Add VAD examples using ALSA for recording (#739)

sherpa-onnx/csrc/CMakeLists.txt

@@ -251,6 +251,7 @@ if(SHERPA_ONNX_HAS_ALSA AND SHERPA_ONNX_ENABLE_BINARY)
   add_executable(sherpa-onnx-keyword-spotter-alsa sherpa-onnx-keyword-spotter-alsa.cc alsa.cc)
   add_executable(sherpa-onnx-alsa-offline sherpa-onnx-alsa-offline.cc alsa.cc)
   add_executable(sherpa-onnx-alsa-offline-speaker-identification sherpa-onnx-alsa-offline-speaker-identification.cc alsa.cc)
+  add_executable(sherpa-onnx-vad-alsa sherpa-onnx-vad-alsa.cc alsa.cc)
 
 
   if(SHERPA_ONNX_ENABLE_TTS)
@@ -259,9 +260,10 @@ if(SHERPA_ONNX_HAS_ALSA AND SHERPA_ONNX_ENABLE_BINARY)
 
   set(exes
     sherpa-onnx-alsa
-    sherpa-onnx-keyword-spotter-alsa
     sherpa-onnx-alsa-offline
     sherpa-onnx-alsa-offline-speaker-identification
+    sherpa-onnx-keyword-spotter-alsa
+    sherpa-onnx-vad-alsa
   )
 
   if(SHERPA_ONNX_ENABLE_TTS)

sherpa-onnx/csrc/sherpa-onnx-vad-alsa.cc

@@ -0,0 +1,132 @@
+// sherpa-onnx/csrc/sherpa-onnx-vad-alsa.cc
+//
+// Copyright (c)  2024  Xiaomi Corporation
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+#include <algorithm>
+
+#include "sherpa-onnx/csrc/alsa.h"
+#include "sherpa-onnx/csrc/circular-buffer.h"
+#include "sherpa-onnx/csrc/voice-activity-detector.h"
+#include "sherpa-onnx/csrc/wave-writer.h"
+
+bool stop = false;
+static void Handler(int32_t sig) {
+  stop = true;
+  fprintf(stderr, "\nCaught Ctrl + C. Exiting...\n");
+}
+
+int32_t main(int32_t argc, char *argv[]) {
+  signal(SIGINT, Handler);
+
+  const char *kUsageMessage = R"usage(
+This program shows how to use VAD in sherpa-onnx.
+
+  ./bin/sherpa-onnx-vad-alsa \
+    --silero-vad-model=/path/to/silero_vad.onnx \
+    device_name
+
+Please download silero_vad.onnx from
+https://github.com/snakers4/silero-vad/blob/master/files/silero_vad.onnx
+
+For instance, use
+wget https://github.com/snakers4/silero-vad/raw/master/files/silero_vad.onnx
+
+The device name specifies which microphone to use in case there are several
+on your system. You can use
+
+  arecord -l
+
+to find all available microphones on your computer. For instance, if it outputs
+
+**** List of CAPTURE Hardware Devices ****
+card 3: UACDemoV10 [UACDemoV1.0], device 0: USB Audio [USB Audio]
+  Subdevices: 1/1
+  Subdevice #0: subdevice #0
+
+and if you want to select card 3 and the device 0 on that card, please use:
+
+  plughw:3,0
+
+as the device_name.
+)usage";
+
+  sherpa_onnx::ParseOptions po(kUsageMessage);
+  sherpa_onnx::VadModelConfig config;
+
+  config.Register(&po);
+  po.Read(argc, argv);
+  if (po.NumArgs() != 1) {
+    fprintf(stderr, "Please provide only 1 argument: the device name\n");
+    po.PrintUsage();
+    exit(EXIT_FAILURE);
+  }
+
+  fprintf(stderr, "%s\n", config.ToString().c_str());
+
+  if (!config.Validate()) {
+    fprintf(stderr, "Errors in config!\n");
+    return -1;
+  }
+
+  std::string device_name = po.GetArg(1);
+  sherpa_onnx::Alsa alsa(device_name.c_str());
+  fprintf(stderr, "Use recording device: %s\n", device_name.c_str());
+
+  int32_t sample_rate = 16000;
+
+  if (alsa.GetExpectedSampleRate() != sample_rate) {
+    fprintf(stderr, "sample rate: %d != %d\n", alsa.GetExpectedSampleRate(),
+            sample_rate);
+    exit(-1);
+  }
+
+  int32_t chunk = 0.1 * alsa.GetActualSampleRate();
+
+  auto vad = std::make_unique<sherpa_onnx::VoiceActivityDetector>(config);
+
+  fprintf(stderr, "Started. Please speak\n");
+
+  int32_t window_size = config.silero_vad.window_size;
+  bool printed = false;
+
+  int32_t k = 0;
+  while (!stop) {
+    {
+      const std::vector<float> &samples = alsa.Read(chunk);
+
+      vad->AcceptWaveform(samples.data(), samples.size());
+
+      if (vad->IsSpeechDetected() && !printed) {
+        printed = true;
+        fprintf(stderr, "\nDetected speech!\n");
+      }
+      if (!vad->IsSpeechDetected()) {
+        printed = false;
+      }
+
+      while (!vad->Empty()) {
+        const auto &segment = vad->Front();
+        float duration =
+            segment.samples.size() / static_cast<float>(sample_rate);
+
+        fprintf(stderr, "Duration: %.3f seconds\n", duration);
+
+        char filename[128];
+        snprintf(filename, sizeof(filename), "seg-%d-%.3fs.wav", k, duration);
+        k += 1;
+        sherpa_onnx::WriteWave(filename, 16000, segment.samples.data(),
+                               segment.samples.size());
+        fprintf(stderr, "Saved to %s\n", filename);
+        fprintf(stderr, "----------\n");
+
+        vad->Pop();
+      }
+    }
+  }
+
+  return 0;
+}

