| /* |
| * Copyright (c) 2011 The WebRTC project authors. All Rights Reserved. |
| * |
| * Use of this source code is governed by a BSD-style license |
| * that can be found in the LICENSE file in the root of the source |
| * tree. An additional intellectual property rights grant can be found |
| * in the file PATENTS. All contributing project authors may |
| * be found in the AUTHORS file in the root of the source tree. |
| */ |
| |
| #include <string.h> |
| #include <math.h> |
| //#include <stdio.h> |
| #include <stdlib.h> |
| #include "noise_suppression.h" |
| #include "ns_core.h" |
| #include "windows_private.h" |
| #include "fft4g.h" |
| #include "signal_processing_library.h" |
| |
| // Set Feature Extraction Parameters |
| void WebRtcNs_set_feature_extraction_parameters(NSinst_t* inst) { |
| //bin size of histogram |
| inst->featureExtractionParams.binSizeLrt = (float)0.1; |
| inst->featureExtractionParams.binSizeSpecFlat = (float)0.05; |
| inst->featureExtractionParams.binSizeSpecDiff = (float)0.1; |
| |
| //range of histogram over which lrt threshold is computed |
| inst->featureExtractionParams.rangeAvgHistLrt = (float)1.0; |
| |
| //scale parameters: multiply dominant peaks of the histograms by scale factor to obtain |
| // thresholds for prior model |
| inst->featureExtractionParams.factor1ModelPars = (float)1.20; //for lrt and spectral diff |
| inst->featureExtractionParams.factor2ModelPars = (float)0.9; //for spectral_flatness: |
| // used when noise is flatter than speech |
| |
| //peak limit for spectral flatness (varies between 0 and 1) |
| inst->featureExtractionParams.thresPosSpecFlat = (float)0.6; |
| |
| //limit on spacing of two highest peaks in histogram: spacing determined by bin size |
| inst->featureExtractionParams.limitPeakSpacingSpecFlat = |
| 2 * inst->featureExtractionParams.binSizeSpecFlat; |
| inst->featureExtractionParams.limitPeakSpacingSpecDiff = |
| 2 * inst->featureExtractionParams.binSizeSpecDiff; |
| |
| //limit on relevance of second peak: |
| inst->featureExtractionParams.limitPeakWeightsSpecFlat = (float)0.5; |
| inst->featureExtractionParams.limitPeakWeightsSpecDiff = (float)0.5; |
| |
| // fluctuation limit of lrt feature |
| inst->featureExtractionParams.thresFluctLrt = (float)0.05; |
| |
| //limit on the max and min values for the feature thresholds |
| inst->featureExtractionParams.maxLrt = (float)1.0; |
| inst->featureExtractionParams.minLrt = (float)0.20; |
| |
| inst->featureExtractionParams.maxSpecFlat = (float)0.95; |
| inst->featureExtractionParams.minSpecFlat = (float)0.10; |
| |
| inst->featureExtractionParams.maxSpecDiff = (float)1.0; |
| inst->featureExtractionParams.minSpecDiff = (float)0.16; |
| |
| //criteria of weight of histogram peak to accept/reject feature |
| inst->featureExtractionParams.thresWeightSpecFlat = (int)(0.3 |
| * (inst->modelUpdatePars[1])); //for spectral flatness |
| inst->featureExtractionParams.thresWeightSpecDiff = (int)(0.3 |
| * (inst->modelUpdatePars[1])); //for spectral difference |
| } |
| |
| // Initialize state |
| int WebRtcNs_InitCore(NSinst_t* inst, WebRtc_UWord32 fs) { |
| int i; |
| //We only support 10ms frames |
| |
| //check for valid pointer |
| if (inst == NULL) { |
| return -1; |
| } |
| |
| // Initialization of struct |
| if (fs == 8000 || fs == 16000 || fs == 32000) { |
| inst->fs = fs; |
| } else { |
| return -1; |
| } |
| inst->windShift = 0; |
| if (fs == 8000) { |
| // We only support 10ms frames |
| inst->blockLen = 80; |
| inst->blockLen10ms = 80; |
| inst->anaLen = 128; |
| inst->window = kBlocks80w128; |
| inst->outLen = 0; |
| } else if (fs == 16000) { |
| // We only support 10ms frames |
| inst->blockLen = 160; |
| inst->blockLen10ms = 160; |
| inst->anaLen = 256; |
| inst->window = kBlocks160w256; |
| inst->outLen = 0; |
| } else if (fs == 32000) { |
| // We only support 10ms frames |
| inst->blockLen = 160; |
| inst->blockLen10ms = 160; |
| inst->anaLen = 256; |
| inst->window = kBlocks160w256; |
| inst->outLen = 0; |
| } |
| inst->magnLen = inst->anaLen / 2 + 1; // Number of frequency bins |
| |
| // Initialize fft work arrays. |
| inst->ip[0] = 0; // Setting this triggers initialization. |
| memset(inst->dataBuf, 0, sizeof(float) * ANAL_BLOCKL_MAX); |
| WebRtc_rdft(inst->anaLen, 1, inst->dataBuf, inst->ip, inst->wfft); |
| |
| memset(inst->dataBuf, 0, sizeof(float) * ANAL_BLOCKL_MAX); |
| memset(inst->syntBuf, 0, sizeof(float) * ANAL_BLOCKL_MAX); |
| |
| //for HB processing |
| memset(inst->dataBufHB, 0, sizeof(float) * ANAL_BLOCKL_MAX); |
| |
| //for quantile noise estimation |
| memset(inst->quantile, 0, sizeof(float) * HALF_ANAL_BLOCKL); |
| for (i = 0; i < SIMULT * HALF_ANAL_BLOCKL; i++) { |
| inst->lquantile[i] = (float)8.0; |
| inst->density[i] = (float)0.3; |
| } |
| |
| for (i = 0; i < SIMULT; i++) { |
| inst->counter[i] = (int)floor((float)(END_STARTUP_LONG * (i + 1)) / (float)SIMULT); |
| } |
| |
| inst->updates = 0; |
| |
| // Wiener filter initialization |
| for (i = 0; i < HALF_ANAL_BLOCKL; i++) { |
| inst->smooth[i] = (float)1.0; |
| } |
| |
| // Set the aggressiveness: default |
| inst->aggrMode = 0; |
| |
| //initialize variables for new method |
| inst->priorSpeechProb = (float)0.5; //prior prob for speech/noise |
| for (i = 0; i < HALF_ANAL_BLOCKL; i++) { |
| inst->magnPrev[i] = (float)0.0; //previous mag spectrum |
| inst->noisePrev[i] = (float)0.0; //previous noise-spectrum |
| inst->logLrtTimeAvg[i] = LRT_FEATURE_THR; //smooth LR ratio (same as threshold) |
| inst->magnAvgPause[i] = (float)0.0; //conservative noise spectrum estimate |
| inst->speechProbHB[i] = (float)0.0; //for estimation of HB in second pass |
| inst->initMagnEst[i] = (float)0.0; //initial average mag spectrum |
| } |
| |
| //feature quantities |
| inst->featureData[0] = SF_FEATURE_THR; //spectral flatness (start on threshold) |
| inst->featureData[1] = (float)0.0; //spectral entropy: not used in this version |
| inst->featureData[2] = (float)0.0; //spectral variance: not used in this version |
| inst->featureData[3] = LRT_FEATURE_THR; //average lrt factor (start on threshold) |
