Siren audio engine in progress - Implementing Impulse-Response based reverb
This commit is contained in:
@@ -94,11 +94,11 @@ void KRAudioManager::renderAudio(UInt32 inNumberFrames, AudioBufferList *ioData)
|
||||
int output_frame = 0;
|
||||
|
||||
while(output_frame < inNumberFrames) {
|
||||
int frames_ready = KRENGINE_FILTER_LENGTH - m_output_sample;
|
||||
int frames_ready = KRENGINE_AUDIO_BLOCK_LENGTH - m_output_sample;
|
||||
if(frames_ready == 0) {
|
||||
renderBlock();
|
||||
m_output_sample = 0;
|
||||
frames_ready = KRENGINE_FILTER_LENGTH;
|
||||
frames_ready = KRENGINE_AUDIO_BLOCK_LENGTH;
|
||||
}
|
||||
|
||||
int frames_processed = inNumberFrames - output_frame;
|
||||
@@ -138,7 +138,7 @@ void KRAudioManager::renderAudio(UInt32 inNumberFrames, AudioBufferList *ioData)
|
||||
|
||||
float *KRAudioManager::getBlockAddress(int block_offset)
|
||||
{
|
||||
return m_output_accumulation + (m_output_accumulation_block_start + block_offset * KRENGINE_FILTER_LENGTH * KRENGINE_MAX_OUTPUT_CHANNELS) % (KRENGINE_REVERB_MAX_SAMPLES * KRENGINE_MAX_OUTPUT_CHANNELS);
|
||||
return m_output_accumulation + (m_output_accumulation_block_start + block_offset * KRENGINE_AUDIO_BLOCK_LENGTH * KRENGINE_MAX_OUTPUT_CHANNELS) % (KRENGINE_REVERB_MAX_SAMPLES * KRENGINE_MAX_OUTPUT_CHANNELS);
|
||||
}
|
||||
|
||||
void KRAudioManager::renderReverbImpulseResponse(KRAudioSample *impulse_response, int impulse_response_block)
|
||||
@@ -150,8 +150,8 @@ void KRAudioManager::renderReverbImpulseResponse(KRAudioSample *impulse_response
|
||||
float *block_data = getBlockAddress(0);
|
||||
float *next_block = getBlockAddress(1);
|
||||
|
||||
float impulse_block_start_index = impulse_response_block * KRENGINE_FILTER_LENGTH;
|
||||
int reverb_block_start_index = (m_reverb_input_next_sample - KRENGINE_FILTER_LENGTH * (impulse_response_block + 1));
|
||||
float impulse_block_start_index = impulse_response_block * KRENGINE_AUDIO_BLOCK_LENGTH;
|
||||
int reverb_block_start_index = (m_reverb_input_next_sample - KRENGINE_AUDIO_BLOCK_LENGTH * (impulse_response_block + 1));
|
||||
if(reverb_block_start_index < 0) {
|
||||
reverb_block_start_index += KRENGINE_REVERB_MAX_SAMPLES;
|
||||
} else {
|
||||
@@ -159,30 +159,29 @@ void KRAudioManager::renderReverbImpulseResponse(KRAudioSample *impulse_response
|
||||
}
|
||||
|
||||
|
||||
|
||||
memcpy(reverb_sample_data_complex.realp, m_reverb_input_samples + reverb_block_start_index, KRENGINE_FILTER_LENGTH * sizeof(float));
|
||||
memset(reverb_sample_data_complex.realp + KRENGINE_FILTER_LENGTH, 0, KRENGINE_FILTER_LENGTH * sizeof(float));
|
||||
memset(reverb_sample_data_complex.imagp, 0, KRENGINE_FILTER_LENGTH * 2 * sizeof(float));
|
||||
memcpy(reverb_sample_data_complex.realp, m_reverb_input_samples + reverb_block_start_index, KRENGINE_AUDIO_BLOCK_LENGTH * sizeof(float));
|
||||
