Add C API for MatchaTTS models (#1675)

@@ -81,6 +81,51 @@ jobs:
             otool -L ./install/lib/libsherpa-onnx-c-api.dylib
           fi
 
+      - name: Test Matcha TTS (zh)
+        shell: bash
+        run: |
+          gcc -o matcha-tts-zh-c-api ./c-api-examples/matcha-tts-zh-c-api.c \
+            -I ./build/install/include \
+            -L ./build/install/lib/ \
+            -l sherpa-onnx-c-api \
+            -l onnxruntime
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/matcha-icefall-zh-baker.tar.bz2
+          tar xvf matcha-icefall-zh-baker.tar.bz2
+          rm matcha-icefall-zh-baker.tar.bz2
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/vocoder-models/hifigan_v2.onnx
+
+          export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
+          export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
+
+          ./matcha-tts-zh-c-api
+
+      - name: Test Matcha TTS (en)
+        shell: bash
+        run: |
+          gcc -o matcha-tts-en-c-api ./c-api-examples/matcha-tts-en-c-api.c \
+            -I ./build/install/include \
+            -L ./build/install/lib/ \
+            -l sherpa-onnx-c-api \
+            -l onnxruntime
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/matcha-icefall-en_US-ljspeech.tar.bz2
+          tar xvf matcha-icefall-en_US-ljspeech.tar.bz2
+          rm matcha-icefall-en_US-ljspeech.tar.bz2
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/vocoder-models/hifigan_v2.onnx
+
+          export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
+          export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
+
+          ./matcha-tts-en-c-api
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: matcha-tts-${{ matrix.os }}
+          path: ./generated-matcha-*.wav
+
       - name: Test vad + Whisper tiny.en
         shell: bash
         run: |

@@ -7,6 +7,12 @@ target_link_libraries(decode-file-c-api sherpa-onnx-c-api cargs)
 if(SHERPA_ONNX_ENABLE_TTS)
   add_executable(offline-tts-c-api offline-tts-c-api.c)
   target_link_libraries(offline-tts-c-api sherpa-onnx-c-api cargs)
+
+  add_executable(matcha-tts-zh-c-api matcha-tts-zh-c-api.c)
+  target_link_libraries(matcha-tts-zh-c-api sherpa-onnx-c-api)
+
+  add_executable(matcha-tts-en-c-api matcha-tts-en-c-api.c)
+  target_link_libraries(matcha-tts-en-c-api sherpa-onnx-c-api)
 endif()
 
 if(SHERPA_ONNX_ENABLE_SPEAKER_DIARIZATION)

+// c-api-examples/matcha-tts-en-c-api.c
+//
+// Copyright (c)  2025  Xiaomi Corporation
+
+// This file shows how to use sherpa-onnx C API
+// for English TTS with MatchaTTS.
+//
+// clang-format off
+/*
+Usage
+
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/matcha-icefall-en_US-ljspeech.tar.bz2
+tar xvf matcha-icefall-en_US-ljspeech.tar.bz2
+rm matcha-icefall-en_US-ljspeech.tar.bz2
+
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/vocoder-models/hifigan_v2.onnx
+
+./matcha-tts-en-c-api
+
+ */
+// clang-format on
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sherpa-onnx/c-api/c-api.h"
+
+static int32_t ProgressCallback(const float *samples, int32_t num_samples,
+                                float progress) {
+  fprintf(stderr, "Progress: %.3f%%\n", progress * 100);
+  // return 1 to continue generating
+  // return 0 to stop generating
+  return 1;
+}
+
+int32_t main(int32_t argc, char *argv[]) {
+  SherpaOnnxOfflineTtsConfig config;
+  memset(&config, 0, sizeof(config));
+  config.model.matcha.acoustic_model =
+      "./matcha-icefall-en_US-ljspeech/model-steps-3.onnx";
+
+  config.model.matcha.vocoder = "./hifigan_v2.onnx";
+
+  config.model.matcha.tokens = "./matcha-icefall-en_US-ljspeech/tokens.txt";
+
+  config.model.matcha.data_dir =
+      "./matcha-icefall-en_US-ljspeech/espeak-ng-data";
+
+  config.model.num_threads = 1;
+
+  // If you don't want to see debug messages, please set it to 0
+  config.model.debug = 1;
+
+  const char *filename = "./generated-matcha-en.wav";
+  const char *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. "
+      "Friends fell out often because life was changing so fast. The easiest "
+      "thing in the world was to lose touch with someone.";
+
+  SherpaOnnxOfflineTts *tts = SherpaOnnxCreateOfflineTts(&config);
+  int32_t sid = 0;
+  float speed = 1.0;  // larger -> faster in speech 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);
+#else
+  const SherpaOnnxGeneratedAudio *audio =
+      SherpaOnnxOfflineTtsGenerateWithProgressCallback(tts, text, sid, speed,
+                                                       ProgressCallback);
+#endif
+
+  SherpaOnnxWriteWave(audio->samples, audio->n, audio->sample_rate, filename);
+
+  SherpaOnnxDestroyOfflineTtsGeneratedAudio(audio);
+  SherpaOnnxDestroyOfflineTts(tts);
+
+  fprintf(stderr, "Input text is: %s\n", text);
+  fprintf(stderr, "Speaker ID is is: %d\n", sid);
+  fprintf(stderr, "Saved to: %s\n", filename);
+
+  return 0;
+}