| inst->featureData[4] = SF_FEATURE_THR; //spectral template diff (start on threshold) |
| inst->featureData[5] = (float)0.0; //normalization for spectral-diff |
| inst->featureData[6] = (float)0.0; //window time-average of input magnitude spectrum |
| |
| //histogram quantities: used to estimate/update thresholds for features |
| for (i = 0; i < HIST_PAR_EST; i++) { |
| inst->histLrt[i] = 0; |
| inst->histSpecFlat[i] = 0; |
| inst->histSpecDiff[i] = 0; |
| } |
| |
| inst->blockInd = -1; //frame counter |
| inst->priorModelPars[0] = LRT_FEATURE_THR; //default threshold for lrt feature |
| inst->priorModelPars[1] = (float)0.5; //threshold for spectral flatness: |
| // determined on-line |
| inst->priorModelPars[2] = (float)1.0; //sgn_map par for spectral measure: |
| // 1 for flatness measure |
| inst->priorModelPars[3] = (float)0.5; //threshold for template-difference feature: |
| // determined on-line |
| inst->priorModelPars[4] = (float)1.0; //default weighting parameter for lrt feature |
| inst->priorModelPars[5] = (float)0.0; //default weighting parameter for |
| // spectral flatness feature |
| inst->priorModelPars[6] = (float)0.0; //default weighting parameter for |
| // spectral difference feature |
| |
| inst->modelUpdatePars[0] = 2; //update flag for parameters: |
| // 0 no update, 1=update once, 2=update every window |
| inst->modelUpdatePars[1] = 500; //window for update |
| inst->modelUpdatePars[2] = 0; //counter for update of conservative noise spectrum |
| //counter if the feature thresholds are updated during the sequence |
| inst->modelUpdatePars[3] = inst->modelUpdatePars[1]; |
| |
| inst->signalEnergy = 0.0; |
| inst->sumMagn = 0.0; |
| inst->whiteNoiseLevel = 0.0; |
| inst->pinkNoiseNumerator = 0.0; |
| inst->pinkNoiseExp = 0.0; |
| |
| WebRtcNs_set_feature_extraction_parameters(inst); // Set feature configuration |
| |
| //default mode |
| WebRtcNs_set_policy_core(inst, 0); |
| |
| |
| memset(inst->outBuf, 0, sizeof(float) * 3 * BLOCKL_MAX); |
| |
| inst->initFlag = 1; |
| return 0; |
| } |
| |
| int WebRtcNs_set_policy_core(NSinst_t* inst, int mode) { |
| // allow for modes:0,1,2,3 |
| if (mode < 0 || mode > 3) { |
| return (-1); |
| } |
| |
| inst->aggrMode = mode; |
| if (mode == 0) { |
| inst->overdrive = (float)1.0; |
| inst->denoiseBound = (float)0.5; |
| inst->gainmap = 0; |
| } else if (mode == 1) { |
| //inst->overdrive = (float)1.25; |
| inst->overdrive = (float)1.0; |
| inst->denoiseBound = (float)0.25; |
| inst->gainmap = 1; |
| } else if (mode == 2) { |
| //inst->overdrive = (float)1.25; |
| inst->overdrive = (float)1.1; |
| inst->denoiseBound = (float)0.125; |
| inst->gainmap = 1; |
| } else if (mode == 3) { |
| //inst->overdrive = (float)1.30; |
| inst->overdrive = (float)1.25; |
| inst->denoiseBound = (float)0.09; |
| inst->gainmap = 1; |
| } |
| return 0; |
| } |
| |
| // Estimate noise |
| void WebRtcNs_NoiseEstimation(NSinst_t* inst, float* magn, float* noise) { |
| int i, s, offset; |
| float lmagn[HALF_ANAL_BLOCKL], delta; |
| |
| if (inst->updates < END_STARTUP_LONG) { |
| inst->updates++; |
| } |
| |
| for (i = 0; i < inst->magnLen; i++) { |
| lmagn[i] = (float)log(magn[i]); |
| } |
| |
| // loop over simultaneous estimates |
| for (s = 0; s < SIMULT; s++) { |
| offset = s * inst->magnLen; |
| |
| // newquantest(...) |
| for (i = 0; i < inst->magnLen; i++) { |
| // compute delta |
| if (inst->density[offset + i] > 1.0) { |
| delta = FACTOR * (float)1.0 / inst->density[offset + i]; |
| } else { |
| delta = FACTOR; |
| } |
| |
| // update log quantile estimate |
| if (lmagn[i] > inst->lquantile[offset + i]) { |
| inst->lquantile[offset + i] += QUANTILE * delta |
| / (float)(inst->counter[s] + 1); |
| } else { |
| inst->lquantile[offset + i] -= ((float)1.0 - QUANTILE) * delta |
| / (float)(inst->counter[s] + 1); |
| } |
| |
| // update density estimate |
| if (fabs(lmagn[i] - inst->lquantile[offset + i]) < WIDTH) { |
| inst->density[offset + i] = ((float)inst->counter[s] * inst->density[offset |
| + i] + (float)1.0 / ((float)2.0 * WIDTH)) / (float)(inst->counter[s] + 1); |
| } |
| } // end loop over magnitude spectrum |
| |
| if (inst->counter[s] >= END_STARTUP_LONG) { |
| inst->counter[s] = 0; |
| if (inst->updates >= END_STARTUP_LONG) { |
| for (i = 0; i < inst->magnLen; i++) { |
| inst->quantile[i] = (float)exp(inst->lquantile[offset + i]); |
| } |
| } |
| } |
| |
| inst->counter[s]++; |
| } // end loop over simultaneous estimates |
| |
| // Sequentially update the noise during startup |
| if (inst->updates < END_STARTUP_LONG) { |
| // Use the last "s" to get noise during startup that differ from zero. |
| for (i = 0; i < inst->magnLen; i++) { |
| inst->quantile[i] = (float)exp(inst->lquantile[offset + i]); |
| } |
| } |
| |
| for (i = 0; i < inst->magnLen; i++) { |
| noise[i] = inst->quantile[i]; |
| } |
| } |
| |
| // Extract thresholds for feature parameters |
| // histograms are computed over some window_size (given by inst->modelUpdatePars[1]) |
| // thresholds and weights are extracted every window |
| // flag 0 means update histogram only, flag 1 means compute the thresholds/weights |
| // threshold and weights are returned in: inst->priorModelPars |
| void WebRtcNs_FeatureParameterExtraction(NSinst_t* inst, int flag) { |
| int i, useFeatureSpecFlat, useFeatureSpecDiff, numHistLrt; |
| int maxPeak1, maxPeak2; |
| int weightPeak1SpecFlat, weightPeak2SpecFlat, weightPeak1SpecDiff, weightPeak2SpecDiff; |
| |
| float binMid, featureSum; |
| float posPeak1SpecFlat, posPeak2SpecFlat, posPeak1SpecDiff, posPeak2SpecDiff; |
| float fluctLrt, avgHistLrt, avgSquareHistLrt, avgHistLrtCompl; |
| |
| //3 features: lrt, flatness, difference |
| //lrt_feature = inst->featureData[3]; |
| //flat_feature = inst->featureData[0]; |
| //diff_feature = inst->featureData[4]; |
| |
| //update histograms |
| if (flag == 0) { |
| // LRT |
| if ((inst->featureData[3] < HIST_PAR_EST * inst->featureExtractionParams.binSizeLrt) |
| && (inst->featureData[3] >= 0.0)) { |