memset(reverb_sample_data_complex.realp + KRENGINE_AUDIO_BLOCK_LENGTH, 0, KRENGINE_AUDIO_BLOCK_LENGTH * sizeof(float));
|
||||
memset(reverb_sample_data_complex.imagp, 0, KRENGINE_AUDIO_BLOCK_LENGTH * 2 * sizeof(float));
|
||||
|
||||
|
||||
|
||||
vDSP_fft_zip(m_fft_setup, &reverb_sample_data_complex, 1, KRENGINE_FILTER_LOG2 + 1, kFFTDirection_Forward);
|
||||
vDSP_fft_zip(m_fft_setup, &reverb_sample_data_complex, 1, KRENGINE_AUDIO_BLOCK_LOG2N + 1, kFFTDirection_Forward);
|
||||
|
||||
float scale = 1.0f / (2 * (KRENGINE_FILTER_LENGTH * 2));
|
||||
float scale = 1.0f / (2 * KRENGINE_AUDIO_BLOCK_LENGTH);
|
||||
|
||||
int impulse_response_channels = 2;
|
||||
for(int channel=0; channel < impulse_response_channels; channel++) {
|
||||
|
||||
impulse_response->sample(impulse_block_start_index, KRENGINE_FILTER_LENGTH, channel, impulse_block_data_complex.realp, 1.0f);
|
||||
memset(impulse_block_data_complex.realp + KRENGINE_FILTER_LENGTH, 0, KRENGINE_FILTER_LENGTH * sizeof(float));
|
||||
memset(impulse_block_data_complex.imagp, 0, KRENGINE_FILTER_LENGTH * 2 * sizeof(float));
|
||||
impulse_response->sample(impulse_block_start_index, KRENGINE_AUDIO_BLOCK_LENGTH, channel, impulse_block_data_complex.realp, 1.0f);
|
||||
memset(impulse_block_data_complex.realp + KRENGINE_AUDIO_BLOCK_LENGTH, 0, KRENGINE_AUDIO_BLOCK_LENGTH * sizeof(float));
|
||||
memset(impulse_block_data_complex.imagp, 0, KRENGINE_AUDIO_BLOCK_LENGTH * 2 * sizeof(float));
|
||||
|
||||
vDSP_fft_zip(m_fft_setup, &impulse_block_data_complex, 1, KRENGINE_FILTER_LOG2 + 1, kFFTDirection_Forward);
|
||||
vDSP_zvmul(&reverb_sample_data_complex, 1, &impulse_block_data_complex, 1, &conv_data_complex, 1, KRENGINE_FILTER_LENGTH * 2, 1);
|
||||
vDSP_fft_zip(m_fft_setup, &conv_data_complex, 1, KRENGINE_FILTER_LOG2 + 1, kFFTDirection_Inverse);
|
||||
vDSP_vsmul(conv_data_complex.realp, 1, &scale, conv_data_complex.realp, 1, KRENGINE_FILTER_LENGTH * 2);
|
||||
vDSP_vadd(block_data + channel, impulse_response_channels, conv_data_complex.realp, 1, block_data + channel, impulse_response_channels, KRENGINE_FILTER_LENGTH);
|
||||
vDSP_vadd(next_block + channel, impulse_response_channels, conv_data_complex.realp + KRENGINE_FILTER_LENGTH, 1, next_block + channel, impulse_response_channels, KRENGINE_FILTER_LENGTH);
|
||||
vDSP_fft_zip(m_fft_setup, &impulse_block_data_complex, 1, KRENGINE_AUDIO_BLOCK_LOG2N + 1, kFFTDirection_Forward);
|
||||
vDSP_zvmul(&reverb_sample_data_complex, 1, &impulse_block_data_complex, 1, &conv_data_complex, 1, KRENGINE_AUDIO_BLOCK_LENGTH * 2, 1);
|
||||
vDSP_fft_zip(m_fft_setup, &conv_data_complex, 1, KRENGINE_AUDIO_BLOCK_LOG2N + 1, kFFTDirection_Inverse);
|
||||
vDSP_vsmul(conv_data_complex.realp, 1, &scale, conv_data_complex.realp, 1, KRENGINE_AUDIO_BLOCK_LENGTH * 2);
|
||||
vDSP_vadd(block_data + channel, impulse_response_channels, conv_data_complex.realp, 1, block_data + channel, impulse_response_channels, KRENGINE_AUDIO_BLOCK_LENGTH);
