15static constexpr uint32_t BUFFER_DURATION_MS = 16;
20 input_stream_info.
get_channels(), this->target_sample_rate_);
35 "Resampler Microphone:\n"
36 " Target Sample Rate: %" PRIu32
" Hz\n"
100 input_stream_info.
get_channels(), this->target_sample_rate_);
102 auto resampler = make_unique<PositionResampler>(input_stream_info.
ms_to_samples(BUFFER_DURATION_MS),
105 esp_audio_libs::resampler::ResamplerConfiguration resample_config = {
106 .source_sample_rate =
static_cast<float>(input_stream_info.
get_sample_rate()),
113 .subsample_interpolate =
false,
114 .number_of_taps = this->
taps_,
115 .number_of_filters = this->
filters_,
118 if (!resampler->initialize(resample_config)) {
119 ESP_LOGE(TAG,
"Not enough memory to resample");
146 for (
uint32_t done = 0; done < total_frames;) {
147 const uint32_t frames = std::min(total_frames - done, max_frames);
149 const int64_t time_us =
161 const size_t input_bytes_per_frame = input_stream_info.
frames_to_bytes(1);
166 const uint8_t *input = capture.
data.data();
169 int64_t frames_consumed = 0;
170 while (input_frames > 0) {
176 esp_audio_libs::resampler::ResamplerResults results = this->
resampler_->resample(
177 input, this->
output_buffer_.data(), std::min(input_frames, max_input_frames), max_output_frames, 0.0f);
179 input += results.frames_used * input_bytes_per_frame;
180 input_frames -= results.frames_used;
181 frames_consumed += results.frames_used;
183 if (results.frames_generated > 0) {
184 int64_t output_time_us = 0;
191 output_time_us = capture.
capture_time_us + (frames_consumed - this->
taps_ / 2) * 1000000 / input_rate +
192 static_cast<int64_t
>(this->
resampler_->get_position() * 1000000.0f / input_rate) -
197 }
else if (results.frames_used == 0) {
void mark_failed()
Mark this component as failed.
void enable_loop()
Enable this component's loop.
void disable_loop()
Disable this component's loop.
size_t ms_to_bytes(uint32_t ms) const
Converts duration to bytes.
size_t frames_to_bytes(uint32_t frames) const
Converts frames to bytes.
uint8_t get_bits_per_sample() const
uint32_t frames_to_microseconds(uint32_t frames) const
Computes the duration, in microseconds, the given amount of frames represents.
uint32_t ms_to_samples(uint32_t ms) const
Converts duration to samples.
uint32_t bytes_to_frames(size_t bytes) const
Convert bytes to frames.
uint8_t get_channels() const
uint32_t ms_to_frames(uint32_t ms) const
Converts duration to frames.
uint32_t get_sample_rate() const
void publish_audio_(std::span< uint8_t > data, const audio::AudioStreamInfo &stream_info, int64_t capture_time_us=0)
While muted, data is zeroed first, so it must be a buffer the platform owns.
DataCallbackManager data_callbacks_
audio::AudioStreamInfo audio_stream_info_
void add_data_callback(F &&data_callback)
Same callbacks as Microphone::add_data_callback; stream_info is the processed format.
audio::AudioStreamInfo get_audio_stream_info()
Gets the AudioStreamInfo of the data after processing.
bool init_resampler_(const audio::AudioStreamInfo &input_stream_info)
Sets up the resampler for the given input format.
microphone::MicrophoneSource * source_
uint8_t active_listeners_
audio::AudioStreamInfo input_stream_info_
uint32_t target_sample_rate_
void process_audio_(const microphone::AudioCapture &capture)
Resamples a chunk of source audio and passes it to the data callbacks. Source microphone task only.
void dump_config() override
std::unique_ptr< PositionResampler > resampler_
std::vector< uint8_t > output_buffer_
ESPHOME_LOG_TAG(TAG, "resampler.microphone")
int64_t capture_time_us
esp_timer_get_time() of the first frame, or 0 if unknown
audio::AudioStreamInfo stream_info
Format of data.
std::span< const uint8_t > data
Only valid during the callback.