| i = (int)(inst->featureData[3] / inst->featureExtractionParams.binSizeLrt); |
| inst->histLrt[i]++; |
| } |
| // Spectral flatness |
| if ((inst->featureData[0] < HIST_PAR_EST |
| * inst->featureExtractionParams.binSizeSpecFlat) |
| && (inst->featureData[0] >= 0.0)) { |
| i = (int)(inst->featureData[0] / inst->featureExtractionParams.binSizeSpecFlat); |
| inst->histSpecFlat[i]++; |
| } |
| // Spectral difference |
| if ((inst->featureData[4] < HIST_PAR_EST |
| * inst->featureExtractionParams.binSizeSpecDiff) |
| && (inst->featureData[4] >= 0.0)) { |
| i = (int)(inst->featureData[4] / inst->featureExtractionParams.binSizeSpecDiff); |
| inst->histSpecDiff[i]++; |
| } |
| } |
| |
| // extract parameters for speech/noise probability |
| if (flag == 1) { |
| //lrt feature: compute the average over inst->featureExtractionParams.rangeAvgHistLrt |
| avgHistLrt = 0.0; |
| avgHistLrtCompl = 0.0; |
| avgSquareHistLrt = 0.0; |
| numHistLrt = 0; |
| for (i = 0; i < HIST_PAR_EST; i++) { |
| binMid = ((float)i + (float)0.5) * inst->featureExtractionParams.binSizeLrt; |
| if (binMid <= inst->featureExtractionParams.rangeAvgHistLrt) { |
| avgHistLrt += inst->histLrt[i] * binMid; |
| numHistLrt += inst->histLrt[i]; |
| } |
| avgSquareHistLrt += inst->histLrt[i] * binMid * binMid; |
| avgHistLrtCompl += inst->histLrt[i] * binMid; |
| } |
| if (numHistLrt > 0) { |
| avgHistLrt = avgHistLrt / ((float)numHistLrt); |
| } |
| avgHistLrtCompl = avgHistLrtCompl / ((float)inst->modelUpdatePars[1]); |
| avgSquareHistLrt = avgSquareHistLrt / ((float)inst->modelUpdatePars[1]); |
| fluctLrt = avgSquareHistLrt - avgHistLrt * avgHistLrtCompl; |
| // get threshold for lrt feature: |
| if (fluctLrt < inst->featureExtractionParams.thresFluctLrt) { |
| //very low fluct, so likely noise |
| inst->priorModelPars[0] = inst->featureExtractionParams.maxLrt; |
| } else { |
| inst->priorModelPars[0] = inst->featureExtractionParams.factor1ModelPars |
| * avgHistLrt; |
| // check if value is within min/max range |
| if (inst->priorModelPars[0] < inst->featureExtractionParams.minLrt) { |
| inst->priorModelPars[0] = inst->featureExtractionParams.minLrt; |
| } |
| if (inst->priorModelPars[0] > inst->featureExtractionParams.maxLrt) { |
| inst->priorModelPars[0] = inst->featureExtractionParams.maxLrt; |
| } |
| } |
| // done with lrt feature |
| |
| // |
| // for spectral flatness and spectral difference: compute the main peaks of histogram |
| maxPeak1 = 0; |
| maxPeak2 = 0; |
| posPeak1SpecFlat = 0.0; |
| posPeak2SpecFlat = 0.0; |
| weightPeak1SpecFlat = 0; |
| weightPeak2SpecFlat = 0; |
| |
| // peaks for flatness |
| for (i = 0; i < HIST_PAR_EST; i++) { |
| binMid = ((float)i + (float)0.5) * inst->featureExtractionParams.binSizeSpecFlat; |
| if (inst->histSpecFlat[i] > maxPeak1) { |
| // Found new "first" peak |
| maxPeak2 = maxPeak1; |
| weightPeak2SpecFlat = weightPeak1SpecFlat; |
| posPeak2SpecFlat = posPeak1SpecFlat; |
| |
| maxPeak1 = inst->histSpecFlat[i]; |
| weightPeak1SpecFlat = inst->histSpecFlat[i]; |
| posPeak1SpecFlat = binMid; |
| } else if (inst->histSpecFlat[i] > maxPeak2) { |
| // Found new "second" peak |
| maxPeak2 = inst->histSpecFlat[i]; |
| weightPeak2SpecFlat = inst->histSpecFlat[i]; |
| posPeak2SpecFlat = binMid; |
| } |
| } |
| |
| //compute two peaks for spectral difference |
| maxPeak1 = 0; |
| maxPeak2 = 0; |
| posPeak1SpecDiff = 0.0; |
| posPeak2SpecDiff = 0.0; |
| weightPeak1SpecDiff = 0; |
| weightPeak2SpecDiff = 0; |
| // peaks for spectral difference |
| for (i = 0; i < HIST_PAR_EST; i++) { |
| binMid = ((float)i + (float)0.5) * inst->featureExtractionParams.binSizeSpecDiff; |
| if (inst->histSpecDiff[i] > maxPeak1) { |
| // Found new "first" peak |
| maxPeak2 = maxPeak1; |
| weightPeak2SpecDiff = weightPeak1SpecDiff; |
| posPeak2SpecDiff = posPeak1SpecDiff; |
| |
| maxPeak1 = inst->histSpecDiff[i]; |
| weightPeak1SpecDiff = inst->histSpecDiff[i]; |
| posPeak1SpecDiff = binMid; |
| } else if (inst->histSpecDiff[i] > maxPeak2) { |
| // Found new "second" peak |
| maxPeak2 = inst->histSpecDiff[i]; |
| weightPeak2SpecDiff = inst->histSpecDiff[i]; |
| posPeak2SpecDiff = binMid; |
| } |
| } |
| |
| // for spectrum flatness feature |
| useFeatureSpecFlat = 1; |
| // merge the two peaks if they are close |
| if ((fabs(posPeak2SpecFlat - posPeak1SpecFlat) |
| < inst->featureExtractionParams.limitPeakSpacingSpecFlat) |
| && (weightPeak2SpecFlat |
| > inst->featureExtractionParams.limitPeakWeightsSpecFlat |
| * weightPeak1SpecFlat)) { |
| weightPeak1SpecFlat += weightPeak2SpecFlat; |
| posPeak1SpecFlat = (float)0.5 * (posPeak1SpecFlat + posPeak2SpecFlat); |
| } |
| //reject if weight of peaks is not large enough, or peak value too small |
| if (weightPeak1SpecFlat < inst->featureExtractionParams.thresWeightSpecFlat |
| || posPeak1SpecFlat < inst->featureExtractionParams.thresPosSpecFlat) { |
| useFeatureSpecFlat = 0; |
| } |
| // if selected, get the threshold |
| if (useFeatureSpecFlat == 1) { |
| // compute the threshold |
| inst->priorModelPars[1] = inst->featureExtractionParams.factor2ModelPars |
| * posPeak1SpecFlat; |
| //check if value is within min/max range |
| if (inst->priorModelPars[1] < inst->featureExtractionParams.minSpecFlat) { |
| inst->priorModelPars[1] = inst->featureExtractionParams.minSpecFlat; |
| } |
| if (inst->priorModelPars[1] > inst->featureExtractionParams.maxSpecFlat) { |
| inst->priorModelPars[1] = inst->featureExtractionParams.maxSpecFlat; |
| } |
| } |
| // done with flatness feature |
| |
| // for template feature |
| useFeatureSpecDiff = 1; |
| // merge the two peaks if they are close |
| if ((fabs(posPeak2SpecDiff - posPeak1SpecDiff) |
| < inst->featureExtractionParams.limitPeakSpacingSpecDiff) |
| && (weightPeak2SpecDiff |
| > inst->featureExtractionParams.limitPeakWeightsSpecDiff |
| * weightPeak1SpecDiff)) { |
| weightPeak1SpecDiff += weightPeak2SpecDiff; |
| posPeak1SpecDiff = (float)0.5 * (posPeak1SpecDiff + posPeak2SpecDiff); |
| } |
| // get the threshold value |
| inst->priorModelPars[3] = inst->featureExtractionParams.factor1ModelPars |