|
||||
vDSP_vadd(next_block + channel, impulse_response_channels, conv_data_complex.realp + KRENGINE_AUDIO_BLOCK_LENGTH, 1, next_block + channel, impulse_response_channels, KRENGINE_AUDIO_BLOCK_LENGTH);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -193,23 +192,23 @@ void KRAudioManager::renderHRTF()
|
||||
|
||||
void KRAudioManager::renderReverb()
|
||||
{
|
||||
float reverb_data[KRENGINE_FILTER_LENGTH];
|
||||
float reverb_data[KRENGINE_AUDIO_BLOCK_LENGTH];
|
||||
|
||||
float *reverb_accum = m_reverb_input_samples + m_reverb_input_next_sample;
|
||||
memset(reverb_accum, 0, sizeof(float) * KRENGINE_FILTER_LENGTH);
|
||||
memset(reverb_accum, 0, sizeof(float) * KRENGINE_AUDIO_BLOCK_LENGTH);
|
||||
for(std::set<KRAudioSource *>::iterator itr=m_activeAudioSources.begin(); itr != m_activeAudioSources.end(); itr++) {
|
||||
KRAudioSource *source = *itr;
|
||||
float reverb_send_level = m_global_reverb_send_level * source->getReverb();
|
||||
if(reverb_send_level > 0.0f) {
|
||||
KRAudioSample *sample = source->getAudioSample();
|
||||
if(sample) {
|
||||
sample->sample((int)((__int64_t)m_audio_frame - source->getStartAudioFrame()), KRENGINE_FILTER_LENGTH, 0, reverb_data, reverb_send_level);
|
||||
vDSP_vadd(reverb_accum, 1, reverb_data, 1, reverb_accum, 1, KRENGINE_FILTER_LENGTH);
|
||||
sample->sample((int)((__int64_t)m_audio_frame - source->getStartAudioFrame()), KRENGINE_AUDIO_BLOCK_LENGTH, 0, reverb_data, reverb_send_level);
|
||||
vDSP_vadd(reverb_accum, 1, reverb_data, 1, reverb_accum, 1, KRENGINE_AUDIO_BLOCK_LENGTH);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
m_reverb_input_next_sample += KRENGINE_FILTER_LENGTH;
|
||||
m_reverb_input_next_sample += KRENGINE_AUDIO_BLOCK_LENGTH;
|
||||
if(m_reverb_input_next_sample >= KRENGINE_REVERB_MAX_SAMPLES) {
|
||||
m_reverb_input_next_sample = 0;
|
||||
}
|
||||
@@ -218,7 +217,7 @@ void KRAudioManager::renderReverb()
|
||||
// KRAudioSample *impulse_response = getContext().getAudioManager()->get("hrtf_kemar_H10e040a");
|
||||
KRAudioSample *impulse_response_sample = getContext().getAudioManager()->get("test_reverb");
|
||||
if(impulse_response_sample) {
|
||||
int impulse_response_blocks = impulse_response_sample->getFrameCount() / KRENGINE_FILTER_LENGTH + 1;
|
||||
int impulse_response_blocks = impulse_response_sample->getFrameCount() / KRENGINE_AUDIO_BLOCK_LENGTH + 1;
|
||||
for(int impulse_response_block=0; impulse_response_block < impulse_response_blocks; impulse_response_block++) {
|
||||
renderReverbImpulseResponse(impulse_response_sample, impulse_response_block);
|
||||
}
|
||||
@@ -231,10 +230,10 @@ void KRAudioManager::renderBlock()
|
||||
|
||||
// Zero out block that was last used, so it will be ready for the next pass through the circular buffer
|
||||
float *block_data = getBlockAddress(0);
|
||||
