples, audio->n, audio->sample_rate, filename);

cxx-api-examples/kokoro-tts-en-cxx-api.cc

@@ -57,12 +57,16 @@ int32_t main(int32_t argc, char *argv[]) {
   auto tts = OfflineTts::Create(config);
   int32_t sid = 0;
   float speed = 1.0;  // larger -> faster in speech speed
+  GenerationConfig gen_config;
+  gen_config.sid = sid;
+  gen_config.speed = speed;
+  gen_config.silence_scale = 0.2f;
 
 #if 0
   // If you don't want to use a callback, then please enable this branch
-  GeneratedAudio audio = tts.Generate(text, sid, speed);
+  GeneratedAudio audio = tts.Generate(text, gen_config);
 #else
-  GeneratedAudio audio = tts.Generate(text, sid, speed, ProgressCallback);
+  GeneratedAudio audio = tts.Generate(text, gen_config, ProgressCallback);
 #endif
 
   WriteWave(filename, {audio.samples, audio.sample_rate});

cxx-api-examples/kokoro-tts-zh-en-cxx-api.cc

@@ -58,12 +58,16 @@ int32_t main(int32_t argc, char *argv[]) {
   auto tts = OfflineTts::Create(config);
   int32_t sid = 50;
   float speed = 1.0;  // larger -> faster in speech speed
+  GenerationConfig gen_config;
+  gen_config.sid = sid;
+  gen_config.speed = speed;
+  gen_config.silence_scale = 0.2f;
 
 #if 0
   // If you don't want to use a callback, then please enable this branch
-  GeneratedAudio audio = tts.Generate(text, sid, speed);
+  GeneratedAudio audio = tts.Generate(text, gen_config);
 #else
-  GeneratedAudio audio = tts.Generate(text, sid, speed, ProgressCallback);
+  GeneratedAudio audio = tts.Generate(text, gen_config, ProgressCallback);
 #endif
 
   WriteWave(filename, {audio.samples, audio.sample_rate});

dart-api-examples/tts/bin/kokoro-en.dart

@@ -58,7 +58,6 @@ void main(List<String> arguments) async {
     voices: voices,
     tokens: tokens,
     dataDir: dataDir,
-    lengthScale: 1 / speed,
   );
 
   final modelConfig = sherpa_onnx.OfflineTtsModelConfig(
@@ -74,7 +73,12 @@ void main(List<String> arguments) async {
   );
 
   final tts = sherpa_onnx.OfflineTts(config);
-  final audio = tts.generate(text: text, sid: sid, speed: speed);
+  final genConfig = sherpa_onnx.OfflineTtsGenerationConfig(
+    sid: sid,
+    speed: speed,
+    silenceScale: config.silenceScale,
+  );
+  final audio = tts.generateWithConfig(text: text, config: genConfig);
   tts.free();
 
   sherpa_onnx.writeWave(

dart-api-examples/tts/bin/kokoro-zh-en.dart

@@ -65,7 +65,6 @@ void main(List<String> arguments) async {
     voices: voices,
     tokens: tokens,
     dataDir: dataDir,
-    lengthScale: 1 / speed,
     lexicon: lexicon,
   );
 
@@ -82,7 +81,12 @@ void main(List<String> arguments) async {
   );
 
   final tts = sherpa_onnx.OfflineTts(config);
-  final audio = tts.generate(text: text, sid: sid, speed: speed);
+  final genConfig = sherpa_onnx.OfflineTtsGenerationConfig(
+    sid: sid,
+    speed: speed,
+    silenceScale: config.silenceScale,
+  );
+  final audio = tts.generateWithConfig(text: text, config: genConfig);
   tts.free();
 
   sherpa_onnx.writeWave(

dotnet-examples/kokoro-tts-play/Program.cs

@@ -35,9 +35,13 @@ static void Main(string[] args)
       "thing in the world was to lose touch with someone.";
 