sherpa-onnx/csrc/sherpa-onnx-vad-microphone.cc

@@ -13,6 +13,7 @@
 #include "sherpa-onnx/csrc/circular-buffer.h"
 #include "sherpa-onnx/csrc/microphone.h"
 #include "sherpa-onnx/csrc/voice-activity-detector.h"
+#include "sherpa-onnx/csrc/wave-writer.h"
 
 bool stop = false;
 std::mutex mutex;
@@ -122,6 +123,7 @@ wget https://github.com/snakers4/silero-vad/raw/master/files/silero_vad.onnx
   int32_t window_size = config.silero_vad.window_size;
   bool printed = false;
 
+  int32_t k = 0;
   while (!stop) {
     {
       std::lock_guard<std::mutex> lock(mutex);
@@ -140,9 +142,19 @@ wget https://github.com/snakers4/silero-vad/raw/master/files/silero_vad.onnx
         }
 
         while (!vad->Empty()) {
-          float duration = vad->Front().samples.size() / sample_rate;
-          vad->Pop();
+          const auto &segment = vad->Front();
+          float duration = segment.samples.size() / sample_rate;
           fprintf(stderr, "Duration: %.3f seconds\n", duration);
+
+          char filename[128];
+          snprintf(filename, sizeof(filename), "seg-%d-%.3fs.wav", k, duration);
+          k += 1;
+          sherpa_onnx::WriteWave(filename, 16000, segment.samples.data(),
+                                 segment.samples.size());
+          fprintf(stderr, "Saved to %s\n", filename);
+          fprintf(stderr, "----------\n");
+
+          vad->Pop();
         }
       }
     }

sherpa-onnx/python/csrc/CMakeLists.txt

@@ -35,6 +35,7 @@ set(srcs
   vad-model-config.cc
   vad-model.cc
   voice-activity-detector.cc
+  wave-writer.cc
 )
 if(SHERPA_ONNX_HAS_ALSA)
   list(APPEND srcs ${CMAKE_SOURCE_DIR}/sherpa-onnx/csrc/alsa.cc alsa.cc)

sherpa-onnx/python/csrc/sherpa-onnx.cc

@@ -26,6 +26,7 @@
 #include "sherpa-onnx/python/csrc/vad-model-config.h"
 #include "sherpa-onnx/python/csrc/vad-model.h"
 #include "sherpa-onnx/python/csrc/voice-activity-detector.h"
+#include "sherpa-onnx/python/csrc/wave-writer.h"
 
 #if SHERPA_ONNX_ENABLE_TTS == 1
 #include "sherpa-onnx/python/csrc/offline-tts.h"
@@ -36,6 +37,8 @@ namespace sherpa_onnx {
 PYBIND11_MODULE(_sherpa_onnx, m) {
   m.doc() = "pybind11 binding of sherpa-onnx";
 
+  PybindWaveWriter(&m);
+
   PybindFeatures(&m);
   PybindOnlineCtcFstDecoderConfig(&m);
   PybindOnlineModelConfig(&m);

sherpa-onnx/python/csrc/wave-writer.cc

@@ -0,0 +1,27 @@
+// sherpa-onnx/python/csrc/wave-writer.cc
+//
+// Copyright (c)  2024  Xiaomi Corporation
+
+#include "sherpa-onnx/python/csrc/wave-writer.h"
+
+#include <string>
+#include <vector>
+
+#include "sherpa-onnx/csrc/wave-writer.h"
+
+namespace sherpa_onnx {
+
+void PybindWaveWriter(py::module *m) {
+  m->def(
+      "write_wave",
+      [](const std::string &filename, const std::vector<float> &samples,
+         int32_t sample_rate) -> bool {
+        bool ok =
+            WriteWave(filename, sample_rate, samples.data(), samples.size());
+
+        return ok;
+      },
+      py::arg("filename"), py::arg("samples"), py::arg("sample_rate"));
+}
+
+}  // namespace sherpa_onnx

sherpa-onnx/python/csrc/wave-writer.h

@@ -0,0 +1,16 @@
+// sherpa-onnx/python/csrc/wave-writer.h
+//
+// Copyright (c)  2024  Xiaomi Corporation
+
+#ifndef SHERPA_ONNX_PYTHON_CSRC_WAVE_WRITER_H_
+#define SHERPA_ONNX_PYTHON_CSRC_WAVE_WRITER_H_
+
+#include "sherpa-onnx/python/csrc/sherpa-onnx.h"
+
+namespace sherpa_onnx {
+
+void PybindWaveWriter(py::module *m);
+
+}
+
+#endif  // SHERPA_ONNX_PYTHON_CSRC_WAVE_WRITER_H_

sherpa-onnx/python/sherpa_onnx/__init__.py

@@ -19,6 +19,7 @@ from _sherpa_onnx import (
     VadModel,
     VadModelConfig,
     VoiceActivityDetector,
+    write_wave,
 )
 
 from .keyword_spotter import KeywordSpotter