+// c-api-examples/matcha-tts-zh-c-api.c
+//
+// Copyright (c)  2025  Xiaomi Corporation
+
+// This file shows how to use sherpa-onnx C API
+// for Chinese TTS with MatchaTTS.
+//
+// clang-format off
+/*
+Usage
+
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/matcha-icefall-zh-baker.tar.bz2
+tar xvf matcha-icefall-zh-baker.tar.bz2
+rm matcha-icefall-zh-baker.tar.bz2
+
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/vocoder-models/hifigan_v2.onnx
+
+./matcha-tts-zh-c-api
+
+ */
+// clang-format on
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sherpa-onnx/c-api/c-api.h"
+
+static int32_t ProgressCallback(const float *samples, int32_t num_samples,
+                                float progress) {
+  fprintf(stderr, "Progress: %.3f%%\n", progress * 100);
+  // return 1 to continue generating
+  // return 0 to stop generating
+  return 1;
+}
+
+int32_t main(int32_t argc, char *argv[]) {
+  SherpaOnnxOfflineTtsConfig config;
+  memset(&config, 0, sizeof(config));
+  config.model.matcha.acoustic_model =
+      "./matcha-icefall-zh-baker/model-steps-3.onnx";
+  config.model.matcha.vocoder = "./hifigan_v2.onnx";
+  config.model.matcha.lexicon = "./matcha-icefall-zh-baker/lexicon.txt";
+  config.model.matcha.tokens = "./matcha-icefall-zh-baker/tokens.txt";
+  config.model.matcha.dict_dir = "./matcha-icefall-zh-baker/dict";
+  config.model.num_threads = 1;
+
+  // If you don't want to see debug messages, please set it to 0
+  config.model.debug = 1;
+
+  // clang-format off
+  config.rule_fsts = "./matcha-icefall-zh-baker/phone.fst,./matcha-icefall-zh-baker/date.fst,./matcha-icefall-zh-baker/number.fst";
+  // clang-format on
+
+  const char *filename = "./generated-matcha-zh.wav";
+  const char *text =
+      "当夜幕降临，星光点点，伴随着微风拂面，我在静谧中感受着时光的流转，思念如"
+      "涟漪荡漾，梦境如画卷展开，我与自然融为一体，沉静在这片宁静的美丽之中，感"
+      "受着生命的奇迹与温柔."
+      "某某银行的副行长和一些行政领导表示，他们去过长江和长白山; "
+      "经济不断增长。2024年12月31号，拨打110或者18920240511。123456块钱。";
+
+  SherpaOnnxOfflineTts *tts = SherpaOnnxCreateOfflineTts(&config);
+  int32_t sid = 0;
+  float speed = 1.0;  // larger -> faster in speech 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);
+#else
+  const SherpaOnnxGeneratedAudio *audio =
+      SherpaOnnxOfflineTtsGenerateWithProgressCallback(tts, text, sid, speed,
+                                                       ProgressCallback);
+#endif
+
+  SherpaOnnxWriteWave(audio->samples, audio->n, audio->sample_rate, filename);
+
+  SherpaOnnxDestroyOfflineTtsGeneratedAudio(audio);
+  SherpaOnnxDestroyOfflineTts(tts);
+
+  fprintf(stderr, "Input text is: %s\n", text);
+  fprintf(stderr, "Speaker ID is is: %d\n", sid);
+  fprintf(stderr, "Saved to: %s\n", filename);
+
+  return 0;
+}