| * posPeak1SpecDiff; |
| //reject if weight of peaks is not large enough |
| if (weightPeak1SpecDiff < inst->featureExtractionParams.thresWeightSpecDiff) { |
| useFeatureSpecDiff = 0; |
| } |
| //check if value is within min/max range |
| if (inst->priorModelPars[3] < inst->featureExtractionParams.minSpecDiff) { |
| inst->priorModelPars[3] = inst->featureExtractionParams.minSpecDiff; |
| } |
| if (inst->priorModelPars[3] > inst->featureExtractionParams.maxSpecDiff) { |
| inst->priorModelPars[3] = inst->featureExtractionParams.maxSpecDiff; |
| } |
| // done with spectral difference feature |
| |
| // don't use template feature if fluctuation of lrt feature is very low: |
| // most likely just noise state |
| if (fluctLrt < inst->featureExtractionParams.thresFluctLrt) { |
| useFeatureSpecDiff = 0; |
| } |
| |
| // select the weights between the features |
| // inst->priorModelPars[4] is weight for lrt: always selected |
| // inst->priorModelPars[5] is weight for spectral flatness |
| // inst->priorModelPars[6] is weight for spectral difference |
| featureSum = (float)(1 + useFeatureSpecFlat + useFeatureSpecDiff); |
| inst->priorModelPars[4] = (float)1.0 / featureSum; |
| inst->priorModelPars[5] = ((float)useFeatureSpecFlat) / featureSum; |
| inst->priorModelPars[6] = ((float)useFeatureSpecDiff) / featureSum; |
| |
| // set hists to zero for next update |
| if (inst->modelUpdatePars[0] >= 1) { |
| for (i = 0; i < HIST_PAR_EST; i++) { |
| inst->histLrt[i] = 0; |
| inst->histSpecFlat[i] = 0; |
| inst->histSpecDiff[i] = 0; |
| } |
| } |
| } // end of flag == 1 |
| } |
| |
| // Compute spectral flatness on input spectrum |
| // magnIn is the magnitude spectrum |
| // spectral flatness is returned in inst->featureData[0] |
| void WebRtcNs_ComputeSpectralFlatness(NSinst_t* inst, float* magnIn) { |
| int i; |
| int shiftLP = 1; //option to remove first bin(s) from spectral measures |
| float avgSpectralFlatnessNum, avgSpectralFlatnessDen, spectralTmp; |
| |
| // comute spectral measures |
| // for flatness |
| avgSpectralFlatnessNum = 0.0; |
| avgSpectralFlatnessDen = inst->sumMagn; |
| for (i = 0; i < shiftLP; i++) { |
| avgSpectralFlatnessDen -= magnIn[i]; |
| } |
| // compute log of ratio of the geometric to arithmetic mean: check for log(0) case |
| for (i = shiftLP; i < inst->magnLen; i++) { |
| if (magnIn[i] > 0.0) { |
| avgSpectralFlatnessNum += (float)log(magnIn[i]); |
| } else { |
| inst->featureData[0] -= SPECT_FL_TAVG * inst->featureData[0]; |
| return; |
| } |
| } |
| //normalize |
| avgSpectralFlatnessDen = avgSpectralFlatnessDen / inst->magnLen; |
| avgSpectralFlatnessNum = avgSpectralFlatnessNum / inst->magnLen; |
| |
| //ratio and inverse log: check for case of log(0) |
| spectralTmp = (float)exp(avgSpectralFlatnessNum) / avgSpectralFlatnessDen; |
| |
| //time-avg update of spectral flatness feature |
| inst->featureData[0] += SPECT_FL_TAVG * (spectralTmp - inst->featureData[0]); |
| // done with flatness feature |
| } |
| |
| // Compute the difference measure between input spectrum and a template/learned noise spectrum |
| // magnIn is the input spectrum |
| // the reference/template spectrum is inst->magnAvgPause[i] |
| // returns (normalized) spectral difference in inst->featureData[4] |
| void WebRtcNs_ComputeSpectralDifference(NSinst_t* inst, float* magnIn) { |
| // avgDiffNormMagn = var(magnIn) - cov(magnIn, magnAvgPause)^2 / var(magnAvgPause) |
| int i; |
| float avgPause, avgMagn, covMagnPause, varPause, varMagn, avgDiffNormMagn; |
| |
| avgPause = 0.0; |
| avgMagn = inst->sumMagn; |
| // compute average quantities |
| for (i = 0; i < inst->magnLen; i++) { |
| //conservative smooth noise spectrum from pause frames |
| avgPause += inst->magnAvgPause[i]; |
| } |
| avgPause = avgPause / ((float)inst->magnLen); |
| avgMagn = avgMagn / ((float)inst->magnLen); |
| |
| covMagnPause = 0.0; |
| varPause = 0.0; |
| varMagn = 0.0; |
| // compute variance and covariance quantities |
| for (i = 0; i < inst->magnLen; i++) { |
| covMagnPause += (magnIn[i] - avgMagn) * (inst->magnAvgPause[i] - avgPause); |
| varPause += (inst->magnAvgPause[i] - avgPause) * (inst->magnAvgPause[i] - avgPause); |
| varMagn += (magnIn[i] - avgMagn) * (magnIn[i] - avgMagn); |
| } |
| covMagnPause = covMagnPause / ((float)inst->magnLen); |
| varPause = varPause / ((float)inst->magnLen); |
| varMagn = varMagn / ((float)inst->magnLen); |
| // update of average magnitude spectrum |
| inst->featureData[6] += inst->signalEnergy; |
| |
| avgDiffNormMagn = varMagn - (covMagnPause * covMagnPause) / (varPause + (float)0.0001); |
| // normalize and compute time-avg update of difference feature |
| avgDiffNormMagn = (float)(avgDiffNormMagn / (inst->featureData[5] + (float)0.0001)); |
| inst->featureData[4] += SPECT_DIFF_TAVG * (avgDiffNormMagn - inst->featureData[4]); |
| } |
| |
| // Compute speech/noise probability |
| // speech/noise probability is returned in: probSpeechFinal |
| //magn is the input magnitude spectrum |
| //noise is the noise spectrum |
| //snrLocPrior is the prior snr for each freq. |
| //snr loc_post is the post snr for each freq. |
| void WebRtcNs_SpeechNoiseProb(NSinst_t* inst, float* probSpeechFinal, float* snrLocPrior, |
| float* snrLocPost) { |
| int i, sgnMap; |
| float invLrt, gainPrior, indPrior; |
| float logLrtTimeAvgKsum, besselTmp; |
| float indicator0, indicator1, indicator2; |
| float tmpFloat1, tmpFloat2; |
| float weightIndPrior0, weightIndPrior1, weightIndPrior2; |
| float threshPrior0, threshPrior1, threshPrior2; |
| float widthPrior, widthPrior0, widthPrior1, widthPrior2; |
| |
| widthPrior0 = WIDTH_PR_MAP; |
| widthPrior1 = (float)2.0 * WIDTH_PR_MAP; //width for pause region: |
| // lower range, so increase width in tanh map |
| widthPrior2 = (float)2.0 * WIDTH_PR_MAP; //for spectral-difference measure |
| |
| //threshold parameters for features |
| threshPrior0 = inst->priorModelPars[0]; |
| threshPrior1 = inst->priorModelPars[1]; |
| threshPrior2 = inst->priorModelPars[3]; |
| |
| //sign for flatness feature |
| sgnMap = (int)(inst->priorModelPars[2]); |
