/* Copyright (C) 2003-2006 Jean-Marc Valin File: mdf.c Echo canceller based on the MDF algorithm (see below) Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 3. The name of the author may not be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ /* The echo canceller is based on the MDF algorithm described in: J. S. Soo, K. K. Pang Multidelay block frequency adaptive filter, IEEE Trans. Acoust. Speech Signal Process., Vol. ASSP-38, No. 2, February 1990. We use the Alternatively Updated MDF (AUMDF) variant. Robustness to double-talk is achieved using a variable learning rate as described in: Valin, J.-M., On Adjusting the Learning Rate in Frequency Domain Echo Cancellation With Double-Talk. IEEE Transactions on Audio, Speech and Language Processing, Vol. 15, No. 3, pp. 1030-1034, 2007. http://people.xiph.org/~jm/papers/valin_taslp2006.pdf There is no explicit double-talk detection, but a continuous variation in the learning rate based on residual echo, double-talk and background noise. About the fixed-point version: All the signals are represented with 16-bit words. The filter weights are represented with 32-bit words, but only the top 16 bits are used in most cases. The lower 16 bits are completely unreliable (due to the fact that the update is done only on the top bits), but help in the adaptation -- probably by removing a "threshold effect" due to quantization (rounding going to zero) when the gradient is small. Another kludge that seems to work good: when performing the weight update, we only move half the way toward the "goal" this seems to reduce the effect of quantization noise in the update phase. This can be seen as applying a gradient descent on a "soft constraint" instead of having a hard constraint. */ #ifdef HAVE_CONFIG_H #include "config.h" #endif #include "arch.h" #include "speex/speex_echo.h" #include "fftwrap.h" #include "pseudofloat.h" #include "math_approx.h" #include "os_support.h" #ifndef M_PI #define M_PI 3.14159265358979323846 #endif #ifdef FIXED_POINT #define WEIGHT_SHIFT 11 #define NORMALIZE_SCALEDOWN 5 #define NORMALIZE_SCALEUP 3 #else #define WEIGHT_SHIFT 0 #endif /* If enabled, the AEC will use a foreground filter and a background filter to be more robust to double-talk and difficult signals in general. The cost is an extra FFT and a matrix-vector multiply */ #define TWO_PATH #ifdef FIXED_POINT static const spx_float_t MIN_LEAK = {20972, -22}; /* Constants for the two-path filter */ static const spx_float_t VAR1_SMOOTH = {23593, -16}; static const spx_float_t VAR2_SMOOTH = {23675, -15}; static const spx_float_t VAR1_UPDATE = {16384, -15}; static const spx_float_t VAR2_UPDATE = {16384, -16}; static const spx_float_t VAR_BACKTRACK = {16384, -12}; #define TOP16(x) ((x)>>16) #else static const spx_float_t MIN_LEAK = .005f; /* Constants for the two-path filter */ static const spx_float_t VAR1_SMOOTH = .36f; static const spx_float_t VAR2_SMOOTH = .7225f; static const spx_float_t VAR1_UPDATE = .5f; static const spx_float_t VAR2_UPDATE = .25f; static const spx_float_t VAR_BACKTRACK = 4.f; #define TOP16(x) (x) #endif #define PLAYBACK_DELAY 2 void speex_echo_get_residual(SpeexEchoState *st, spx_word32_t *Yout, int len); /** Speex echo cancellation state. */ struct SpeexEchoState_ { int frame_size; /**< Number of samples processed each time */ int window_size; int M; int cancel_count; int adapted; int saturated; int screwed_up; spx_int32_t sampling_rate; spx_word16_t spec_average; spx_word16_t beta0; spx_word16_t beta_max; spx_word32_t sum_adapt; spx_word16_t leak_estimate; spx_word16_t *e; /* scratch */ spx_word16_t *x; /* Far-end input buffer (2N) */ spx_word16_t *X; /* Far-end buffer (M+1 frames) in frequency domain */ spx_word16_t *input; /* scratch */ spx_word16_t *y; /* scratch */ spx_word16_t *last_y; spx_word16_t *Y; /* scratch */ spx_word16_t *E; spx_word32_t *PHI; /* scratch */ spx_word32_t *W; /* (Background) filter weights */ #ifdef TWO_PATH spx_word16_t *foreground; /* Foreground filter weights */ spx_word32_t Davg1; /* 1st recursive average of the residual power difference */ spx_word32_t Davg2; /* 