@@ -1058,6 +1058,7 @@ static sherpa_onnx::OfflineTtsConfig GetOfflineTtsConfig(
     const SherpaOnnxOfflineTtsConfig *config) {
   sherpa_onnx::OfflineTtsConfig tts_config;
 
+  // vits
   tts_config.model.vits.model = SHERPA_ONNX_OR(config->model.vits.model, "");
   tts_config.model.vits.lexicon =
       SHERPA_ONNX_OR(config->model.vits.lexicon, "");
@@ -1073,6 +1074,24 @@ static sherpa_onnx::OfflineTtsConfig GetOfflineTtsConfig(
   tts_config.model.vits.dict_dir =
       SHERPA_ONNX_OR(config->model.vits.dict_dir, "");
 
+  // matcha
+  tts_config.model.matcha.acoustic_model =
+      SHERPA_ONNX_OR(config->model.matcha.acoustic_model, "");
+  tts_config.model.matcha.vocoder =
+      SHERPA_ONNX_OR(config->model.matcha.vocoder, "");
+  tts_config.model.matcha.lexicon =
+      SHERPA_ONNX_OR(config->model.matcha.lexicon, "");
+  tts_config.model.matcha.tokens =
+      SHERPA_ONNX_OR(config->model.matcha.tokens, "");
+  tts_config.model.matcha.data_dir =
+      SHERPA_ONNX_OR(config->model.matcha.data_dir, "");
+  tts_config.model.matcha.noise_scale =
+      SHERPA_ONNX_OR(config->model.matcha.noise_scale, 0.667);
+  tts_config.model.matcha.length_scale =
+      SHERPA_ONNX_OR(config->model.matcha.length_scale, 1.0);
+  tts_config.model.matcha.dict_dir =
+      SHERPA_ONNX_OR(config->model.matcha.dict_dir, "");
+
   tts_config.model.num_threads = SHERPA_ONNX_OR(config->model.num_threads, 1);
   tts_config.model.debug = config->model.debug;
   tts_config.model.provider = SHERPA_ONNX_OR(config->model.provider, "cpu");
@@ -1082,7 +1101,7 @@ static sherpa_onnx::OfflineTtsConfig GetOfflineTtsConfig(
 
   tts_config.rule_fsts = SHERPA_ONNX_OR(config->rule_fsts, "");
   tts_config.rule_fars = SHERPA_ONNX_OR(config->rule_fars, "");
-  tts_config.max_num_sentences = SHERPA_ONNX_OR(config->max_num_sentences, 2);
+  tts_config.max_num_sentences = SHERPA_ONNX_OR(config->max_num_sentences, 1);
 
   if (tts_config.model.debug) {
 #if __OHOS__

@@ -894,15 +894,28 @@ SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsVitsModelConfig {
 
   float noise_scale;
   float noise_scale_w;
-  float length_scale;  // < 1, faster in speed; > 1, slower in speed
+  float length_scale;  // < 1, faster in speech speed; > 1, slower in speed
   const char *dict_dir;
 } SherpaOnnxOfflineTtsVitsModelConfig;
 
+SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsMatchaModelConfig {
+  const char *acoustic_model;
+  const char *vocoder;
+  const char *lexicon;
+  const char *tokens;
+  const char *data_dir;
+
+  float noise_scale;
+  float length_scale;  // < 1, faster in speech speed; > 1, slower in speed
+  const char *dict_dir;
+} SherpaOnnxOfflineTtsMatchaModelConfig;
+
 SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsModelConfig {
   SherpaOnnxOfflineTtsVitsModelConfig vits;
   int32_t num_threads;
   int32_t debug;
   const char *provider;
+  SherpaOnnxOfflineTtsMatchaModelConfig matcha;
 } SherpaOnnxOfflineTtsModelConfig;
 
 SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsConfig {

@@ -30,7 +30,7 @@ struct OfflineTtsConfig {
   // Maximum number of sentences that we process at a time.
   // This is to avoid OOM for very long input text.
   // If you set it to -1, then we process all sentences in a single batch.
-  int32_t max_num_sentences = 2;
+  int32_t max_num_sentences = 1;
 
   OfflineTtsConfig() = default;
   OfflineTtsConfig(const OfflineTtsModelConfig &model,