     // mapping of sid to voice name
-    // 0->af, 1->af_bella, 2->af_nicole, 3->af_sarah, 4->af_sky, 5->am_adam
-    // 6->am_michael, 7->bf_emma, 8->bf_isabella, 9->bm_george, 10->bm_lewis
-    var sid = 0;
+    // 0->af, 1->af_bella, 2->af_nicole, 3->af_sarah, 4->af_sky, 5->am_adam
+    // 6->am_michael, 7->bf_emma, 8->bf_isabella, 9->bm_george, 10->bm_lewis
+    var sid = 0;
+    OfflineTtsGenerationConfig genConfig = new OfflineTtsGenerationConfig();
+    genConfig.Sid = sid;
+    genConfig.Speed = speed;
+    genConfig.SilenceScale = 0.2f;
 
 
     Console.WriteLine(PortAudio.VersionInfo.versionText);
@@ -73,7 +77,7 @@ static void Main(string[] args)
     // https://learn.microsoft.com/en-us/dotnet/standard/collections/thread-safe/blockingcollection-overview
     var dataItems = new BlockingCollection<float[]>();
 
-    var MyCallback = (IntPtr samples, int n, float progress) =>
+    var MyCallback = (IntPtr samples, int n, float progress, IntPtr arg) =>
     {
       Console.WriteLine($"Progress {progress*100}%");
 
@@ -165,9 +169,9 @@ IntPtr userData
 
     stream.Start();
 
-    var callback = new OfflineTtsCallbackProgress(MyCallback);
-
-    var audio = tts.GenerateWithCallbackProgress(text, speed, sid, callback);
+    var callback = new OfflineTtsCallbackProgressWithArg(MyCallback);
+
+    var audio = tts.GenerateWithConfig(text, genConfig, callback);
     var outputFilename = "./generated-kokoro-0.wav";
     var ok = audio.SaveToWaveFile(outputFilename);
 

dotnet-examples/kokoro-tts/Program.cs

@@ -38,7 +38,12 @@ static void TestZhEn()
 
     var sid = 50;
 
-    var MyCallback = (IntPtr samples, int n, float progress) =>
+    OfflineTtsGenerationConfig genConfig = new OfflineTtsGenerationConfig();
+    genConfig.Sid = sid;
+    genConfig.Speed = speed;
+    genConfig.SilenceScale = 0.2f;
+
+    var MyCallback = (IntPtr samples, int n, float progress, IntPtr arg) =>
     {
       float[] data = new float[n];
       Marshal.Copy(samples, data, 0, n);
@@ -51,9 +56,9 @@ static void TestZhEn()
       return 1;
     };
 
-    var callback = new OfflineTtsCallbackProgress(MyCallback);
-
-    var audio = tts.GenerateWithCallbackProgress(text, speed, sid, callback);
+    var callback = new OfflineTtsCallbackProgressWithArg(MyCallback);
+
+    var audio = tts.GenerateWithConfig(text, genConfig, callback);
 
     var outputFilename = "./generated-kokoro-zh-en.wav";
     var ok = audio.SaveToWaveFile(outputFilename);
@@ -93,7 +98,12 @@ static void TestEn()
     // 6->am_michael, 7->bf_emma, 8->bf_isabella, 9->bm_george, 10->bm_lewis
     var sid = 0;
 
-    var MyCallback = (IntPtr samples, int n, float progress) =>
+    OfflineTtsGenerationConfig genConfig = new OfflineTtsGenerationConfig();
+    genConfig.Sid = sid;
+    genConfig.Speed = speed;
+    genConfig.SilenceScale = 0.2f;
+
+    var MyCallback = (IntPtr samples, int n, float progress, IntPtr arg) =>
     {
       float[] data = new float[n];
       Marshal.Copy(samples, data, 0, n);
@@ -106,9 +116,9 @@ static void TestEn()
       return 1;
     };
 
