Support specifying max speech duration for VAD. (#1348)

python-api-examples/generate-subtitles.py

@@ -406,7 +406,14 @@ def main():
 
     config = sherpa_onnx.VadModelConfig()
     config.silero_vad.model = args.silero_vad_model
-    config.silero_vad.min_silence_duration = 0.25
+    config.silero_vad.threshold = 0.5
+    config.silero_vad.min_silence_duration = 0.25  # seconds
+    config.silero_vad.min_speech_duration = 0.25  # seconds
+
+    # If the current segment is larger than this value, then it increases
+    # the threshold to 0.9 internally. After detecting this segment,
+    # it resets the threshold to its original value.
+    config.silero_vad.max_speech_duration = 5  # seconds
     config.sample_rate = args.sample_rate
 
     window_size = config.silero_vad.window_size

sherpa-onnx/csrc/silero-vad-model-config.cc

@@ -28,6 +28,12 @@ void SileroVadModelConfig::Register(ParseOptions *po) {
                "In seconds.  In the end of each silence chunk wait for "
                "--silero-vad-min-speech-duration seconds before separating it");
 
+  po->Register(
+      "silero-vad-max-speech-duration", &max_speech_duration,
+      "In seconds. If a speech segment is longer than this value, then we "
+      "increase the threshold to 0.9. After finishing detecting the segment, "
+      "the threshold value is reset to its original value.");
+
   po->Register(
       "silero-vad-window-size", &window_size,
       "In samples. Audio chunks of --silero-vad-window-size samples are fed "
@@ -63,6 +69,33 @@ bool SileroVadModelConfig::Validate() const {
     return false;
   }
 
+  if (min_silence_duration <= 0) {
+    SHERPA_ONNX_LOGE(
+        "Please use a larger value for --silero-vad-min-silence-duration. "
+        "Given: "
+        "%f",
+        min_silence_duration);
+    return false;
+  }
+
+  if (min_speech_duration <= 0) {
+    SHERPA_ONNX_LOGE(
+        "Please use a larger value for --silero-vad-min-speech-duration. "
+        "Given: "
+        "%f",
+        min_speech_duration);
+    return false;
+  }
+
+  if (max_speech_duration <= 0) {
+    SHERPA_ONNX_LOGE(
+        "Please use a larger value for --silero-vad-max-speech-duration. "
+        "Given: "
+        "%f",
+        max_speech_duration);
+    return false;
+  }
+
   return true;
 }
 
@@ -74,6 +107,7 @@ std::string SileroVadModelConfig::ToString() const {
   os << "threshold=" << threshold << ", ";
   os << "min_silence_duration=" << min_silence_duration << ", ";
   os << "min_speech_duration=" << min_speech_duration << ", ";
+  os << "max_speech_duration=" << max_speech_duration << ", ";
   os << "window_size=" << window_size << ")";
 
   return os.str();

sherpa-onnx/csrc/silero-vad-model-config.h

@@ -27,6 +27,11 @@ struct SileroVadModelConfig {
   // 256, 512, 768 samples for 800 Hz
   int32_t window_size = 512;  // in samples
 
+  // If a speech segment is longer than this value, then we increase
+  // the threshold to 0.9. After finishing detecting the segment,
+  // the threshold value is reset to its original value.
+  float max_speech_duration = 20;  // in seconds
+
   SileroVadModelConfig() = default;
 
   void Register(ParseOptions *po);

sherpa-onnx/csrc/voice-activity-detector.cc

@@ -18,14 +18,18 @@ class VoiceActivityDetector::Impl {
   explicit Impl(const VadModelConfig &config, float buffer_size_in_seconds = 60)
       : model_(VadModel::Create(config)),
         config_(config),
-        buffer_(buffer_size_in_seconds * config.sample_rate) {}
+        buffer_(buffer_size_in_seconds * config.sample_rate) {
+    Init();
+  }
 
 #if __ANDROID_API__ >= 9
   Impl(AAssetManager *mgr, const VadModelConfig &config,
        float buffer_size_in_seconds = 60)
       : model_(VadModel::Create(mgr, config)),
         config_(config),
-        buffer_(buffer_size_in_seconds * config.sample_rate) {}
+        buffer_(buffer_size_in_seconds * config.sample_rate) {
+    Init();
+  }
 #endif
 
   void AcceptWaveform(const float *samples, int32_t n) {
@@ -145,6 +149,15 @@ class VoiceActivityDetector::Impl {
 
   const VadModelConfig &GetConfig() const { return config_; }
 
+ private:
+  void Init() {
+    // TODO(fangjun): Currently, we support only one vad model.
+    // If a new vad model is added, we need to change the place
+    // where max_speech_duration is placed.
+    max_utterance_length_ =
+        config_.sample_rate * config_.silero_vad.max_speech_duration;
+  }
+
  private:
   std::queue<SpeechSegment> segments_;
 
@@ -153,9 +166,9 @@ class VoiceActivityDetector::Impl {
   CircularBuffer buffer_;
   std::vector<float> last_;
 
-  int max_utterance_length_ = 16000 * 20;  // in samples
+  int max_utterance_length_ = -1;  // in samples
   float new_min_silence_duration_s_ = 0.1;
-  float new_threshold_ = 1.10;
+  float new_threshold_ = 0.90;
 
   int32_t start_ = -1;
 };

sherpa-onnx/python/csrc/silero-vad-model-config.cc

@@ -17,25 +17,29 @@ void PybindSileroVadModelConfig(py::module *m) {
       .def(py::init<>())
       .def(py::init([](const std::string &model, float threshold,
                        float min_silence_duration, float min_speech_duration,
-                       int32_t window_size) -> std::unique_ptr<PyClass> {
+                       int32_t window_size,
+                       float max_speech_duration) -> std::unique_ptr<PyClass> {
              auto ans = std::make_unique<PyClass>();
 
              ans->model = model;
              ans->threshold = threshold;
              ans->min_silence_duration = min_silence_duration;
              ans->min_speech_duration = min_speech_duration;
              ans->window_size = window_size;
+             ans->max_speech_duration = max_speech_duration;
 
              return ans;
            }),
            py::arg("model"), py::arg("threshold") = 0.5,
            py::arg("min_silence_duration") = 0.5,
-           py::arg("min_speech_duration") = 0.25, py::arg("window_size") = 512)
+           py::arg("min_speech_duration") = 0.25, py::arg("window_size") = 512,
+           py::arg("max_speech_duration") = 20)
       .def_readwrite("model", &PyClass::model)
       .def_readwrite("threshold", &PyClass::threshold)
       .def_readwrite("min_silence_duration", &PyClass::min_silence_duration)
       .def_readwrite("min_speech_duration", &PyClass::min_speech_duration)
       .def_readwrite("window_size", &PyClass::window_size)
+      .def_readwrite("max_speech_duration", &PyClass::max_speech_duration)
       .def("__str__", &PyClass::ToString)
       .def("validate", &PyClass::Validate);
 }