2nd recursive average of the residual power difference */ spx_float_t Dvar1; /* Estimated variance of 1st estimator */ spx_float_t Dvar2; /* Estimated variance of 2nd estimator */ #endif spx_word32_t *power; /* Power of the far-end signal */ spx_float_t *power_1;/* Inverse power of far-end */ spx_word16_t *wtmp; /* scratch */ #ifdef FIXED_POINT spx_word16_t *wtmp2; /* scratch */ #endif spx_word32_t *Rf; /* scratch */ spx_word32_t *Yf; /* scratch */ spx_word32_t *Xf; /* scratch */ spx_word32_t *Eh; spx_word32_t *Yh; spx_float_t Pey; spx_float_t Pyy; spx_word16_t *window; spx_word16_t *prop; void *fft_table; spx_word16_t memX, memD, memE; spx_word16_t preemph; spx_word16_t notch_radius; spx_mem_t notch_mem[2]; /* NOTE: If you only use speex_echo_cancel() and want to save some memory, remove this */ spx_int16_t *play_buf; int play_buf_pos; int play_buf_started; }; static inline void filter_dc_notch16(const spx_int16_t *in, spx_word16_t radius, spx_word16_t *out, int len, spx_mem_t *mem) { int i; spx_word16_t den2; #ifdef FIXED_POINT den2 = MULT16_16_Q15(radius,radius) + MULT16_16_Q15(QCONST16(.7,15),MULT16_16_Q15(32767-radius,32767-radius)); #else den2 = radius*radius + .7*(1-radius)*(1-radius); #endif /*printf ("%d %d %d %d %d %d\n", num[0], num[1], num[2], den[0], den[1], den[2]);*/ for (i=0;i<len;i++) { spx_word16_t vin = in[i]; spx_word32_t vout = mem[0] + SHL32(EXTEND32(vin),15); #ifdef FIXED_POINT mem[0] = mem[1] + SHL32(SHL32(-EXTEND32(vin),15) + MULT16_32_Q15(radius,vout),1); #else mem[0] = mem[1] + 2*(-vin + radius*vout); #endif mem[1] = SHL32(EXTEND32(vin),15) - MULT16_32_Q15(den2,vout); out[i] = SATURATE32(PSHR32(MULT16_32_Q15(radius,vout),15),32767); } } /* This inner product is slightly different from the codec version because of fixed-point */ static inline spx_word32_t mdf_inner_prod(const spx_word16_t *x, const spx_word16_t *y, int len) { spx_word32_t sum=0; len >>= 1; while(len--) { spx_word32_t part=0; part = MAC16_16(part,*x++,*y++); part = MAC16_16(part,*x++,*y++); /* HINT: If you had a 40-bit accumulator, you could shift only at the end */ sum = ADD32(sum,SHR32(part,6)); } return sum; } /** Compute power spectrum of a half-complex (packed) vector */ static inline void power_spectrum(const spx_word16_t *X, spx_word32_t *ps, int N) { int i, j; ps[0]=MULT16_16(X[0],X[0]); for (i=1,j=1;i<N-1;i+=2,j++) { ps[j] = MULT16_16(X[i],X[i]) + MULT16_16(X[i+1],X[i+1]); } ps[j]=MULT16_16(X[i],X[i]); } /** Compute cross-power spectrum of a half-complex (packed) vectors and add to acc */ #ifdef FIXED_POINT static inline void spectral_mul_accum(const spx_word16_t *X, const spx_word32_t *Y, spx_word16_t *acc, int N, int M) { int i,j; spx_word32_t tmp1=0,tmp2=0; for (j=0;j<M;j++) { tmp1 = MAC16_16(tmp1, X[j*N],TOP16(Y[j*N])); } acc[0] = PSHR32(tmp1,WEIGHT_SHIFT); for (i=1;i<N-1;i+=2) { tmp1 = tmp2 = 0; for (j=0;j<M;j++) { tmp1 = SUB32(MAC16_16(tmp1, X[j*N+i],TOP16(Y[j*N+i])), MULT16_16(X[j*N+i+1],TOP16(Y[j*N+i+1]))); tmp2 = MAC16_16(MAC16_16(tmp2, X[j*N+i+1],TOP16(Y[j*N+i])), X[j*N+i], TOP16(Y[j*N+i+1])); } acc[i] = PSHR32(tmp1,WEIGHT_SHIFT); acc[i+1] = PSHR32(tmp2,WEIGHT_SHIFT); } tmp1 = tmp2 = 0; for (j=0;j<M;j++) { tmp1 = MAC16_16(tmp1, X[(j+1)*N-1],TOP16(Y[(j+1)*N-1])); } acc[N-1] = PSHR32(tmp1,WEIGHT_SHIFT); } static inline void spectral_mul_accum16(const spx_word16_t *X, const spx_word16_t *Y, spx_word16_t *acc, int N, int M) { int i,j; spx_word32_t tmp1=0,tmp2=0; for (j=0;j<M;j++) { tmp1 = MAC16_16(tmp1, X[j*N],Y[j*N]); } acc[0] = PSHR32(tmp1,WEIGHT_SHIFT); for (i=1;i<N-1;i+=2) { tmp1 = tmp2 = 0; for (j=0;j<M;j++) { tmp1 = SUB32(MAC16_16(tmp1, X[j*N+i],Y[j*N+i]), MULT16_16(X[j*N+i+1],Y[j*N+i+1])); tmp2 = MAC16_16(MAC16_16(tmp2, X[j*N+i+1],Y[j*N+i]), X[j*N+i], Y[j*N+i+1]); } acc[i] = PSHR32(tmp1,WEIGHT_SHIFT); acc[i+1] = PSHR32(tmp2,WEIGHT_SHIFT); } tmp1 = tmp2 = 0; for (j=0;j<M;j++) { tmp1 = MAC16_16(tmp1, X[(j+1)*N-1],Y[(j+1)*N-1]); } acc[N-1] = PSHR32(tmp1,WEIGHT_SHIFT); } #else static inline void spectral_mul_accum(const spx_word16_t *X, const spx_word32_t *Y, spx_word16_t *acc, int N, int M) { int i,j; for (i=0;i<N;i++) acc[i] = 