Audio pipeline: the real-time task that pulls PCM from a source and pushes it to an Audio sink interface. More...

#include "esp_err.h"
#include "audio_sink.h"
#include <stdint.h>

Enumerations
enum	pipeline_end_reason_t { PIPE_END_EOF , PIPE_END_ERROR , PIPE_END_UNSUPPORTED }
	Why a file playback ended, reported to the track-end callback. More...

Functions
esp_err_t	pipeline_init (void)
	Create the pinned audio task and its command queue.
void	pipeline_set_track_end_cb (void(*cb)(pipeline_end_reason_t reason))
	Register the callback fired when a file playback ends on its own (EOF or error).
esp_err_t	pipeline_play_file (const char *path)
	Play an audio file: decode it to the current sink at the file's own format.
esp_err_t	pipeline_pause (void)
	Pause file playback: stop feeding the sink while keeping the stream position.
esp_err_t	pipeline_resume (void)
	Resume a playback paused with pipeline_pause(): restart the sink and keep decoding.
void	pipeline_set_sink (const audio_sink_t *sink)
	Select the output backend used by the next playback.
esp_err_t	pipeline_switch_sink (void)
	Re-route the current playback onto the sink last set by pipeline_set_sink().
void	pipeline_get_position (uint32_t elapsed_ms, uint32_t total_ms)
	Read a tear-free snapshot of the current playback position.
esp_err_t	pipeline_play_tone (uint32_t freq_hz)
	Bring-up/diagnostic: stream a continuous sine to the current sink.
esp_err_t	pipeline_stop (void)
	Stop playback and power the output path down. Safe to call when already idle.

Detailed Description

Audio pipeline: the real-time task that pulls PCM from a source and pushes it to an Audio sink interface.

The pipeline owns one high-priority FreeRTOS task pinned to the application core (kept off core 0, where the Wi-Fi/BT stack runs). The task drives a producer->consumer loop: fill a buffer from the current source, hand it to sink->write(), repeat. Only the fill step depends on the source: the decoder for file playback, or a synthetic sine (bring-up/diagnostic, below). Everything else (task, buffering, sink sequencing) is shared, so the tone exercises the real pipeline without needing the SD card.

Commands are posted from any task and serviced by the audio task; the public calls below never block on the audio path.

Note: Développé avec l'assistance de Claude Opus 4.8 (Anthropic), sous la direction de Y. Dos Santos : spécification, revue et validation sur cible par l'auteur.

Enumerations

Functions

Detailed Description