AudioDevice* AudioDevice::create(){
int ret;
AudioDeviceImpl *impl = new AudioDeviceImpl();
impl->audio = webrtc::CreateAudioDeviceModule(0, webrtc::AudioDeviceModule::kPlatformDefaultAudio);
if(!impl->audio){
goto err;
}
ret = impl->audio->Init();
if(ret == -1){
goto err;
}
WebRtcVad_Create(&impl->vad);
if(!impl->vad){
goto err;
}
ret = WebRtcVad_Init(impl->vad);
if(ret == -1){
goto err;
}
{
uint32_t device_rate = 0;
ret = impl->audio->RecordingSampleRate(&device_rate);
if(ret == -1){
goto err;
}
impl->set_input_sample_rate(device_rate);
}
{
uint32_t device_rate = 0;
ret = impl->audio->PlayoutSampleRate(&device_rate);
if(ret == -1){
goto err;
}
impl->set_output_sample_rate(device_rate);
}
ret = impl->audio->RegisterAudioCallback(impl);
if(ret == -1){
goto err;
}
return impl;
err:
delete impl;
return NULL;
}
RTC::ReturnCode_t WebRTCVAD::onInitialize()
{
RTC_DEBUG(("onInitialize start"));
RTC_INFO(("WebRTCVAD : WebRTC based noise reduction component"));
RTC_INFO((" Copyright (C) 2010-2011 Yosuke Matsusaka and AIST-OpenHRI development team"));
RTC_INFO((" Version %s", VERSION));
// Registration: InPort/OutPort/Service
// <rtc-template block="registration">
// Set InPort buffers
addInPort("AudioDataIn", m_ninIn);
m_ninIn.setDescription(_("Audio data input."));
/* setiing datalistener event */
m_ninIn.addConnectorDataListener(ON_BUFFER_WRITE,
new DataListener("ON_BUFFER_WRITE", this));
// Set OutPort buffer
addOutPort("AudioDataOut", m_foutOut);
m_foutOut.setDescription(_("Audio data output."));
// Set service provider to Ports
// Set service consumers to Ports
// Set CORBA Service Ports
// </rtc-template>
bindParameter("FilterLength", m_bufferlen, "5");//ADDED
WebRtcVad_Create(&handle);
WebRtcVad_Init(handle);
WebRtcVad_set_mode(handle, 2); // agressive adaptation mode
RTC_DEBUG(("onInitialize finish"));
return RTC::RTC_OK;
}
Filter_Audio *new_filter_audio(uint32_t fs)
{
if (fs == 0) {
return NULL;
}
Filter_Audio *f_a = calloc(sizeof(Filter_Audio), 1);
if (!f_a) {
return NULL;
}
f_a->fs = fs;
if (fs != 16000)
fs = 32000;
init_highpass_filter_zam(&f_a->hpfa, 100, (float) f_a->fs);
init_highpass_filter_zam(&f_a->hpfb, 100, (float) f_a->fs);
unsigned int lowpass_filter_frequency = 12000;
if (f_a->fs > (lowpass_filter_frequency * 2)) {
init_lowpass_filter_zam(&f_a->lpfa, lowpass_filter_frequency, (float) f_a->fs);
init_lowpass_filter_zam(&f_a->lpfb, lowpass_filter_frequency, (float) f_a->fs);
f_a->lowpass_enabled = 1;
}
if (WebRtcAgc_Create(&f_a->gain_control) == -1) {
free(f_a);
return NULL;
}
if (WebRtcNsx_Create(&f_a->noise_sup_x) == -1) {
WebRtcAgc_Free(f_a->gain_control);
free(f_a);
return NULL;
}
if (WebRtcAec_Create(&f_a->echo_cancellation) == -1) {
WebRtcAgc_Free(f_a->gain_control);
WebRtcNsx_Free(f_a->noise_sup_x);
free(f_a);
return NULL;
}
if (WebRtcVad_Create(&f_a->Vad_handle) == -1){
WebRtcAec_Free(f_a->echo_cancellation);
WebRtcAgc_Free(f_a->gain_control);
WebRtcNsx_Free(f_a->noise_sup_x);
free(f_a);
return NULL;
}
WebRtcAgc_config_t gain_config;
gain_config.targetLevelDbfs = 1;
gain_config.compressionGaindB = 20;
gain_config.limiterEnable = kAgcTrue;
if (WebRtcAgc_Init(f_a->gain_control, 0, 255, kAgcModeAdaptiveDigital, fs) == -1 || WebRtcAgc_set_config(f_a->gain_control, gain_config) == -1) {
kill_filter_audio(f_a);
return NULL;
}
if (WebRtcNsx_Init(f_a->noise_sup_x, fs) == -1 || WebRtcNsx_set_policy(f_a->noise_sup_x, 2) == -1) {
kill_filter_audio(f_a);
return NULL;
}
AecConfig echo_config;
echo_config.nlpMode = kAecNlpAggressive;
echo_config.skewMode = kAecFalse;
echo_config.metricsMode = kAecFalse;
echo_config.delay_logging = kAecFalse;
if (WebRtcAec_Init(f_a->echo_cancellation, fs, f_a->fs) == -1 || WebRtcAec_set_config(f_a->echo_cancellation, echo_config) == -1) {
kill_filter_audio(f_a);
return NULL;
}
int vad_mode = 1; //Aggressiveness mode (0, 1, 2, or 3).