0; for (j=0;j<M;j++) { acc[0] += X[0]*Y[0]; for (i=1;i<N-1;i+=2) { acc[i] += (X[i]*Y[i] - X[i+1]*Y[i+1]); acc[i+1] += (X[i+1]*Y[i] + X[i]*Y[i+1]); } acc[i] += X[i]*Y[i]; X += N; Y += N; } } #define spectral_mul_accum16 spectral_mul_accum #endif /** Compute weighted cross-power spectrum of a half-complex (packed) vector with conjugate */ static inline void weighted_spectral_mul_conj(const spx_float_t *w, const spx_float_t p, const spx_word16_t *X, const spx_word16_t *Y, spx_word32_t *prod, int N) { int i, j; spx_float_t W; W = FLOAT_AMULT(p, w[0]); prod[0] = FLOAT_MUL32(W,MULT16_16(X[0],Y[0])); for (i=1,j=1;i<N-1;i+=2,j++) { W = FLOAT_AMULT(p, w[j]); prod[i] = FLOAT_MUL32(W,MAC16_16(MULT16_16(X[i],Y[i]), X[i+1],Y[i+1])); prod[i+1] = FLOAT_MUL32(W,MAC16_16(MULT16_16(-X[i+1],Y[i]), X[i],Y[i+1])); } W = FLOAT_AMULT(p, w[j]); prod[i] = FLOAT_MUL32(W,MULT16_16(X[i],Y[i])); } static inline void mdf_adjust_prop(const spx_word32_t *W, int N, int M, spx_word16_t *prop) { int i, j; spx_word16_t max_sum = 1; spx_word32_t prop_sum = 1; for (i=0;i<M;i++) { spx_word32_t tmp = 1; for (j=0;j<N;j++) tmp += MULT16_16(EXTRACT16(SHR32(W[i*N+j],18)), EXTRACT16(SHR32(W[i*N+j],18))); #ifdef FIXED_POINT /* Just a security in case an overflow were to occur */ tmp = MIN32(ABS32(tmp), 536870912); #endif prop[i] = spx_sqrt(tmp); if (prop[i] > max_sum) max_sum = prop[i]; } for (i=0;i<M;i++) { prop[i] += MULT16_16_Q15(QCONST16(.1f,15),max_sum); prop_sum += EXTEND32(prop[i]); } for (i=0;i<M;i++) { prop[i] = DIV32(MULT16_16(QCONST16(.99f,15), prop[i]),prop_sum); /*printf ("%f ", prop[i]);*/ } /*printf ("\n");*/ } #ifdef DUMP_ECHO_CANCEL_DATA #include <stdio.h> static FILE *rFile=NULL, *pFile=NULL, *oFile=NULL; static void dump_audio(const spx_int16_t *rec, const spx_int16_t *play, const spx_int16_t *out, int len) { if (!(rFile && pFile && oFile)) { speex_fatal("Dump files not open"); } fwrite(rec, sizeof(spx_int16_t), len, rFile); fwrite(play, sizeof(spx_int16_t), len, pFile); fwrite(out, sizeof(spx_int16_t), len, oFile); } #endif /** Creates a new echo canceller state */ SpeexEchoState *speex_echo_state_init(int frame_size, int filter_length) { int i,N,M; SpeexEchoState *st = (SpeexEchoState *)speex_alloc(sizeof(SpeexEchoState)); #ifdef DUMP_ECHO_CANCEL_DATA if (rFile || pFile || oFile) speex_fatal("Opening dump files twice"); rFile = fopen("aec_rec.sw", "wb"); pFile = fopen("aec_play.sw", "wb"); oFile = fopen("aec_out.sw", "wb"); #endif st->frame_size = frame_size; st->window_size = 2*frame_size; N = st->window_size; M = st->M = (filter_length+st->frame_size-1)/frame_size; st->cancel_count=0; st->sum_adapt = 0; st->saturated = 0; st->screwed_up = 0; /* This is the default sampling rate */ st->sampling_rate = 8000; st->spec_average = DIV32_16(SHL32(EXTEND32(st->frame_size), 15), st->sampling_rate); #ifdef FIXED_POINT st->beta0 = DIV32_16(SHL32(EXTEND32(st->frame_size), 16), st->sampling_rate); st->beta_max = DIV32_16(SHL32(EXTEND32(st->frame_size), 14), st->sampling_rate); #else st->beta0 = (2.0f*st->frame_size)/st->sampling_rate; st->beta_max = (.5f*st->frame_size)/st->sampling_rate; #endif st->leak_estimate = 0; st->fft_table = spx_fft_init(N); st->e = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->x = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->input = (spx_word16_t*)speex_alloc(st->frame_size*sizeof(spx_word16_t)); st->y = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->last_y = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->Yf = (spx_word32_t*)speex_alloc((st->frame_size+1)*sizeof(spx_word32_t)); st->Rf = (spx_word32_t*)speex_alloc((st->frame_size+1)*sizeof(spx_word32_t)); st->Xf = (spx_word32_t*)speex_alloc((st->frame_size+1)*sizeof(spx_word32_t)); st->Yh = (spx_word32_t*)speex_alloc((st->frame_size+1)*sizeof(spx_word32_t)); st->Eh = (spx_word32_t*)speex_alloc((st->frame_size+1)*sizeof(spx_word32_t)); st->X = (spx_word16_t*)speex_alloc((M+1)*N*sizeof(spx_word16_t)); st->Y = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->E = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->W = (spx_word32_t*)speex_alloc(M*N*sizeof(spx_word32_t)); #ifdef