memset(block_data, 0, KRENGINE_FILTER_LENGTH * KRENGINE_MAX_OUTPUT_CHANNELS * sizeof(float));
|
||||
memset(block_data, 0, KRENGINE_AUDIO_BLOCK_LENGTH * KRENGINE_MAX_OUTPUT_CHANNELS * sizeof(float));
|
||||
|
||||
// Advance to the next block, and wrap around
|
||||
m_output_accumulation_block_start = (m_output_accumulation_block_start + KRENGINE_FILTER_LENGTH * KRENGINE_MAX_OUTPUT_CHANNELS) % (KRENGINE_REVERB_MAX_SAMPLES * KRENGINE_MAX_OUTPUT_CHANNELS);
|
||||
m_output_accumulation_block_start = (m_output_accumulation_block_start + KRENGINE_AUDIO_BLOCK_LENGTH * KRENGINE_MAX_OUTPUT_CHANNELS) % (KRENGINE_REVERB_MAX_SAMPLES * KRENGINE_MAX_OUTPUT_CHANNELS);
|
||||
|
||||
// ----====---- Render Direct / HRTF audio ----====----
|
||||
renderHRTF();
|
||||
@@ -244,7 +243,7 @@ void KRAudioManager::renderBlock()
|
||||
renderReverb();
|
||||
|
||||
// ----====---- Advance audio sources ----====----
|
||||
m_audio_frame += KRENGINE_FILTER_LENGTH;
|
||||
m_audio_frame += KRENGINE_AUDIO_BLOCK_LENGTH;
|
||||
}
|
||||
|
||||
// audio render procedure, don't allocate memory, don't take any locks, don't waste time
|
||||
@@ -277,7 +276,7 @@ void KRSetAUCanonical(AudioStreamBasicDescription &desc, UInt32 nChannels, bool
|
||||
void KRAudioManager::initSiren()
|
||||
{
|
||||
if(m_auGraph == NULL) {
|
||||
m_output_sample = KRENGINE_FILTER_LENGTH;
|
||||
m_output_sample = KRENGINE_AUDIO_BLOCK_LENGTH;
|
||||
|
||||
// initialize double-buffer for reverb input
|
||||
int buffer_size = sizeof(float) * KRENGINE_REVERB_MAX_SAMPLES; // Reverb input is a single channel, circular buffered
|
||||
|
||||
@@ -44,12 +44,12 @@ const int KRENGINE_AUDIO_MAX_POOL_SIZE = 32;
|
||||
const int KRENGINE_AUDIO_MAX_BUFFER_SIZE = 64*1024;
|
||||
const int KRENGINE_AUDIO_BUFFERS_PER_SOURCE = 3;
|
||||
|
||||
const int KRENGINE_FILTER_LENGTH = 128; // Size for FFT's used in HRTF convolution
|
||||
const int KRENGINE_FILTER_LOG2 = 7; // 2 ^ 7 = 128
|
||||
const int KRENGINE_AUDIO_BLOCK_LENGTH = 256; // Length of one block to process. Determines the latency of the audio system and sets size for FFT's used in HRTF convolution
|
||||
const int KRENGINE_AUDIO_BLOCK_LOG2N = 8; // 2 ^ KRENGINE_AUDIO_BLOCK_LOG2N = KRENGINE_AUDIO_BLOCK_LENGTH
|
||||
|
||||
const int KRENGINE_REVERB_WORKSPACE_SIZE = KRENGINE_FILTER_LENGTH * 2;
|
||||
const int KRENGINE_REVERB_WORKSPACE_SIZE = 1024;
|
||||
|
||||
const int KRENGINE_REVERB_MAX_SAMPLES = 435200; // At least 10s reverb impulse response length, divisible by KRENGINE_REVERB_FILTER_LENGTH
|
||||
const int KRENGINE_REVERB_MAX_SAMPLES = 435200; // At least 10s reverb impulse response length, divisible by KRENGINE_AUDIO_BLOCK_LENGTH
|
||||
const int KRENGINE_MAX_REVERB_IMPULSE_MIX = 16; // Maximum number of impulse response filters that can be mixed simultaneously
|
||||
const int KRENGINE_MAX_OUTPUT_CHANNELS = 2;
|
||||
|
||||
|
||||
Reference in New Issue
Block a user