configure GPIO and sample rates

main/Application.cc

@@ -7,9 +7,6 @@
 #include "silk_resampler.h"
 
 #define TAG "application"
-#define INPUT_SAMPLE_RATE 16000
-#define DECODE_SAMPLE_RATE 24000
-#define OUTPUT_SAMPLE_RATE 24000
 
 
 Application::Application() {
@@ -27,10 +24,10 @@ Application::Application() {
         }
     }
     
-    opus_encoder_.Configure(INPUT_SAMPLE_RATE, 1);
+    opus_encoder_.Configure(CONFIG_AUDIO_INPUT_SAMPLE_RATE, 1);
-    opus_decoder_ = opus_decoder_create(DECODE_SAMPLE_RATE, 1, NULL);
+    opus_decoder_ = opus_decoder_create(opus_decode_sample_rate_, 1, NULL);
-    if (DECODE_SAMPLE_RATE != OUTPUT_SAMPLE_RATE) {
+    if (opus_decode_sample_rate_ != CONFIG_AUDIO_OUTPUT_SAMPLE_RATE) {
-        assert(0 == silk_resampler_init(&resampler_state_, DECODE_SAMPLE_RATE, OUTPUT_SAMPLE_RATE, 1));
+        assert(0 == silk_resampler_init(&resampler_state_, opus_decode_sample_rate_, CONFIG_AUDIO_OUTPUT_SAMPLE_RATE, 1));
     }
 }
 
@@ -59,7 +56,7 @@ Application::~Application() {
 }
 
 void Application::Start() {
-    audio_device_.Start(INPUT_SAMPLE_RATE, OUTPUT_SAMPLE_RATE);
+    audio_device_.Start(CONFIG_AUDIO_INPUT_SAMPLE_RATE, CONFIG_AUDIO_OUTPUT_SAMPLE_RATE);
     audio_device_.OnStateChanged([this]() {
         if (audio_device_.playing()) {
             SetChatState(kChatStateSpeaking);
@@ -154,7 +151,7 @@ void Application::StartCommunication() {
             .total_ch_num = 1,
             .mic_num = 1,
             .ref_num = 0,
-            .sample_rate = INPUT_SAMPLE_RATE
+            .sample_rate = CONFIG_AUDIO_INPUT_SAMPLE_RATE,
         },
         .debug_init = false,
         .debug_hook = {{ AFE_DEBUG_HOOK_MASE_TASK_IN, NULL }, { AFE_DEBUG_HOOK_FETCH_TASK_IN, NULL }},
@@ -195,7 +192,7 @@ void Application::StartDetection() {
             .total_ch_num = 1,
             .mic_num = 1,
             .ref_num = 0,
-            .sample_rate = INPUT_SAMPLE_RATE
+            .sample_rate = CONFIG_AUDIO_INPUT_SAMPLE_RATE
         },
         .debug_init = false,
         .debug_hook = {{ AFE_DEBUG_HOOK_MASE_TASK_IN, NULL }, { AFE_DEBUG_HOOK_FETCH_TASK_IN, NULL }},
@@ -335,11 +332,11 @@ void Application::AudioEncodeTask() {
 }
 
 void Application::AudioDecodeTask() {
-    int frame_size = DECODE_SAMPLE_RATE / 1000 * opus_duration_ms_;
-
     while (true) {
         AudioPacket* packet;
         xQueueReceive(audio_decode_queue_, &packet, portMAX_DELAY);
+
+        int frame_size = opus_decode_sample_rate_ / 1000 * opus_duration_ms_;
         packet->pcm.resize(frame_size);
 
         int ret = opus_decode(opus_decoder_, packet->opus.data(), packet->opus.size(), packet->pcm.data(), frame_size, 0);
@@ -349,8 +346,8 @@ void Application::AudioDecodeTask() {
             continue;
         }
 
-        if (DECODE_SAMPLE_RATE != OUTPUT_SAMPLE_RATE) {
+        if (opus_decode_sample_rate_ != CONFIG_AUDIO_OUTPUT_SAMPLE_RATE) {
-            int target_size = frame_size * OUTPUT_SAMPLE_RATE / DECODE_SAMPLE_RATE;
+            int target_size = frame_size * CONFIG_AUDIO_OUTPUT_SAMPLE_RATE / opus_decode_sample_rate_;
             std::vector<int16_t> resampled(target_size);
             assert(0 == silk_resampler(&resampler_state_, resampled.data(), packet->pcm.data(), frame_size));
             packet->pcm = std::move(resampled);
@@ -360,6 +357,19 @@ void Application::AudioDecodeTask() {
     }
 }
 