| |
| //weight parameters for features |
| weightIndPrior0 = inst->priorModelPars[4]; |
| weightIndPrior1 = inst->priorModelPars[5]; |
| weightIndPrior2 = inst->priorModelPars[6]; |
| |
| // compute feature based on average LR factor |
| // this is the average over all frequencies of the smooth log lrt |
| logLrtTimeAvgKsum = 0.0; |
| for (i = 0; i < inst->magnLen; i++) { |
| tmpFloat1 = (float)1.0 + (float)2.0 * snrLocPrior[i]; |
| tmpFloat2 = (float)2.0 * snrLocPrior[i] / (tmpFloat1 + (float)0.0001); |
| besselTmp = (snrLocPost[i] + (float)1.0) * tmpFloat2; |
| inst->logLrtTimeAvg[i] += LRT_TAVG * (besselTmp - (float)log(tmpFloat1) |
| - inst->logLrtTimeAvg[i]); |
| logLrtTimeAvgKsum += inst->logLrtTimeAvg[i]; |
| } |
| logLrtTimeAvgKsum = (float)logLrtTimeAvgKsum / (inst->magnLen); |
| inst->featureData[3] = logLrtTimeAvgKsum; |
| // done with computation of LR factor |
| |
| // |
| //compute the indicator functions |
| // |
| |
| // average lrt feature |
| widthPrior = widthPrior0; |
| //use larger width in tanh map for pause regions |
| if (logLrtTimeAvgKsum < threshPrior0) { |
| widthPrior = widthPrior1; |
| } |
| // compute indicator function: sigmoid map |
| indicator0 = (float)0.5 * ((float)tanh(widthPrior * |
| (logLrtTimeAvgKsum - threshPrior0)) + (float)1.0); |
| |
| //spectral flatness feature |
| tmpFloat1 = inst->featureData[0]; |
| widthPrior = widthPrior0; |
| //use larger width in tanh map for pause regions |
| if (sgnMap == 1 && (tmpFloat1 > threshPrior1)) { |
| widthPrior = widthPrior1; |
| } |
| if (sgnMap == -1 && (tmpFloat1 < threshPrior1)) { |
| widthPrior = widthPrior1; |
| } |
| // compute indicator function: sigmoid map |
| indicator1 = (float)0.5 * ((float)tanh((float)sgnMap * |
| widthPrior * (threshPrior1 - tmpFloat1)) + (float)1.0); |
| |
| //for template spectrum-difference |
| tmpFloat1 = inst->featureData[4]; |
| widthPrior = widthPrior0; |
| //use larger width in tanh map for pause regions |
| if (tmpFloat1 < threshPrior2) { |
| widthPrior = widthPrior2; |
| } |
| // compute indicator function: sigmoid map |
| indicator2 = (float)0.5 * ((float)tanh(widthPrior * (tmpFloat1 - threshPrior2)) |
| + (float)1.0); |
| |
| //combine the indicator function with the feature weights |
| indPrior = weightIndPrior0 * indicator0 + weightIndPrior1 * indicator1 + weightIndPrior2 |
| * indicator2; |
| // done with computing indicator function |
| |
| //compute the prior probability |
| inst->priorSpeechProb += PRIOR_UPDATE * (indPrior - inst->priorSpeechProb); |
| // make sure probabilities are within range: keep floor to 0.01 |
| if (inst->priorSpeechProb > 1.0) { |
| inst->priorSpeechProb = (float)1.0; |
| } |
| if (inst->priorSpeechProb < 0.01) { |
| inst->priorSpeechProb = (float)0.01; |
| } |
| |
| //final speech probability: combine prior model with LR factor: |
| gainPrior = ((float)1.0 - inst->priorSpeechProb) / (inst->priorSpeechProb + (float)0.0001); |
| for (i = 0; i < inst->magnLen; i++) { |
| invLrt = (float)exp(-inst->logLrtTimeAvg[i]); |
| invLrt = (float)gainPrior * invLrt; |
| probSpeechFinal[i] = (float)1.0 / ((float)1.0 + invLrt); |
| } |
| } |
| |
| int WebRtcNs_ProcessCore(NSinst_t* inst, |
| short* speechFrame, |
| short* speechFrameHB, |
| short* outFrame, |
| short* outFrameHB) { |
| // main routine for noise reduction |
| |
| int flagHB = 0; |
| int i; |
| const int kStartBand = 5; // Skip first frequency bins during estimation. |
| int updateParsFlag; |
| |
| float energy1, energy2, gain, factor, factor1, factor2; |
| float signalEnergy, sumMagn; |
| float snrPrior, currentEstimateStsa; |
| float tmpFloat1, tmpFloat2, tmpFloat3, probSpeech, probNonSpeech; |
| float gammaNoiseTmp, gammaNoiseOld; |
| float noiseUpdateTmp, fTmp, dTmp; |
| float fin[BLOCKL_MAX], fout[BLOCKL_MAX]; |
| float winData[ANAL_BLOCKL_MAX]; |
| float magn[HALF_ANAL_BLOCKL], noise[HALF_ANAL_BLOCKL]; |
| float theFilter[HALF_ANAL_BLOCKL], theFilterTmp[HALF_ANAL_BLOCKL]; |
| float snrLocPost[HALF_ANAL_BLOCKL], snrLocPrior[HALF_ANAL_BLOCKL]; |
| float probSpeechFinal[HALF_ANAL_BLOCKL], previousEstimateStsa[HALF_ANAL_BLOCKL]; |
| float real[ANAL_BLOCKL_MAX], imag[HALF_ANAL_BLOCKL]; |
| // Variables during startup |
| float sum_log_i = 0.0; |
| float sum_log_i_square = 0.0; |
| float sum_log_magn = 0.0; |
| float sum_log_i_log_magn = 0.0; |
| float parametric_noise = 0.0; |
| float parametric_exp = 0.0; |
| float parametric_num = 0.0; |
| |
| // SWB variables |
| int deltaBweHB = 1; |
| int deltaGainHB = 1; |
| float decayBweHB = 1.0; |
| float gainMapParHB = 1.0; |
| float gainTimeDomainHB = 1.0; |
| float avgProbSpeechHB, avgProbSpeechHBTmp, avgFilterGainHB, gainModHB; |
| |
| // Check that initiation has been done |
| if (inst->initFlag != 1) { |
| return (-1); |
| } |
| // Check for valid pointers based on sampling rate |
| if (inst->fs == 32000) { |
| if (speechFrameHB == NULL) { |
| return -1; |
| } |
| flagHB = 1; |
| // range for averaging low band quantities for H band gain |
| deltaBweHB = (int)inst->magnLen / 4; |
| deltaGainHB = deltaBweHB; |
| } |
| // |
| updateParsFlag = inst->modelUpdatePars[0]; |
| // |
| |
| //for LB do all processing |
| // convert to float |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| fin[i] = (float)speechFrame[i]; |
| } |
| // update analysis buffer for L band |
| memcpy(inst->dataBuf, inst->dataBuf + inst->blockLen10ms, |
| sizeof(float) * (inst->anaLen - inst->blockLen10ms)); |
| memcpy(inst->dataBuf + inst->anaLen - inst->blockLen10ms, fin, |
| sizeof(float) * inst->blockLen10ms); |
| |
| if (flagHB == 1) { |
| // convert to float |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| fin[i] = (float)speechFrameHB[i]; |
| } |
| // update analysis buffer for H band |
| memcpy(inst->dataBufHB, inst->dataBufHB + inst->blockLen10ms, |
| sizeof(float) * (inst->anaLen - inst->blockLen10ms)); |
| memcpy(inst->dataBufHB + inst->anaLen - inst->blockLen10ms, fin, |
| sizeof(float) * inst->blockLen10ms); |
| } |
| |
| // check if processing needed |
| if (inst->outLen == 0) { |