TWO_PATH st->foreground = (spx_word16_t*)speex_alloc(M*N*sizeof(spx_word16_t)); #endif st->PHI = (spx_word32_t*)speex_alloc(N*sizeof(spx_word32_t)); st->power = (spx_word32_t*)speex_alloc((frame_size+1)*sizeof(spx_word32_t)); st->power_1 = (spx_float_t*)speex_alloc((frame_size+1)*sizeof(spx_float_t)); st->window = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); st->prop = (spx_word16_t*)speex_alloc(M*sizeof(spx_word16_t)); st->wtmp = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); #ifdef FIXED_POINT st->wtmp2 = (spx_word16_t*)speex_alloc(N*sizeof(spx_word16_t)); for (i=0;i<N>>1;i++) { st->window[i] = (16383-SHL16(spx_cos(DIV32_16(MULT16_16(25736,i<<1),N)),1)); st->window[N-i-1] = st->window[i]; } #else for (i=0;i<N;i++) st->window[i] = .5-.5*cos(2*M_PI*i/N); #endif for (i=0;i<=st->frame_size;i++) st->power_1[i] = FLOAT_ONE; for (i=0;i<N*M;i++) st->W[i] = 0; { spx_word32_t sum = 0; /* Ratio of ~10 between adaptation rate of first and last block */ spx_word16_t decay = SHR32(spx_exp(NEG16(DIV32_16(QCONST16(2.4,11),M))),1); st->prop[0] = QCONST16(.7, 15); sum = EXTEND32(st->prop[0]); for (i=1;i<M;i++) { st->prop[i] = MULT16_16_Q15(st->prop[i-1], decay); sum = ADD32(sum, EXTEND32(st->prop[i])); } for (i=M-1;i>=0;i--) { st->prop[i] = DIV32(MULT16_16(QCONST16(.8,15), st->prop[i]),sum); } } st->memX=st->memD=st->memE=0; st->preemph = QCONST16(.9,15); if (st->sampling_rate<12000) st->notch_radius = QCONST16(.9, 15); else if (st->sampling_rate<24000) st->notch_radius = QCONST16(.982, 15); else st->notch_radius = QCONST16(.992, 15); st->notch_mem[0] = st->notch_mem[1] = 0; st->adapted = 0; st->Pey = st->Pyy = FLOAT_ONE; #ifdef TWO_PATH st->Davg1 = st->Davg2 = 0; st->Dvar1 = st->Dvar2 = FLOAT_ZERO; #endif st->play_buf = (spx_int16_t*)speex_alloc((PLAYBACK_DELAY+1)*st->frame_size*sizeof(spx_int16_t)); st->play_buf_pos = PLAYBACK_DELAY*st->frame_size; st->play_buf_started = 0; return st; } /** Resets echo canceller state */ void speex_echo_state_reset(SpeexEchoState *st) { int i, M, N; st->cancel_count=0; st->screwed_up = 0; N = st->window_size; M = st->M; for (i=0;i<N*M;i++) st->W[i] = 0; #ifdef TWO_PATH for (i=0;i<N*M;i++) st->foreground[i] = 0; #endif for (i=0;i<N*(M+1);i++) st->X[i] = 0; for (i=0;i<=st->frame_size;i++) { st->power[i] = 0; st->power_1[i] = FLOAT_ONE; st->Eh[i] = 0; st->Yh[i] = 0; } for (i=0;i<st->frame_size;i++) { st->last_y[i] = 0; } for (i=0;i<N;i++) { st->E[i] = 0; st->x[i] = 0; } st->notch_mem[0] = st->notch_mem[1] = 0; st->memX=st->memD=st->memE=0; st->saturated = 0; st->adapted = 0; st->sum_adapt = 0; st->Pey = st->Pyy = FLOAT_ONE; #ifdef TWO_PATH st->Davg1 = st->Davg2 = 0; st->Dvar1 = st->Dvar2 = FLOAT_ZERO; #endif for (i=0;i<3*st->frame_size;i++) st->play_buf[i] = 0; st->play_buf_pos = PLAYBACK_DELAY*st->frame_size; st->play_buf_started = 0; } /** Destroys an echo canceller state */ void speex_echo_state_destroy(SpeexEchoState *st) { spx_fft_destroy(st->fft_table); speex_free(st->e); speex_free(st->x); speex_free(st->input); speex_free(st->y); speex_free(st->last_y); speex_free(st->Yf); speex_free(st->Rf); speex_free(st->Xf); speex_free(st->Yh); speex_free(st->Eh); speex_free(st->X); speex_free(st->Y); speex_free(st->E); speex_free(st->W); #ifdef TWO_PATH speex_free(st->foreground); #endif speex_free(st->PHI); speex_free(st->power); speex_free(st->power_1); speex_free(st->window); speex_free(st->prop); speex_free(st->wtmp); #ifdef FIXED_POINT speex_free(st->wtmp2); #endif speex_free(st->play_buf); speex_free(st); #ifdef DUMP_ECHO_CANCEL_DATA fclose(rFile); fclose(pFile); fclose(oFile); rFile = pFile = oFile = NULL; #endif } void speex_echo_capture(SpeexEchoState *st, const spx_int16_t *rec, spx_int16_t *out) { int i; /*speex_warning_int("capture with fill level ", st->play_buf_pos/st->frame_size);*/ st->play_buf_started = 1; if (st->play_buf_pos>=st->frame_size) { speex_echo_cancellation(st, rec, st->play_buf, out); st->play_buf_pos -= st->frame_size; for (i=0;i<st->play_buf_pos;i++) st->play_buf[i] = st->play_buf[i+st->frame_size]; } else { speex_warning("No playback frame available (your application is buggy and/or got xruns)"); if (st->play_buf_pos!