+void Application::SetDecodeSampleRate(int sample_rate) {
+    if (opus_decode_sample_rate_ == sample_rate) {
+        return;
+    }
+
+    opus_decoder_destroy(opus_decoder_);
+    opus_decode_sample_rate_ = sample_rate;
+    opus_decoder_ = opus_decoder_create(opus_decode_sample_rate_, 1, NULL);
+    if (opus_decode_sample_rate_ != CONFIG_AUDIO_OUTPUT_SAMPLE_RATE) {
+        assert(0 == silk_resampler_init(&resampler_state_, opus_decode_sample_rate_, CONFIG_AUDIO_OUTPUT_SAMPLE_RATE, 1));
+    }
+}
+
 void Application::StartWebSocketClient() {
     if (ws_client_ != nullptr) {
         delete ws_client_;
@@ -379,7 +389,7 @@ void Application::StartWebSocketClient() {
         message += "\"type\":\"hello\", \"version\":\"1.0\",";
         message += "\"wakeup_model\":\"" + std::string(wakenet_model_) + "\",";
         message += "\"audio_params\":{";
-        message += "\"format\":\"opus\", \"sample_rate\":" + std::to_string(INPUT_SAMPLE_RATE) + ", \"channels\":1";
+        message += "\"format\":\"opus\", \"sample_rate\":" + std::to_string(CONFIG_AUDIO_INPUT_SAMPLE_RATE) + ", \"channels\":1";
         message += "}}";
         ws_client_->Send(message);
     });
@@ -403,6 +413,10 @@ void Application::StartWebSocketClient() {
                     auto state = cJSON_GetObjectItem(root, "state");
                     if (strcmp(state->valuestring, "start") == 0) {
                         packet->type = kAudioPacketTypeStart;
+                        auto sample_rate = cJSON_GetObjectItem(root, "sample_rate");
+                        if (sample_rate != NULL) {
+                            SetDecodeSampleRate(sample_rate->valueint);
+                        }
                     } else if (strcmp(state->valuestring, "stop") == 0) {
                         packet->type = kAudioPacketTypeStop;
                     } else if (strcmp(state->valuestring, "sentence_end") == 0) {

main/Application.h

@@ -62,8 +62,10 @@ private:
     OpusDecoder* opus_decoder_ = nullptr;
 
     int opus_duration_ms_ = 60;
+    int opus_decode_sample_rate_ = CONFIG_AUDIO_OUTPUT_SAMPLE_RATE;
     silk_resampler_state_struct resampler_state_;
 
+    void SetDecodeSampleRate(int sample_rate);
     void SetChatState(ChatState state);
     void StartDetection();
     void StartCommunication();

main/AudioDevice.cc

@@ -23,15 +23,14 @@ AudioDevice::~AudioDevice() {
 }
 
 void AudioDevice::Start(int input_sample_rate, int output_sample_rate) {
-    assert(input_sample_rate == 16000);
     input_sample_rate_ = input_sample_rate;
     output_sample_rate_ = output_sample_rate;
 
-    if (output_sample_rate == 16000) {
+#ifdef CONFIG_AUDIO_DEVICE_I2S_SIMPLEX
-        CreateDuplexChannels();
-    } else {
         CreateSimplexChannels();
-    }
+#else
+        CreateDuplexChannels();
+#endif
 
     ESP_ERROR_CHECK(i2s_channel_enable(tx_handle_));
     ESP_ERROR_CHECK(i2s_channel_enable(rx_handle_));
@@ -77,10 +76,10 @@ void AudioDevice::CreateDuplexChannels() {
         },
         .gpio_cfg = {
             .mclk = I2S_GPIO_UNUSED,
-            .bclk = GPIO_NUM_5,
+            .bclk = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_BCLK,
-            .ws = GPIO_NUM_4,
+            .ws = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_WS,
-            .dout = GPIO_NUM_6,
+            .dout = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_DOUT,
-            .din = GPIO_NUM_3,
+            .din = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_DIN,
             .invert_flags = {
                 .mclk_inv = false,
                 .bclk_inv = false,
@@ -93,6 +92,7 @@ void AudioDevice::CreateDuplexChannels() {
     ESP_LOGI(TAG, "Duplex channels created");
 }
 
