Say "Computer" -> chime + listening -> speak -> AFE VAD detects you stopped -> auto-sends. No taps. Touch still works as a manual override. - esp-sr 2.4.6 added; wn9_computer_tts model packed into a new "model" flash partition (MODEL_IN_FLASH). App moved to 8M, model 4M. - audio.c: replaced the on-demand capture_task with an AFE pipeline — feed_task is the SOLE mic reader (-> afe->feed); detect_task fetches, watches wakeup_state for the wake word and vad_state for end of speech, and forwards AFE-cleaned audio upstream during an utterance. One mic reader ever. - short rising chime acknowledges the wake audibly. Fixes from adversarial review before trusting it: 1. utterance framing (blocking WS sends) moved OFF the AFE fetch thread onto an app_task event queue (EV_TOUCH/EV_WAKE/EV_SPEECH_END) — a 1.5s send could stall fetch and drop the first ~1.5s of speech. 2. app_task is now the single serializer of start/end -> no TOCTOU double-start (was: two utterance_start on a tap during wake). 3. VAD accounting resets on every streaming (re)start (wake OR tap), not just wake -> a tapped utterance can no longer end instantly on stale silence. 4. chime/reply set s_playing (+DMA tail hold) and detect_task skips the mic while s_playing -> our own audio no longer streams into STT or false-triggers the wake at a playback boundary (no AEC yet). 5. NULL-checked AFE create + feed buffer; tasks only start if AFE is up. Verified on hardware: model loads, AFE inits with the Computer word, boots and connects clean, no crash/wedge. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
165 lines
5.0 KiB
C
165 lines
5.0 KiB
C
/* DeskLock — living-room face and voice for the Tatlock butler.
|
|
*
|
|
* Boot: display + face -> audio (gong) -> Wi-Fi (C6/ESP-Hosted) -> gateway WS.
|
|
* Interaction (phase 2): touch-to-talk — tap to speak, tap again to finish.
|
|
*
|
|
* Touch is DECOUPLED from the LVGL thread: the touch callback only signals a
|
|
* semaphore; all the real work (blocking WebSocket sends, audio, face changes)
|
|
* runs in app_task. A blocking WS send on the LVGL render thread stalls the
|
|
* MIPI-DSI flush -> blue/garbage frame + task-watchdog reboot.
|
|
*/
|
|
|
|
#include "esp_log.h"
|
|
#include "esp_system.h"
|
|
#include "freertos/FreeRTOS.h"
|
|
#include "freertos/queue.h"
|
|
#include "freertos/semphr.h"
|
|
#include "freertos/task.h"
|
|
#include "bsp/esp-bsp.h"
|
|
|
|
#include "desklock.h"
|
|
|
|
/* 1 = L1 wifi driver diagnosis (SoftAP, no STA/gateway). 0 = normal app. */
|
|
#define WIFI_DIAG_MODE 0
|
|
|
|
static const char *TAG = "desklock";
|
|
|
|
/* All utterance framing (blocking WS sends, chime, face) runs on app_task, never
|
|
* on the LVGL touch callback nor the AFE detect/fetch thread. The wake word, VAD,
|
|
* and touch just POST an event; app_task is the single serializer of start/end,
|
|
* which removes both the fetch-thread stall and the double-start race. */
|
|
typedef enum { EV_TOUCH, EV_WAKE, EV_SPEECH_END } app_event_t;
|
|
static QueueHandle_t s_events;
|
|
|
|
static void post_event(app_event_t ev)
|
|
{
|
|
if (s_events != NULL) {
|
|
xQueueSend(s_events, &ev, 0);
|
|
}
|
|
}
|
|
|
|
/* --- these three are the fast, non-blocking hooks called from other tasks --- */
|
|
void app_on_touch(void) { post_event(EV_TOUCH); } /* LVGL touch callback */
|
|
void app_on_wake(void) { post_event(EV_WAKE); } /* AFE detect_task: wake word */
