70 lines
2.8 KiB
C
70 lines
2.8 KiB
C
#pragma once
|
|
|
|
#include <stdbool.h>
|
|
#include <stddef.h>
|
|
#include <stdint.h>
|
|
|
|
#include "esp_err.h"
|
|
|
|
#ifdef __cplusplus
|
|
extern "C" {
|
|
#endif
|
|
|
|
/* 上行回调接收已经编码的 Opus 包；返回错误会被采集任务记录并按策略处理。 */
|
|
typedef esp_err_t (*multimodal_audio_uplink_cb_t)(const uint8_t *data, size_t len, void *ctx);
|
|
/* PCM 观察者运行在采集任务上下文，必须有界且不可阻塞。 */
|
|
typedef void (*multimodal_audio_pcm_observer_t)(const int16_t *pcm,
|
|
size_t samples,
|
|
uint32_t sample_rate_hz,
|
|
void *ctx);
|
|
|
|
typedef enum {
|
|
MULTIMODAL_AUDIO_DOWNSTREAM_PLAYBACK_STARTED = 0,
|
|
MULTIMODAL_AUDIO_DOWNSTREAM_PLAYBACK_COMPLETED,
|
|
MULTIMODAL_AUDIO_DOWNSTREAM_PLAYBACK_STOPPED,
|
|
MULTIMODAL_AUDIO_DOWNSTREAM_PLAYBACK_STALLED,
|
|
} multimodal_audio_downstream_playback_event_type_t;
|
|
|
|
/* generation 将开始、完成、停止或卡顿事件绑定到唯一一次下行响应。 */
|
|
typedef struct {
|
|
multimodal_audio_downstream_playback_event_type_t type;
|
|
uint32_t playback_generation;
|
|
} multimodal_audio_downstream_playback_event_t;
|
|
|
|
typedef void (*multimodal_audio_downstream_playback_callback_t)(
|
|
const multimodal_audio_downstream_playback_event_t *event,
|
|
void *ctx);
|
|
|
|
esp_err_t multimodal_audio_init(void);
|
|
void multimodal_audio_deinit(void);
|
|
|
|
/* 上行和下行工作任务共享 codec/I2S 资源，但生命周期和停止请求彼此独立。 */
|
|
esp_err_t multimodal_audio_start_uplink(multimodal_audio_uplink_cb_t cb, void *cb_ctx);
|
|
esp_err_t multimodal_audio_stop_uplink(int timeout_ms);
|
|
bool multimodal_audio_is_uplink_running(void);
|
|
bool multimodal_audio_is_initialized(void);
|
|
esp_err_t multimodal_audio_set_speaker_output_enabled(bool enabled);
|
|
esp_err_t multimodal_audio_register_pcm_observer(
|
|
multimodal_audio_pcm_observer_t observer,
|
|
void *observer_ctx);
|
|
esp_err_t multimodal_audio_unregister_pcm_observer(
|
|
multimodal_audio_pcm_observer_t observer,
|
|
void *observer_ctx,
|
|
int timeout_ms);
|
|
|
|
/* 下行响应先 begin 建代、持续送入 Opus 包，最后 finish 标记服务端数据结束。 */
|
|
esp_err_t multimodal_audio_handle_downstream_opus(const uint8_t *data, size_t len);
|
|
esp_err_t multimodal_audio_begin_downstream_response(uint32_t *out_playback_generation);
|
|
esp_err_t multimodal_audio_finish_downstream_response(void);
|
|
esp_err_t multimodal_audio_wait_downstream_playback_complete(
|
|
uint32_t playback_generation,
|
|
int timeout_ms);
|
|
esp_err_t multimodal_audio_stop_downstream_playback(void);
|
|
esp_err_t multimodal_audio_register_downstream_playback_callback(
|
|
multimodal_audio_downstream_playback_callback_t cb,
|
|
void *cb_ctx);
|
|
|
|
#ifdef __cplusplus
|
|
}
|
|
#endif
|