=0) { speex_warning("internal playback buffer corruption?"); st->play_buf_pos = 0; } for (i=0;i<st->frame_size;i++) out[i] = rec[i]; } } void speex_echo_playback(SpeexEchoState *st, const spx_int16_t *play) { /*speex_warning_int("playback with fill level ", st->play_buf_pos/st->frame_size);*/ if (!st->play_buf_started) { speex_warning("discarded first playback frame"); return; } if (st->play_buf_pos<=PLAYBACK_DELAY*st->frame_size) { int i; for (i=0;i<st->frame_size;i++) st->play_buf[st->play_buf_pos+i] = play[i]; st->play_buf_pos += st->frame_size; if (st->play_buf_pos <= (PLAYBACK_DELAY-1)*st->frame_size) { speex_warning("Auto-filling the buffer (your application is buggy and/or got xruns)"); for (i=0;i<st->frame_size;i++) st->play_buf[st->play_buf_pos+i] = play[i]; st->play_buf_pos += st->frame_size; } } else { speex_warning("Had to discard a playback frame (your application is buggy and/or got xruns)"); } } /** Performs echo cancellation on a frame (deprecated, last arg now ignored) */ void speex_echo_cancel(SpeexEchoState *st, const spx_int16_t *in, const spx_int16_t *far_end, spx_int16_t *out, spx_int32_t *Yout) { speex_echo_cancellation(st, in, far_end, out); } /** Performs echo cancellation on a frame */ void speex_echo_cancellation(SpeexEchoState *st, const spx_int16_t *in, const spx_int16_t *far_end, spx_int16_t *out) { int i,j; int N,M; spx_word32_t Syy,See,Sxx,Sdd, Sff; #ifdef TWO_PATH spx_word32_t Dbf; int update_foreground; #endif spx_word32_t Sey; spx_word16_t ss, ss_1; spx_float_t Pey = FLOAT_ONE, Pyy=FLOAT_ONE; spx_float_t alpha, alpha_1; spx_word16_t RER; spx_word32_t tmp32; N = st->window_size; M = st->M; st->cancel_count++; #ifdef FIXED_POINT ss=DIV32_16(11469,M); ss_1 = SUB16(32767,ss); #else ss=.35/M; ss_1 = 1-ss; #endif /* Apply a notch filter to make sure DC doesn't end up causing problems */ filter_dc_notch16(in, st->notch_radius, st->input, st->frame_size, st->notch_mem); /* Copy input data to buffer and apply pre-emphasis */ for (i=0;i<st->frame_size;i++) { spx_word32_t tmp32; tmp32 = SUB32(EXTEND32(far_end[i]), EXTEND32(MULT16_16_P15(st->preemph, st->memX))); #ifdef FIXED_POINT /* If saturation occurs here, we need to freeze adaptation for M+1 frames (not just one) */ if (tmp32 > 32767) { tmp32 = 32767; st->saturated = M+1; } if (tmp32 < -32767) { tmp32 = -32767; st->saturated = M+1; } #endif st->x[i+st->frame_size] = EXTRACT16(tmp32); st->memX = far_end[i]; tmp32 = SUB32(EXTEND32(st->input[i]), EXTEND32(MULT16_16_P15(st->preemph, st->memD))); #ifdef FIXED_POINT if (tmp32 > 32767) { tmp32 = 32767; if (st->saturated == 0) st->saturated = 1; } if (tmp32 < -32767) { tmp32 = -32767; if (st->saturated == 0) st->saturated = 1; } #endif st->memD = st->input[i]; st->input[i] = tmp32; } /* Shift memory: this could be optimized eventually*/ for (j=M-1;j>=0;j--) { for (i=0;i<N;i++) st->X[(j+1)*N+i] = st->X[j*N+i]; } /* Convert x (far end) to frequency domain */ spx_fft(st->fft_table, st->x, &st->X[0]); for (i=0;i<N;i++) st->last_y[i] = st->x[i]; Sxx = mdf_inner_prod(st->x+st->frame_size, st->x+st->frame_size, st->frame_size); for (i=0;i<st->frame_size;i++) st->x[i] = st->x[i+st->frame_size]; /* From here on, the top part of x is used as scratch space */ #ifdef TWO_PATH /* Compute foreground filter */ spectral_mul_accum16(st->X, st->foreground, st->Y, N, M); spx_ifft(st->fft_table, st->Y, st->e); for (i=0;i<st->frame_size;i++) st->e[i] = SUB16(st->input[i], st->e[i+st->frame_size]); Sff = mdf_inner_prod(st->e, st->e, st->frame_size); #endif /* Adjust proportional adaption rate */ mdf_adjust_prop (st->W, N, M, st->prop); /* Compute weight gradient */ if (st->saturated == 0) { for (j=M-1;j>=0;j--) { weighted_spectral_mul_conj(st->power_1, FLOAT_SHL(PSEUDOFLOAT(st->prop[j]),-15), &st->X[(j+1)*N], st->E, st->PHI, N); for (i=0;i<N;i++) st->W[j*N+i] = ADD32(st->W[j*N+i], st->PHI[i]); } } else { st->saturated--; } /* Update weight to prevent circular convolution (MDF / AUMDF) */ for (j=0;j<M;j++) { /* This is a variant of the Alternatively Updated