| // windowing |
| energy1 = 0.0; |
| for (i = 0; i < inst->anaLen; i++) { |
| winData[i] = inst->window[i] * inst->dataBuf[i]; |
| energy1 += winData[i] * winData[i]; |
| } |
| if (energy1 == 0.0) { |
| // synthesize the special case of zero input |
| // we want to avoid updating statistics in this case: |
| // Updating feature statistics when we have zeros only will cause thresholds to |
| // move towards zero signal situations. This in turn has the effect that once the |
| // signal is "turned on" (non-zero values) everything will be treated as speech |
| // and there is no noise suppression effect. Depending on the duration of the |
| // inactive signal it takes a considerable amount of time for the system to learn |
| // what is noise and what is speech. |
| |
| // read out fully processed segment |
| for (i = inst->windShift; i < inst->blockLen + inst->windShift; i++) { |
| fout[i - inst->windShift] = inst->syntBuf[i]; |
| } |
| // update synthesis buffer |
| memcpy(inst->syntBuf, inst->syntBuf + inst->blockLen, |
| sizeof(float) * (inst->anaLen - inst->blockLen)); |
| memset(inst->syntBuf + inst->anaLen - inst->blockLen, 0, |
| sizeof(float) * inst->blockLen); |
| |
| // out buffer |
| inst->outLen = inst->blockLen - inst->blockLen10ms; |
| if (inst->blockLen > inst->blockLen10ms) { |
| for (i = 0; i < inst->outLen; i++) { |
| inst->outBuf[i] = fout[i + inst->blockLen10ms]; |
| } |
| } |
| // convert to short |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| dTmp = fout[i]; |
| if (dTmp < WEBRTC_SPL_WORD16_MIN) { |
| dTmp = WEBRTC_SPL_WORD16_MIN; |
| } else if (dTmp > WEBRTC_SPL_WORD16_MAX) { |
| dTmp = WEBRTC_SPL_WORD16_MAX; |
| } |
| outFrame[i] = (short)dTmp; |
| } |
| |
| // for time-domain gain of HB |
| if (flagHB == 1) { |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| dTmp = inst->dataBufHB[i]; |
| if (dTmp < WEBRTC_SPL_WORD16_MIN) { |
| dTmp = WEBRTC_SPL_WORD16_MIN; |
| } else if (dTmp > WEBRTC_SPL_WORD16_MAX) { |
| dTmp = WEBRTC_SPL_WORD16_MAX; |
| } |
| outFrameHB[i] = (short)dTmp; |
| } |
| } // end of H band gain computation |
| // |
| return 0; |
| } |
| |
| // |
| inst->blockInd++; // Update the block index only when we process a block. |
| // FFT |
| WebRtc_rdft(inst->anaLen, 1, winData, inst->ip, inst->wfft); |
| |
| imag[0] = 0; |
| real[0] = winData[0]; |
| magn[0] = (float)(fabs(real[0]) + 1.0f); |
| imag[inst->magnLen - 1] = 0; |
| real[inst->magnLen - 1] = winData[1]; |
| magn[inst->magnLen - 1] = (float)(fabs(real[inst->magnLen - 1]) + 1.0f); |
| signalEnergy = (float)(real[0] * real[0]) + |
| (float)(real[inst->magnLen - 1] * real[inst->magnLen - 1]); |
| sumMagn = magn[0] + magn[inst->magnLen - 1]; |
| if (inst->blockInd < END_STARTUP_SHORT) { |
| inst->initMagnEst[0] += magn[0]; |
| inst->initMagnEst[inst->magnLen - 1] += magn[inst->magnLen - 1]; |
| tmpFloat2 = log((float)(inst->magnLen - 1)); |
| sum_log_i = tmpFloat2; |
| sum_log_i_square = tmpFloat2 * tmpFloat2; |
| tmpFloat1 = log(magn[inst->magnLen - 1]); |
| sum_log_magn = tmpFloat1; |
| sum_log_i_log_magn = tmpFloat2 * tmpFloat1; |
| } |
| for (i = 1; i < inst->magnLen - 1; i++) { |
| real[i] = winData[2 * i]; |
| imag[i] = winData[2 * i + 1]; |
| // magnitude spectrum |
| fTmp = real[i] * real[i]; |
| fTmp += imag[i] * imag[i]; |
| signalEnergy += fTmp; |
| magn[i] = ((float)sqrt(fTmp)) + 1.0f; |
| sumMagn += magn[i]; |
| if (inst->blockInd < END_STARTUP_SHORT) { |
| inst->initMagnEst[i] += magn[i]; |
| if (i >= kStartBand) { |
| tmpFloat2 = log((float)i); |
| sum_log_i += tmpFloat2; |
| sum_log_i_square += tmpFloat2 * tmpFloat2; |
| tmpFloat1 = log(magn[i]); |
| sum_log_magn += tmpFloat1; |
| sum_log_i_log_magn += tmpFloat2 * tmpFloat1; |
| } |
| } |
| } |
| signalEnergy = signalEnergy / ((float)inst->magnLen); |
| inst->signalEnergy = signalEnergy; |
| inst->sumMagn = sumMagn; |
| |
| //compute spectral flatness on input spectrum |
| WebRtcNs_ComputeSpectralFlatness(inst, magn); |
| // quantile noise estimate |
| WebRtcNs_NoiseEstimation(inst, magn, noise); |
| //compute simplified noise model during startup |
| if (inst->blockInd < END_STARTUP_SHORT) { |
| // Estimate White noise |
| inst->whiteNoiseLevel += sumMagn / ((float)inst->magnLen) * inst->overdrive; |
| // Estimate Pink noise parameters |
| tmpFloat1 = sum_log_i_square * ((float)(inst->magnLen - kStartBand)); |
| tmpFloat1 -= (sum_log_i * sum_log_i); |
| tmpFloat2 = (sum_log_i_square * sum_log_magn - sum_log_i * sum_log_i_log_magn); |
| tmpFloat3 = tmpFloat2 / tmpFloat1; |
| // Constrain the estimated spectrum to be positive |
| if (tmpFloat3 < 0.0f) { |
| tmpFloat3 = 0.0f; |
| } |
| inst->pinkNoiseNumerator += tmpFloat3; |
| tmpFloat2 = (sum_log_i * sum_log_magn); |
| tmpFloat2 -= ((float)(inst->magnLen - kStartBand)) * sum_log_i_log_magn; |
| tmpFloat3 = tmpFloat2 / tmpFloat1; |
| // Constrain the pink noise power to be in the interval [0, 1]; |
| if (tmpFloat3 < 0.0f) { |
| tmpFloat3 = 0.0f; |
| } |
| if (tmpFloat3 > 1.0f) { |
| tmpFloat3 = 1.0f; |
| } |
| inst->pinkNoiseExp += tmpFloat3; |
| |
| // Calculate frequency independent parts of parametric noise estimate. |
| if (inst->pinkNoiseExp == 0.0f) { |
| // Use white noise estimate |
| parametric_noise = inst->whiteNoiseLevel; |
| } else { |
| // Use pink noise estimate |
| parametric_num = exp(inst->pinkNoiseNumerator / (float)(inst->blockInd + 1)); |
| parametric_num *= (float)(inst->blockInd + 1); |
| parametric_exp = inst->pinkNoiseExp / (float)(inst->blockInd + 1); |
| parametric_noise = parametric_num / pow((float)kStartBand, parametric_exp); |
| } |
| for (i = 0; i < inst->magnLen; i++) { |
| // Estimate the background noise using the white and pink noise parameters |
| if ((inst->pinkNoiseExp > 0.0f) && (i >= kStartBand)) { |
| // Use pink noise estimate |
| parametric_noise = parametric_num / pow((float)i, parametric_exp); |
| } |
| theFilterTmp[i] = (inst->initMagnEst[i] - inst->overdrive * parametric_noise); |
| theFilterTmp[i] /= (inst->initMagnEst[i] + (float)0.0001); |
| // Weight quantile noise with modeled noise |