-    var callback = new OfflineTtsCallbackProgress(MyCallback);
-
-    var audio = tts.GenerateWithCallbackProgress(text, speed, sid, callback);
+    var callback = new OfflineTtsCallbackProgressWithArg(MyCallback);
+
+    var audio = tts.GenerateWithConfig(text, genConfig, callback);
 
     var outputFilename = "./generated-kokoro-en.wav";
     var ok = audio.SaveToWaveFile(outputFilename);

java-api-examples/NonStreamingTtsKokoroEn.java

@@ -38,8 +38,12 @@ public static void main(String[] args) {
 
     int sid = 0;
     float speed = 1.0f;
+    GenerationConfig genConfig = new GenerationConfig();
+    genConfig.setSid(sid);
+    genConfig.setSpeed(speed);
+    genConfig.setSilenceScale(0.2f);
     long start = System.currentTimeMillis();
-    GeneratedAudio audio = tts.generate(text, sid, speed);
+    GeneratedAudio audio = tts.generateWithConfigAndCallback(text, genConfig, samples -> {});
     long stop = System.currentTimeMillis();
 
     float timeElapsedSeconds = (stop - start) / 1000.0f;

java-api-examples/NonStreamingTtsKokoroZhEn.java

@@ -40,8 +40,12 @@ public static void main(String[] args) {
 
     int sid = 0; // this model has 53 speakers. You can use sid in the range 0-52
     float speed = 1.0f;
+    GenerationConfig genConfig = new GenerationConfig();
+    genConfig.setSid(sid);
+    genConfig.setSpeed(speed);
+    genConfig.setSilenceScale(0.2f);
     long start = System.currentTimeMillis();
-    GeneratedAudio audio = tts.generate(text, sid, speed);
+    GeneratedAudio audio = tts.generateWithConfigAndCallback(text, genConfig, samples -> {});
     long stop = System.currentTimeMillis();
 
     float timeElapsedSeconds = (stop - start) / 1000.0f;

nodejs-addon-examples/test_tts_non_streaming_kokoro_en.js

@@ -27,9 +27,15 @@ const tts = createOfflineTts();
 const text =
     'Today as always, men fall into two groups: slaves and free men. Whoever does not have two-thirds of his day for himself, is a slave, whatever he may be: a statesman, a businessman, an official, or a scholar.';
 
+const generationConfig = new sherpa_onnx.GenerationConfig({
+  sid: 6,
+  speed: 1.0,
+  silenceScale: 0.2,
+});
+
 
 let start = Date.now();
-const audio = tts.generate({text: text, sid: 6, speed: 1.0});
+const audio = tts.generate({text, generationConfig});
 let stop = Date.now();
 const elapsed_seconds = (stop - start) / 1000;
 const duration = audio.samples.length / audio.sampleRate;

nodejs-addon-examples/test_tts_non_streaming_kokoro_zh_en.js

@@ -29,8 +29,14 @@ const tts = createOfflineTts();
 const text =
     '中英文语音合成测试。This is generated by next generation Kaldi using Kokoro without Misaki. 你觉得中英文说的如何呢？';
 
+const generationConfig = new sherpa_onnx.GenerationConfig({
+  sid: 48,
+  speed: 1.0,
+  silenceScale: 0.2,
+});
+
 let start = Date.now();
-const audio = tts.generate({text: text, sid: 48, speed: 1.0});
+const audio = tts.generate({text, generationConfig});
 let stop = Date.now();
 const elapsed_seconds = (stop - start) / 1000;
 const duration = audio.samples.length / audio.sampleRate;

nodejs-examples/test-offline-tts-kokoro-en.js

@@ -28,10 +28,15 @@ function createOfflineTts() {
 const tts = createOfflineTts();
 const speakerId = 0;
 const speed = 1.0;
+const generationConfig = {
+  sid: speakerId,
+  speed: speed,
+  silenceScale: 0.2,
+};
 const text =
     'Today as always, men fall into two groups: slaves and free men. Whoever does not have two-thirds of his day for himself, is a slave, whatever he may be: a statesman, a businessman, an official, or a scholar.';
 