+#ifdef CONFIG_AUDIO_DEVICE_I2S_SIMPLEX
 void AudioDevice::CreateSimplexChannels() {
     // Create a new channel for speaker
     i2s_chan_config_t chan_cfg = {
@@ -127,9 +127,9 @@ void AudioDevice::CreateSimplexChannels() {
         },
         .gpio_cfg = {
             .mclk = I2S_GPIO_UNUSED,
-            .bclk = GPIO_NUM_5,
+            .bclk = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_BCLK,
-            .ws = GPIO_NUM_4,
+            .ws = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_WS,
-            .dout = GPIO_NUM_6,
+            .dout = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_DOUT,
             .din = I2S_GPIO_UNUSED,
             .invert_flags = {
                 .mclk_inv = false,
@@ -144,13 +144,14 @@ void AudioDevice::CreateSimplexChannels() {
     chan_cfg.id = I2S_NUM_1;
     ESP_ERROR_CHECK(i2s_new_channel(&chan_cfg, nullptr, &rx_handle_));
     std_cfg.clk_cfg.sample_rate_hz = (uint32_t)input_sample_rate_;
-    std_cfg.gpio_cfg.bclk = GPIO_NUM_11;
+    std_cfg.gpio_cfg.bclk = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_MIC_GPIO_BCLK;
-    std_cfg.gpio_cfg.ws = GPIO_NUM_10;
+    std_cfg.gpio_cfg.ws = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_MIC_GPIO_WS;
     std_cfg.gpio_cfg.dout = I2S_GPIO_UNUSED;
-    std_cfg.gpio_cfg.din = GPIO_NUM_3;
+    std_cfg.gpio_cfg.din = (gpio_num_t)CONFIG_AUDIO_DEVICE_I2S_GPIO_DIN;
     ESP_ERROR_CHECK(i2s_channel_init_std_mode(rx_handle_, &std_cfg));
     ESP_LOGI(TAG, "Simplex channels created");
 }
+#endif
 
 void AudioDevice::Write(const int16_t* data, int samples) {
     int32_t buffer[samples];

main/Kconfig.projbuild

@@ -19,4 +19,60 @@ config BUILTIN_LED_GPIO
     help
         GPIO number of the builtin LED.
 
+config AUDIO_INPUT_SAMPLE_RATE
+    int "Audio Input Sample Rate"
+    default 16000
+    help
+        Audio input sample rate.
+
+config AUDIO_OUTPUT_SAMPLE_RATE
+    int "Audio Output Sample Rate"
+    default 24000
+    help
+        Audio output sample rate.
+
+config AUDIO_DEVICE_I2S_GPIO_BCLK
+    int "I2S GPIO BCLK"
+    default 5
+    help
+        GPIO number of the I2S BCLK.
+
+config AUDIO_DEVICE_I2S_GPIO_WS
+    int "I2S GPIO WS"
+    default 4
+    help
+        GPIO number of the I2S WS.
+
+config AUDIO_DEVICE_I2S_GPIO_DOUT
+    int "I2S GPIO DOUT"
+    default 6
+    help
+        GPIO number of the I2S DOUT.
+
+config AUDIO_DEVICE_I2S_GPIO_DIN
+    int "I2S GPIO DIN"
+    default 3
+    help
+        GPIO number of the I2S DIN.
+    
+config AUDIO_DEVICE_I2S_SIMPLEX
+    bool "I2S Simplex"
+    default n
+    help
+        Enable I2S Simplex mode.
+    
+config AUDIO_DEVICE_I2S_MIC_GPIO_BCLK
+    int "I2S MIC GPIO BCLK"
+    default 11
+    depends on AUDIO_DEVICE_I2S_SIMPLEX
+    help
+        GPIO number of the I2S MIC BCLK.
+    
+config AUDIO_DEVICE_I2S_MIC_GPIO_WS
+    int "I2S MIC GPIO WS"
+    default 10
+    depends on AUDIO_DEVICE_I2S_SIMPLEX
+    help
+        GPIO number of the I2S MIC WS.
+
 endmenu