| noise[i] *= (inst->blockInd); |
| tmpFloat2 = parametric_noise * (END_STARTUP_SHORT - inst->blockInd); |
| noise[i] += (tmpFloat2 / (float)(inst->blockInd + 1)); |
| noise[i] /= END_STARTUP_SHORT; |
| } |
| } |
| //compute average signal during END_STARTUP_LONG time: |
| // used to normalize spectral difference measure |
| if (inst->blockInd < END_STARTUP_LONG) { |
| inst->featureData[5] *= inst->blockInd; |
| inst->featureData[5] += signalEnergy; |
| inst->featureData[5] /= (inst->blockInd + 1); |
| } |
| |
| #ifdef PROCESS_FLOW_0 |
| if (inst->blockInd > END_STARTUP_LONG) { |
| //option: average the quantile noise: for check with AEC2 |
| for (i = 0; i < inst->magnLen; i++) { |
| noise[i] = (float)0.6 * inst->noisePrev[i] + (float)0.4 * noise[i]; |
| } |
| for (i = 0; i < inst->magnLen; i++) { |
| // Wiener with over sub-substraction: |
| theFilter[i] = (magn[i] - inst->overdrive * noise[i]) / (magn[i] + (float)0.0001); |
| } |
| } |
| #else |
| //start processing at frames == converged+1 |
| // |
| // STEP 1: compute prior and post snr based on quantile noise est |
| // |
| |
| // compute DD estimate of prior SNR: needed for new method |
| for (i = 0; i < inst->magnLen; i++) { |
| // post snr |
| snrLocPost[i] = (float)0.0; |
| if (magn[i] > noise[i]) { |
| snrLocPost[i] = magn[i] / (noise[i] + (float)0.0001) - (float)1.0; |
| } |
| // previous post snr |
| // previous estimate: based on previous frame with gain filter |
| previousEstimateStsa[i] = inst->magnPrev[i] / (inst->noisePrev[i] + (float)0.0001) |
| * (inst->smooth[i]); |
| // DD estimate is sum of two terms: current estimate and previous estimate |
| // directed decision update of snrPrior |
| snrLocPrior[i] = DD_PR_SNR * previousEstimateStsa[i] + ((float)1.0 - DD_PR_SNR) |
| * snrLocPost[i]; |
| // post and prior snr needed for step 2 |
| } // end of loop over freqs |
| #ifdef PROCESS_FLOW_1 |
| for (i = 0; i < inst->magnLen; i++) { |
| // gain filter |
| tmpFloat1 = inst->overdrive + snrLocPrior[i]; |
| tmpFloat2 = (float)snrLocPrior[i] / tmpFloat1; |
| theFilter[i] = (float)tmpFloat2; |
| } // end of loop over freqs |
| #endif |
| // done with step 1: dd computation of prior and post snr |
| |
| // |
| //STEP 2: compute speech/noise likelihood |
| // |
| #ifdef PROCESS_FLOW_2 |
| // compute difference of input spectrum with learned/estimated noise spectrum |
| WebRtcNs_ComputeSpectralDifference(inst, magn); |
| // compute histograms for parameter decisions (thresholds and weights for features) |
| // parameters are extracted once every window time (=inst->modelUpdatePars[1]) |
| if (updateParsFlag >= 1) { |
| // counter update |
| inst->modelUpdatePars[3]--; |
| // update histogram |
| if (inst->modelUpdatePars[3] > 0) { |
| WebRtcNs_FeatureParameterExtraction(inst, 0); |
| } |
| // compute model parameters |
| if (inst->modelUpdatePars[3] == 0) { |
| WebRtcNs_FeatureParameterExtraction(inst, 1); |
| inst->modelUpdatePars[3] = inst->modelUpdatePars[1]; |
| // if wish to update only once, set flag to zero |
| if (updateParsFlag == 1) { |
| inst->modelUpdatePars[0] = 0; |
| } else { |
| // update every window: |
| // get normalization for spectral difference for next window estimate |
| inst->featureData[6] = inst->featureData[6] |
| / ((float)inst->modelUpdatePars[1]); |
| inst->featureData[5] = (float)0.5 * (inst->featureData[6] |
| + inst->featureData[5]); |
| inst->featureData[6] = (float)0.0; |
| } |
| } |
| } |
| // compute speech/noise probability |
| WebRtcNs_SpeechNoiseProb(inst, probSpeechFinal, snrLocPrior, snrLocPost); |
| // time-avg parameter for noise update |
| gammaNoiseTmp = NOISE_UPDATE; |
| for (i = 0; i < inst->magnLen; i++) { |
| probSpeech = probSpeechFinal[i]; |
| probNonSpeech = (float)1.0 - probSpeech; |
| // temporary noise update: |
| // use it for speech frames if update value is less than previous |
| noiseUpdateTmp = gammaNoiseTmp * inst->noisePrev[i] + ((float)1.0 - gammaNoiseTmp) |
| * (probNonSpeech * magn[i] + probSpeech * inst->noisePrev[i]); |
| // |
| // time-constant based on speech/noise state |
| gammaNoiseOld = gammaNoiseTmp; |
| gammaNoiseTmp = NOISE_UPDATE; |
| // increase gamma (i.e., less noise update) for frame likely to be speech |
| if (probSpeech > PROB_RANGE) { |
| gammaNoiseTmp = SPEECH_UPDATE; |
| } |
| // conservative noise update |
| if (probSpeech < PROB_RANGE) { |
| inst->magnAvgPause[i] += GAMMA_PAUSE * (magn[i] - inst->magnAvgPause[i]); |
| } |
| // noise update |
| if (gammaNoiseTmp == gammaNoiseOld) { |
| noise[i] = noiseUpdateTmp; |
| } else { |
| noise[i] = gammaNoiseTmp * inst->noisePrev[i] + ((float)1.0 - gammaNoiseTmp) |
| * (probNonSpeech * magn[i] + probSpeech * inst->noisePrev[i]); |
| // allow for noise update downwards: |
| // if noise update decreases the noise, it is safe, so allow it to happen |
| if (noiseUpdateTmp < noise[i]) { |
| noise[i] = noiseUpdateTmp; |
| } |
| } |
| } // end of freq loop |
| // done with step 2: noise update |
| |
| // |
| // STEP 3: compute dd update of prior snr and post snr based on new noise estimate |
| // |
| for (i = 0; i < inst->magnLen; i++) { |
| // post and prior snr |
| currentEstimateStsa = (float)0.0; |
| if (magn[i] > noise[i]) { |
| currentEstimateStsa = magn[i] / (noise[i] + (float)0.0001) - (float)1.0; |
| } |
| // DD estimate is sume of two terms: current estimate and previous estimate |
| // directed decision update of snrPrior |
| snrPrior = DD_PR_SNR * previousEstimateStsa[i] + ((float)1.0 - DD_PR_SNR) |
| * currentEstimateStsa; |
| // gain filter |
| tmpFloat1 = inst->overdrive + snrPrior; |
| tmpFloat2 = (float)snrPrior / tmpFloat1; |
| theFilter[i] = (float)tmpFloat2; |
| } // end of loop over freqs |
| // done with step3 |
| #endif |
| #endif |
| |
| for (i = 0; i < inst->magnLen; i++) { |
| // flooring bottom |
| if (theFilter[i] < inst->denoiseBound) { |
| theFilter[i] = inst->denoiseBound; |
| } |
| // flooring top |
| if (theFilter[i] > (float)1.0) { |
| theFilter[i] = 1.0; |
| } |
| if (inst->blockInd < END_STARTUP_SHORT) { |
| // flooring bottom |