|
|
void app_on_speech_end(void) { post_event(EV_SPEECH_END); } /* AFE detect_task: VAD silence */
|
|
|
|
/* --- the actual work, all on app_task --- */
|
|
static void start_utterance(void)
|
|
{
|
|
if (!gw_connected() || audio_is_playing() || audio_capture_active()
|
|
|| face_get() == FACE_EFFORT) {
|
|
return;
|
|
}
|
|
face_activity();
|
|
audio_play_chime(); /* audible "I heard you" */
|
|
gw_send_text("{\"type\":\"utterance_start\"}"); /* must precede any audio frames */
|
|
audio_capture_start(); /* now detect_task forwards audio */
|
|
face_set(FACE_LISTENING);
|
|
ESP_LOGI(TAG, "listening…");
|
|
}
|
|
|
|
static void end_utterance(void)
|
|
{
|
|
if (!audio_capture_active()) {
|
|
return;
|
|
}
|
|
audio_capture_stop();
|
|
gw_send_text("{\"type\":\"utterance_end\"}");
|
|
face_set(FACE_PENSIVE);
|
|
ESP_LOGI(TAG, "utterance sent");
|
|
}
|
|
|
|
/* WS link dropped: abandon any half-utterance so the next wake starts fresh. */
|
|
void app_on_disconnect(void)
|
|
{
|
|
audio_capture_stop();
|
|
}
|
|
|
|
static void app_task(void *arg)
|
|
{
|
|
(void)arg;
|
|
app_event_t ev;
|
|
for (;;) {
|
|
if (xQueueReceive(s_events, &ev, portMAX_DELAY) != pdTRUE) {
|
|
continue;
|
|
}
|
|
if (ev == EV_TOUCH) {
|
|
vTaskDelay(pdMS_TO_TICKS(60)); /* debounce */
|
|
xQueueReset(s_events); /* coalesce a bouncy tap burst */
|
|
face_activity();
|
|
if (!gw_connected()) {
|
|
audio_capture_stop();
|
|
} else if (!audio_capture_active()) {
|
|
start_utterance(); /* tap = manual wake */
|
|
} else {
|
|
end_utterance(); /* tap = manual end */
|
|
}
|
|
} else if (ev == EV_WAKE) {
|
|
start_utterance();
|
|
} else if (ev == EV_SPEECH_END) {
|
|
end_utterance();
|
|
}
|
|
}
|
|
}
|
|
|
|
void app_on_playback_done(void)
|
|
{
|
|
face_set(FACE_IDLE);
|
|
}
|
|
|
|
#if SDIO_TX_SELFTEST
|
|
/* Temporary autonomous SDIO-TX stress test: stream mic audio upstream for 40s
|
|
* right after connecting (no user tap needed) to prove the #167 alignment fix.
|
|
* Remove once verified. */
|
|
static void sdio_tx_selftest_task(void *arg)
|
|
{
|
|
(void)arg;
|
|
vTaskDelay(pdMS_TO_TICKS(4000));
|
|
ESP_LOGW("selftest", "SDIO TX STRESS START: streaming mic audio 40s");
|
|
audio_capture_start();
|
|
for (int i = 0; i < 40; i++) {
|
|
vTaskDelay(pdMS_TO_TICKS(1000));
|
|
ESP_LOGW("selftest", "SDIO TX STRESS: alive %ds (no wedge)", i + 1);
|
|
}
|
|
audio_capture_stop();
|
|
ESP_LOGW("selftest", "SDIO TX STRESS SURVIVED 40s — #167 FIX CONFIRMED");
|
|
vTaskDelete(NULL);
|
|
}
|
|
|
|
void sdio_tx_selftest_kick(void)
|
|
{
|
|
static bool started;
|
|
if (!started) {
|
|
started = true;
|
|
xTaskCreate(sdio_tx_selftest_task, "sdiotest", 4096, NULL, 4, NULL);
|
|
}
|
|
}
|
|
#endif
|
|
|
|
void app_main(void)
|
|
{
|
|
ESP_LOGI(TAG, "DeskLock starting");
|
|
|
|
s_events = xQueueCreate(8, sizeof(app_event_t));
|
|
|
|
bsp_display_start();
|
|
bsp_display_backlight_on();
|
|
face_init();
|
|
|
|
audio_init();
|
|
if (esp_reset_reason() == ESP_RST_POWERON) {
|
|
audio_play_gong(); /* cold boot only — recovery reboots stay silent */
|
|
}
|
|
|
|
xTaskCreate(app_task, "app", 4096, NULL, 5, NULL);
|
|
|
|
#if WIFI_DIAG_MODE
|
|
wifi_diag_start();
|
|
#else
|
|
net_start();
|
|
c6_ota_start(); /* one-shot C6 radio firmware update — remove once proven */
|
|
#endif
|
|
|
|
ESP_LOGI(TAG, "DeskLock up");
|
|
}
|