-const audio = tts.generate({text: text, sid: speakerId, speed: speed});
+const audio = tts.generateWithConfig(text, generationConfig);
 tts.save('./test-kokoro-en.wav', audio);
 console.log('Saved to test-kokoro-en.wav successfully.');
 tts.free();

nodejs-examples/test-offline-tts-kokoro-zh-en.js

@@ -30,10 +30,15 @@ function createOfflineTts() {
 const tts = createOfflineTts();
 const speakerId = 49;
 const speed = 1.0;
+const generationConfig = {
+  sid: speakerId,
+  speed: speed,
+  silenceScale: 0.2,
+};
 const text =
     '中英文语音合成测试。This is generated by next generation Kaldi using Kokoro without Misaki. 你觉得中英文说的如何呢？';
 
-const audio = tts.generate({text: text, sid: speakerId, speed: speed});
+const audio = tts.generateWithConfig(text, generationConfig);
 tts.save('./test-kokoro-zh-en-49.wav', audio);
 console.log('Saved to test-kokoro-zh-en-49.wav successfully.');
 tts.free();

pascal-api-examples/tts/kokoro-en-playback.pas

@@ -50,10 +50,11 @@
   Param: TPaStreamParameters;
   Stream: PPaStream;
   Wave: TSherpaOnnxWave;
+  GenerationConfig: TSherpaOnnxGenerationConfig;
 
 function GenerateCallback(
       Samples: pcfloat; N: cint32;
-      Arg: Pointer): cint; cdecl;
+      Progress: cfloat; Arg: Pointer): cint; cdecl;
 begin
   EnterCriticalSection(CriticalSection);
   try
@@ -207,8 +208,13 @@ function GetOfflineTts: TSherpaOnnxOfflineTts;
 
   Text := 'Friends fell out often because life was changing so fast. The easiest thing in the world was to lose touch with someone.';
 
-  Audio :=  Tts.Generate(Text, SpeakerId, Speed,
-    PSherpaOnnxGeneratedAudioCallbackWithArg(@GenerateCallback), nil);
+  GenerationConfig := Default(TSherpaOnnxGenerationConfig);
+  GenerationConfig.SilenceScale := 0.2;
+  GenerationConfig.Speed := Speed;
+  GenerationConfig.Sid := SpeakerId;
+
+  Audio :=  Tts.Generate(Text, GenerationConfig,
+    @GenerateCallback, nil);
   FinishedGeneration := True;
   SherpaOnnxWriteWave('./kokoro-en-playback-7.wav', Audio.Samples, Audio.SampleRate);
   WriteLn('Saved to ./kokoro-en-playback-7.wav');
@@ -236,4 +242,3 @@ function GetOfflineTts: TSherpaOnnxOfflineTts;
       Exit;
     end;
 end.
-

pascal-api-examples/tts/kokoro-en.pas

@@ -34,6 +34,7 @@ function GetOfflineTts: TSherpaOnnxOfflineTts;
 var
   Tts: TSherpaOnnxOfflineTts;
   Audio: TSherpaOnnxGeneratedAudio;
+  GenerationConfig: TSherpaOnnxGenerationConfig;
 
   Text: AnsiString;
   Speed: Single = 1.0;  {Use a larger value to speak faster}
@@ -46,10 +47,14 @@ function GetOfflineTts: TSherpaOnnxOfflineTts;
 
   Text := 'Friends fell out often because life was changing so fast. The easiest thing in the world was to lose touch with someone.';
 
-  Audio :=  Tts.Generate(Text, SpeakerId, Speed);
+  GenerationConfig := Default(TSherpaOnnxGenerationConfig);
+  GenerationConfig.SilenceScale := 0.2;
+  GenerationConfig.Speed := Speed;
+  GenerationConfig.Sid := SpeakerId;
+
+  Audio :=  Tts.Generate(Text, GenerationConfig, NIL, NIL);
   SherpaOnnxWriteWave('./kokoro-en-8.wav', Audio.Samples, Audio.SampleRate);
   WriteLn('Saved to ./kokoro-en-8.wav');
 
   FreeAndNil(Tts);
 end.
-