if (WebRtcVad_Init(f_a->Vad_handle) == -1 || WebRtcVad_set_mode(f_a->Vad_handle,vad_mode) == -1){
kill_filter_audio(f_a);
return NULL;
}
f_a->echo_enabled = 1;
f_a->gain_enabled = 1;
f_a->noise_enabled = 1;
f_a->vad_enabled = 1;
int quality = 4;
if (f_a->fs != 16000) {
f_a->downsampler = speex_resampler_init(1, f_a->fs, 32000, quality, 0);
f_a->upsampler = speex_resampler_init(1, 32000, f_a->fs, quality, 0);
/* quality doesn't need to be high for this one. */
quality = 0;
f_a->downsampler_echo = speex_resampler_init(1, f_a->fs, 16000, quality, 0);
if (!f_a->upsampler || !f_a->downsampler || !f_a->downsampler_echo) {
kill_filter_audio(f_a);
return NULL;
}
}
return f_a;
}
int main(int argc, char* argv[])
{
FILE * infile, *outfile, *statefile;
short res=0,errtype;
/*float time=0.0;*/
WebRtcVad_Create(&vinst);
WebRtcVad_Init(vinst);
short size;
int samps=0;
if (argc < 5){
printf("Usage:\n CNG.exe infile outfile samplingfreq(Hz) interval(ms) order\n\n");
return(0);
}
infile=fopen(argv[1],"rb");
if (infile==NULL){
printf("file %s does not exist\n",argv[1]);
return(0);
}
outfile=fopen(argv[2],"wb");
statefile=fopen("CNGVAD.d","wb");
if (outfile==NULL){
printf("file %s could not be created\n",argv[2]);
return(0);
}
unsigned int fs=16000;
short frameLen=fs/50;
#ifndef ASSIGN
res=WebRtcCng_CreateEnc(&e_inst);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in initialization: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
res=WebRtcCng_CreateDec(&d_inst);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeDec(d_inst);
fprintf(stderr,"\n\n Error in initialization: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
#else
// Test the Assign-functions
int Esize, Dsize;
void *Eaddr, *Daddr;
res=WebRtcCng_AssignSizeEnc(&Esize);
res=WebRtcCng_AssignSizeDec(&Dsize);
Eaddr=malloc(Esize);
Daddr=malloc(Dsize);
res=WebRtcCng_AssignEnc(&e_inst, Eaddr);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in initialization: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
res=WebRtcCng_AssignDec(&d_inst, Daddr);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeDec(d_inst);
fprintf(stderr,"\n\n Error in initialization: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
#endif
res=WebRtcCng_InitEnc(e_inst,atoi(argv[3]),atoi(argv[4]),atoi(argv[5]));
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in initialization: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
res=WebRtcCng_InitDec(d_inst);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeDec(d_inst);
fprintf(stderr,"\n\n Error in initialization: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
static bool firstSilent=true;
int numSamp=0;
int speech=0;
int silent=0;
long cnt=0;
while(fread(anaSpeech,2,frameLen,infile)==frameLen){
cnt++;
if (cnt==60){
cnt=60;
}
/* time+=(float)frameLen/fs;
numSamp+=frameLen;
float temp[640];
for(unsigned int j=0;j<frameLen;j++)
temp[j]=(float)anaSpeech[j]; */
// if(!WebRtcVad_Process(vinst, fs, anaSpeech, frameLen)){
if(1){ // Do CNG coding of entire file
// if(!((anaSpeech[0]==0)&&(anaSpeech[1]==0)&&(anaSpeech[2]==0))){
if(firstSilent){
res = WebRtcCng_Encode(e_inst, anaSpeech, frameLen/2, SIDpkt,&size,1);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in encoder: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
firstSilent=false;
res=WebRtcCng_Encode(e_inst, &anaSpeech[frameLen/2], frameLen/2, SIDpkt,&size,1);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in encoder: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
}
else{
res=WebRtcCng_Encode(e_inst, anaSpeech, frameLen/2, SIDpkt,&size,0);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in encoder: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
res=WebRtcCng_Encode(e_inst, &anaSpeech[frameLen/2], frameLen/2, SIDpkt,&size,0);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeEnc(e_inst);
fprintf(stderr,"\n\n Error in encoder: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
}
if(size>0){
res=WebRtcCng_UpdateSid(d_inst,SIDpkt, size);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeDec(d_inst);
fprintf(stderr,"\n\n Error in decoder: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
}
res=WebRtcCng_Generate(d_inst,genSpeech, frameLen,0);
if (res < 0) {
/* exit if returned with error */
errtype=WebRtcCng_GetErrorCodeDec(d_inst);
fprintf(stderr,"\n\n Error in decoder: %d.\n\n", errtype);
exit(EXIT_FAILURE);
}
memcpy(state,anaSpeech,2*frameLen);
}
else{
firstSilent=true;
memcpy(genSpeech,anaSpeech,2*frameLen);
memset(anaSpeech,0,frameLen*2);
memset(state,0,frameLen*2);
}
fwrite(genSpeech,2,frameLen,outfile);
fwrite(state,2,frameLen,statefile);
}
fclose(infile);
fclose(outfile);
fclose(statefile);
return 0;
}