| if (theFilterTmp[i] < inst->denoiseBound) { |
| theFilterTmp[i] = inst->denoiseBound; |
| } |
| // flooring top |
| if (theFilterTmp[i] > (float)1.0) { |
| theFilterTmp[i] = 1.0; |
| } |
| // Weight the two suppression filters |
| theFilter[i] *= (inst->blockInd); |
| theFilterTmp[i] *= (END_STARTUP_SHORT - inst->blockInd); |
| theFilter[i] += theFilterTmp[i]; |
| theFilter[i] /= (END_STARTUP_SHORT); |
| } |
| // smoothing |
| #ifdef PROCESS_FLOW_0 |
| inst->smooth[i] *= SMOOTH; // value set to 0.7 in define.h file |
| inst->smooth[i] += ((float)1.0 - SMOOTH) * theFilter[i]; |
| #else |
| inst->smooth[i] = theFilter[i]; |
| #endif |
| real[i] *= inst->smooth[i]; |
| imag[i] *= inst->smooth[i]; |
| } |
| // keep track of noise and magn spectrum for next frame |
| for (i = 0; i < inst->magnLen; i++) { |
| inst->noisePrev[i] = noise[i]; |
| inst->magnPrev[i] = magn[i]; |
| } |
| // back to time domain |
| winData[0] = real[0]; |
| winData[1] = real[inst->magnLen - 1]; |
| for (i = 1; i < inst->magnLen - 1; i++) { |
| winData[2 * i] = real[i]; |
| winData[2 * i + 1] = imag[i]; |
| } |
| WebRtc_rdft(inst->anaLen, -1, winData, inst->ip, inst->wfft); |
| |
| for (i = 0; i < inst->anaLen; i++) { |
| real[i] = 2.0f * winData[i] / inst->anaLen; // fft scaling |
| } |
| |
| //scale factor: only do it after END_STARTUP_LONG time |
| factor = (float)1.0; |
| if (inst->gainmap == 1 && inst->blockInd > END_STARTUP_LONG) { |
| factor1 = (float)1.0; |
| factor2 = (float)1.0; |
| |
| energy2 = 0.0; |
| for (i = 0; i < inst->anaLen; i++) { |
| energy2 += (float)real[i] * (float)real[i]; |
| } |
| gain = (float)sqrt(energy2 / (energy1 + (float)1.0)); |
| |
| #ifdef PROCESS_FLOW_2 |
| // scaling for new version |
| if (gain > B_LIM) { |
| factor1 = (float)1.0 + (float)1.3 * (gain - B_LIM); |
| if (gain * factor1 > (float)1.0) { |
| factor1 = (float)1.0 / gain; |
| } |
| } |
| if (gain < B_LIM) { |
| //don't reduce scale too much for pause regions: |
| // attenuation here should be controlled by flooring |
| if (gain <= inst->denoiseBound) { |
| gain = inst->denoiseBound; |
| } |
| factor2 = (float)1.0 - (float)0.3 * (B_LIM - gain); |
| } |
| //combine both scales with speech/noise prob: |
| // note prior (priorSpeechProb) is not frequency dependent |
| factor = inst->priorSpeechProb * factor1 + ((float)1.0 - inst->priorSpeechProb) |
| * factor2; |
| #else |
| if (gain > B_LIM) { |
| factor = (float)1.0 + (float)1.3 * (gain - B_LIM); |
| } else { |
| factor = (float)1.0 + (float)2.0 * (gain - B_LIM); |
| } |
| if (gain * factor > (float)1.0) { |
| factor = (float)1.0 / gain; |
| } |
| #endif |
| } // out of inst->gainmap==1 |
| |
| // synthesis |
| for (i = 0; i < inst->anaLen; i++) { |
| inst->syntBuf[i] += factor * inst->window[i] * (float)real[i]; |
| } |
| // read out fully processed segment |
| for (i = inst->windShift; i < inst->blockLen + inst->windShift; i++) { |
| fout[i - inst->windShift] = inst->syntBuf[i]; |
| } |
| // update synthesis buffer |
| memcpy(inst->syntBuf, inst->syntBuf + inst->blockLen, |
| sizeof(float) * (inst->anaLen - inst->blockLen)); |
| memset(inst->syntBuf + inst->anaLen - inst->blockLen, 0, |
| sizeof(float) * inst->blockLen); |
| |
| // out buffer |
| inst->outLen = inst->blockLen - inst->blockLen10ms; |
| if (inst->blockLen > inst->blockLen10ms) { |
| for (i = 0; i < inst->outLen; i++) { |
| inst->outBuf[i] = fout[i + inst->blockLen10ms]; |
| } |
| } |
| } // end of if out.len==0 |
| else { |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| fout[i] = inst->outBuf[i]; |
| } |
| memcpy(inst->outBuf, inst->outBuf + inst->blockLen10ms, |
| sizeof(float) * (inst->outLen - inst->blockLen10ms)); |
| memset(inst->outBuf + inst->outLen - inst->blockLen10ms, 0, |
| sizeof(float) * inst->blockLen10ms); |
| inst->outLen -= inst->blockLen10ms; |
| } |
| |
| // convert to short |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| dTmp = fout[i]; |
| if (dTmp < WEBRTC_SPL_WORD16_MIN) { |
| dTmp = WEBRTC_SPL_WORD16_MIN; |
| } else if (dTmp > WEBRTC_SPL_WORD16_MAX) { |
| dTmp = WEBRTC_SPL_WORD16_MAX; |
| } |
| outFrame[i] = (short)dTmp; |
| } |
| |
| // for time-domain gain of HB |
| if (flagHB == 1) { |
| for (i = 0; i < inst->magnLen; i++) { |
| inst->speechProbHB[i] = probSpeechFinal[i]; |
| } |
| // average speech prob from low band |
| // avg over second half (i.e., 4->8kHz) of freq. spectrum |
| avgProbSpeechHB = 0.0; |
| for (i = inst->magnLen - deltaBweHB - 1; i < inst->magnLen - 1; i++) { |
| avgProbSpeechHB += inst->speechProbHB[i]; |
| } |
| avgProbSpeechHB = avgProbSpeechHB / ((float)deltaBweHB); |
| // average filter gain from low band |
| // average over second half (i.e., 4->8kHz) of freq. spectrum |
| avgFilterGainHB = 0.0; |
| for (i = inst->magnLen - deltaGainHB - 1; i < inst->magnLen - 1; i++) { |
| avgFilterGainHB += inst->smooth[i]; |
| } |
| avgFilterGainHB = avgFilterGainHB / ((float)(deltaGainHB)); |
| avgProbSpeechHBTmp = (float)2.0 * avgProbSpeechHB - (float)1.0; |
| // gain based on speech prob: |
| gainModHB = (float)0.5 * ((float)1.0 + (float)tanh(gainMapParHB * avgProbSpeechHBTmp)); |
| //combine gain with low band gain |
| gainTimeDomainHB = (float)0.5 * gainModHB + (float)0.5 * avgFilterGainHB; |
| if (avgProbSpeechHB >= (float)0.5) { |
| gainTimeDomainHB = (float)0.25 * gainModHB + (float)0.75 * avgFilterGainHB; |
| } |
| gainTimeDomainHB = gainTimeDomainHB * decayBweHB; |
| //make sure gain is within flooring range |
| // flooring bottom |
| if (gainTimeDomainHB < inst->denoiseBound) { |
| gainTimeDomainHB = inst->denoiseBound; |
| } |
| // flooring top |
| if (gainTimeDomainHB > (float)1.0) { |
| gainTimeDomainHB = 1.0; |
| } |
| //apply gain |
| for (i = 0; i < inst->blockLen10ms; i++) { |
| dTmp = gainTimeDomainHB * inst->dataBufHB[i]; |
| if (dTmp < WEBRTC_SPL_WORD16_MIN) { |
| dTmp = WEBRTC_SPL_WORD16_MIN; |
| } else if (dTmp > WEBRTC_SPL_WORD16_MAX) { |
| dTmp = WEBRTC_SPL_WORD16_MAX; |
| } |
| outFrameHB[i] = (short)dTmp; |
| } |
| } // end of H band gain computation |
| // |
| |
| return 0; |
| } |