MDF (AUMDF) */ /* Remove the "if" to make this an MDF filter */ if (j==0 || st->cancel_count%(M-1) == j-1) { #ifdef FIXED_POINT for (i=0;i<N;i++) st->wtmp2[i] = EXTRACT16(PSHR32(st->W[j*N+i],NORMALIZE_SCALEDOWN+16)); spx_ifft(st->fft_table, st->wtmp2, st->wtmp); for (i=0;i<st->frame_size;i++) { st->wtmp[i]=0; } for (i=st->frame_size;i<N;i++) { st->wtmp[i]=SHL16(st->wtmp[i],NORMALIZE_SCALEUP); } spx_fft(st->fft_table, st->wtmp, st->wtmp2); /* The "-1" in the shift is a sort of kludge that trades less efficient update speed for decrease noise */ for (i=0;i<N;i++) st->W[j*N+i] -= SHL32(EXTEND32(st->wtmp2[i]),16+NORMALIZE_SCALEDOWN-NORMALIZE_SCALEUP-1); #else spx_ifft(st->fft_table, &st->W[j*N], st->wtmp); for (i=st->frame_size;i<N;i++) { st->wtmp[i]=0; } spx_fft(st->fft_table, st->wtmp, &st->W[j*N]); #endif } } /* Compute filter response Y */ spectral_mul_accum(st->X, st->W, st->Y, N, M); spx_ifft(st->fft_table, st->Y, st->y); #ifdef TWO_PATH /* Difference in response, this is used to estimate the variance of our residual power estimate */ for (i=0;i<st->frame_size;i++) st->e[i] = SUB16(st->e[i+st->frame_size], st->y[i+st->frame_size]); Dbf = 10+mdf_inner_prod(st->e, st->e, st->frame_size); #endif for (i=0;i<st->frame_size;i++) st->e[i] = SUB16(st->input[i], st->y[i+st->frame_size]); See = mdf_inner_prod(st->e, st->e, st->frame_size); #ifndef TWO_PATH Sff = See; #endif #ifdef TWO_PATH /* Logic for updating the foreground filter */ /* For two time windows, compute the mean of the energy difference, as well as the variance */ st->Davg1 = ADD32(MULT16_32_Q15(QCONST16(.6f,15),st->Davg1), MULT16_32_Q15(QCONST16(.4f,15),SUB32(Sff,See))); st->Davg2 = ADD32(MULT16_32_Q15(QCONST16(.85f,15),st->Davg2), MULT16_32_Q15(QCONST16(.15f,15),SUB32(Sff,See))); st->Dvar1 = FLOAT_ADD(FLOAT_MULT(VAR1_SMOOTH, st->Dvar1), FLOAT_MUL32U(MULT16_32_Q15(QCONST16(.4f,15),Sff), MULT16_32_Q15(QCONST16(.4f,15),Dbf))); st->Dvar2 = FLOAT_ADD(FLOAT_MULT(VAR2_SMOOTH, st->Dvar2), FLOAT_MUL32U(MULT16_32_Q15(QCONST16(.15f,15),Sff), MULT16_32_Q15(QCONST16(.15f,15),Dbf))); /* Equivalent float code: st->Davg1 = .6*st->Davg1 + .4*(Sff-See); st->Davg2 = .85*st->Davg2 + .15*(Sff-See); st->Dvar1 = .36*st->Dvar1 + .16*Sff*Dbf; st->Dvar2 = .7225*st->Dvar2 + .0225*Sff*Dbf; */ update_foreground = 0; /* Check if we have a statistically significant reduction in the residual echo */ /* Note that this is *not* Gaussian, so we need to be careful about the longer tail */ if (FLOAT_GT(FLOAT_MUL32U(SUB32(Sff,See),ABS32(SUB32(Sff,See))), FLOAT_MUL32U(Sff,Dbf))) update_foreground = 1; else if (FLOAT_GT(FLOAT_MUL32U(st->Davg1, ABS32(st->Davg1)), FLOAT_MULT(VAR1_UPDATE,(st->Dvar1)))) update_foreground = 1; else if (FLOAT_GT(FLOAT_MUL32U(st->Davg2, ABS32(st->Davg2)), FLOAT_MULT(VAR2_UPDATE,(st->Dvar2)))) update_foreground = 1; /* Do we update? */ if (update_foreground) { st->Davg1 = st->Davg2 = 0; st->Dvar1 = st->Dvar2 = FLOAT_ZERO; /* Copy background filter to foreground filter */ for (i=0;i<N*M;i++) st->foreground[i] = EXTRACT16(PSHR32(st->W[i],16)); /* Apply a smooth transition so as to not introduce blocking artifacts */ for (i=0;i<st->frame_size;i++) st->e[i+st->frame_size] = MULT16_16_Q15(st->window[i+st->frame_size],st->e[i+st->frame_size]) + MULT16_16_Q15(st->window[i],st->y[i+st->frame_size]); } else { int reset_background=0; /* Otherwise, check if the background filter is significantly worse */ if (FLOAT_GT(FLOAT_MUL32U(NEG32(SUB32(Sff,See)),ABS32(SUB32(Sff,See))), FLOAT_MULT(VAR_BACKTRACK,FLOAT_MUL32U(Sff,Dbf)))) reset_background = 1; if (FLOAT_GT(FLOAT_MUL32U(NEG32(st->Davg1), ABS32(st->Davg1)), FLOAT_MULT(VAR_BACKTRACK,st->Dvar1))) reset_background = 1; if (FLOAT_GT(FLOAT_MUL32U(NEG32(st->Davg2), ABS32(st->Davg2)), FLOAT_MULT(VAR_BACKTRACK,st->Dvar2))) reset_background = 1; if (reset_background) { /* Copy foreground filter to background filter */ for (i=0;i<N*M;i++) st->W[i] = SHL32(EXTEND32(st->foreground[i]),16); /* We also need to copy the output so as to get correct adaptation */ for (i=0;i<st->frame_size;i++) st->y[i+st->frame_size] = st->e[i+st->frame_size]; for (i=0;i<st->frame_size;i++) st->e[i] = SUB16(st->input[i], st->y[i+st->frame_size]); See = Sff; st->Davg1 = st->Davg2 = 0; st->Dvar1 = st->Dvar2 = FLOAT_ZERO; } } #endif /* Compute error signal (for the output with de-emphasis) */ for (i=0;i<st->frame_size;i++) { spx_word32_t tmp_out; #ifdef TWO_PATH tmp_out = SUB32(EXTEND32(st->input[i]), EXTEND32(st->e[i+st->frame_size])); #else tmp_out = SUB32(EXTEND32(st->input[i]), EXTEND32(st->y[i+st->frame_size])); #endif /* Saturation */ if (tmp_out>32767) tmp_out = 32767; else if (tmp_out<-32768) tmp_out = -32768; tmp_out = ADD32(tmp_out, EXTEND32(MULT16_16_P15(st->preemph, st->memE))); /* This is an arbitrary test for saturation in the microphone signal */ if (in[i] <= -32000 || in[i] >= 32000) { tmp_out = 0; if (st->saturated == 0) st->saturated = 1; } out[i] = (spx_int16_t)tmp_out; st->memE = tmp_out; } #ifdef DUMP_ECHO_CANCEL_DATA dump_audio(in, far_end, out, st->frame_size); #endif /* Compute error signal (filter update version) */ for (i=0;i<st->frame_size;i++) { st->e[i+st->frame_size] = st->e[i]; st->e[i] = 0; } /* Compute a bunch of correlations */ Sey = mdf_inner_prod(st->e+st->frame_size, st->y+st->frame_size, st->frame_size); Syy = mdf_inner_prod(st->y+st->frame_size, st->y+st->frame_size, st->frame_size); Sdd = mdf_inner_prod(st->input, st->input, st->frame_size); /*printf ("%f %f %f %f\n", Sff, See, Syy, Sdd, st->update_cond);*/ /* Do some sanity check */ if (!(Syy>=0 && Sxx>=0 && See >= 0) #ifndef FIXED_POINT || !(Sff < N*1e9 && Syy < N*1e9 && Sxx < N*1e9) #endif ) { /* Things have gone really bad */ st->screwed_up += 50; for (i=0;i<st->frame_size;i++) out[i] = 0; } else if (SHR32(Sff, 2) > ADD32(Sdd, SHR32(MULT16_16(N, 10000),6))) { /* AEC seems to add lots of echo instead of removing it, let's see if it will improve */ st->screwed_up++; } else { /* Everything's fine */ st->screwed_up=0; } if (st->screwed_up>=50) { speex_warning("The echo canceller started acting funny and got slapped (reset). It swears it will behave now."); speex_echo_state_reset(st); return; } /* Add a small noise floor to make sure not to have problems when dividing */ See = MAX32(See, SHR32(MULT16_16(N, 100),6)); /* Convert error to frequency domain */ spx_fft(st->fft_table, st->e, st->E); for (i=0;i<st->frame_size;i++) st->y[i] = 0; spx_fft(st->fft_table, st->y, st->Y); /* Compute power spectrum of far end (X), error (E) and filter response (Y) */ power_spectrum(st->E, st->Rf, N); power_spectrum(st->Y, st->Yf, N); power_spectrum(st->X, st->Xf, N); /* Smooth far end energy estimate over time */ for (j=0;j<=st->frame_size;j++) st->power[j] = MULT16_32_Q15(ss_1,st->power[j]) + 1 + MULT16_32_Q15(ss,st->Xf[j]); /* Enable this to compute the power based only on the tail (would need to compute more efficiently to make this really useful */ if (0) { float scale2 = .5f/M; for (j=0;j<=st->frame_size;j++) st->power[j] = 100; for (i=0;i<M;i++) { power_spectrum(&st->X[i*N], st->Xf, N); for (j=0;j<=st->frame_size;j++) st->power[j] += scale2*st->Xf[j]; } } /* Compute filtered spectra and (cross-)correlations */ for (j=st->frame_size;j>=0;j--) { spx_float_t Eh, Yh; Eh = PSEUDOFLOAT(st->Rf[j] - st->Eh[j]); Yh = PSEUDOFLOAT(st->Yf[j] - st->Yh[j]); Pey = FLOAT_ADD(Pey,FLOAT_MULT(Eh,Yh)); Pyy = FLOAT_ADD(Pyy,FLOAT_MULT(Yh,Yh)); #ifdef FIXED_POINT st->Eh[j] = MAC16_32_Q15(MULT16_32_Q15(SUB16(32767,st->spec_average),st->Eh[j]), st->spec_average, st->Rf[j]); st->Yh[j] = MAC16_32_Q15(MULT16_32_Q15(SUB16(32767,st->spec_average),st->Yh[j]), st->spec_average, st->Yf[j]); #else st->Eh[j] = (1-st->spec_average)*st->Eh[j] + st->spec_average*st->Rf[j]; st->Yh[j] = (1-st->spec_average)*st->Yh[j] + st->spec_average*st->Yf[j]; #endif } Pyy = FLOAT_SQRT(Pyy); Pey = FLOAT_DIVU(Pey,Pyy); /* Compute correlation updatete rate */ tmp32 = MULT16_32_Q15(st->beta0,Syy); if (tmp32 > MULT16_32_Q15(st->beta_max,See)) tmp32 = MULT16_32_Q15(st->beta_max,See); alpha = FLOAT_DIV32(tmp32, See); alpha_1 = FLOAT_SUB(FLOAT_ONE, alpha); /* Update correlations (recursive average) */ st->Pey = FLOAT_ADD(FLOAT_MULT(alpha_1,st->Pey) , FLOAT_MULT(alpha,Pey)); st->Pyy = FLOAT_ADD(FLOAT_MULT(alpha_1,st->Pyy) , FLOAT_MULT(alpha,Pyy)); if (FLOAT_LT(st->Pyy, FLOAT_ONE)) st->Pyy = FLOAT_ONE; /* We don't really hope to get better than 33 dB (MIN_LEAK-3dB) attenuation anyway */ if (FLOAT_LT(st->Pey, FLOAT_MULT(MIN_LEAK,st->Pyy))) st->Pey = FLOAT_MULT(MIN_LEAK,st->Pyy); if (FLOAT_GT(st->Pey, st->Pyy)) st->Pey = st->Pyy; /* leak_estimate is the linear regression result */ st->leak_estimate = FLOAT_EXTRACT16(FLOAT_SHL(FLOAT_DIVU(st->Pey, st->Pyy),14)); /* This looks like a stupid bug, but it's right (because we convert from Q14 to Q15) */ if (st->leak_estimate > 16383) st->leak_estimate = 32767; else st->leak_estimate = SHL16(st->leak_estimate,1); /*printf ("%f\n", st->leak_estimate);*/ /* Compute Residual to Error Ratio */ #ifdef FIXED_POINT tmp32 = MULT16_32_Q15(st->leak_estimate,Syy); tmp32 = ADD32(SHR32(Sxx,13), ADD32(tmp32, SHL32(tmp32,1))); /* Check for y in e (lower bound on RER) */ { spx_float_t bound = PSEUDOFLOAT(Sey); bound = FLOAT_DIVU(FLOAT_MULT(bound, bound), PSEUDOFLOAT(ADD32(1,Syy))); if (FLOAT_GT(bound, PSEUDOFLOAT(See))) tmp32 = See; else if (tmp32 < FLOAT_EXTRACT32(bound)) tmp32 = FLOAT_EXTRACT32(bound); } if (tmp32 > SHR32(See,1)) tmp32 = SHR32(See,1); RER = FLOAT_EXTRACT16(FLOAT_SHL(FLOAT_DIV32(tmp32,See),15)); #else RER = (.0001*Sxx + 3.*MULT16_32_Q15(st->leak_estimate,Syy)) / See; /* Check for y in e (lower bound on RER) */ if (RER < Sey*Sey/(1+See*Syy)) RER = Sey*Sey/(1+See*Syy); if (RER > .5) RER = .5; #endif /* We consider that the filter has had minimal adaptation if the following is true*/ if (!st->adapted && st->sum_adapt > SHL32(EXTEND32(M),15) && MULT16_32_Q15(st->leak_estimate,Syy) > MULT16_32_Q15(QCONST16(.03f,15),Syy)) { st->adapted = 1; } if (st->adapted) { /* Normal learning rate calculation once we're past the minimal adaptation phase */ for (i=0;i<=st->frame_size;i++) { spx_word32_t r, e; /* Compute frequency-domain adaptation mask */ r = MULT16_32_Q15(st->leak_estimate,SHL32(st->Yf[i],3)); e = SHL32(st->Rf[i],3)+1; #ifdef FIXED_POINT if (r>SHR32(e,1)) r = SHR32(e,1); #else if (r>.5*e) r = .5*e; #endif r = MULT16_32_Q15(QCONST16(.7,15),r) + MULT16_32_Q15(QCONST16(.3,15),(spx_word32_t)(MULT16_32_Q15(RER,e))); /*st->power_1[i] = adapt_rate*r/(e*(1+st->power[i]));*/ st->power_1[i] = FLOAT_SHL(FLOAT_DIV32_FLOAT(r,FLOAT_MUL32U(e,st->power[i]+10)),WEIGHT_SHIFT+16); } } else { /* Temporary adaption rate if filter is not yet adapted enough */ spx_word16_t adapt_rate=0; if (Sxx > SHR32(MULT16_16(N, 1000),6)) { tmp32 = MULT16_32_Q15(QCONST16(.25f, 15), Sxx); #ifdef FIXED_POINT if (tmp32 > SHR32(See,2)) tmp32 = SHR32(See,2); #else if (tmp32 > .25*See) tmp32 = .25*See; #endif adapt_rate = FLOAT_EXTRACT16(FLOAT_SHL(FLOAT_DIV32(tmp32, See),15)); } for (i=0;i<=st->frame_size;i++) st->power_1[i] = FLOAT_SHL(FLOAT_DIV32(EXTEND32(adapt_rate),ADD32(st->power[i],10)),WEIGHT_SHIFT+1); /* How much have we adapted so far? */ st->sum_adapt = ADD32(st->sum_adapt,adapt_rate); } /* Save residual echo so it can be used by the nonlinear processor */ if (st->adapted) { /* If the filter is adapted, take the filtered echo */ for (i=0;i<st->frame_size;i++) st->last_y[i] = st->last_y[st->frame_size+i]; for (i=0;i<st->frame_size;i++) st->last_y[st->frame_size+i] = in[i]-out[i]; } else { /* If filter isn't adapted yet, all we can do is take the far end signal directly */ /* moved earlier: for (i=0;i<N;i++) st->last_y[i] = st->x[i];*/ } } /* Compute spectrum of estimated echo for use in an echo post-filter */ void speex_echo_get_residual(SpeexEchoState *st, spx_word32_t *residual_echo, int len) { int i; spx_word16_t leak2; int N; N = st->window_size; /* Apply hanning window (should pre-compute it)*/ for (i=0;i<N;i++) st->y[i] = MULT16_16_Q15(st->window[i],st->last_y[i]); /* Compute power spectrum of the echo */ spx_fft(st->fft_table, st->y, st->Y); power_spectrum(st->Y, residual_echo, N); #ifdef FIXED_POINT if (st->leak_estimate > 16383) leak2 = 32767; else leak2 = SHL16(st->leak_estimate, 1); #else if (st->leak_estimate>.5) leak2 = 1; else leak2 = 2*st->leak_estimate; #endif /* Estimate residual echo */ for (i=0;i<=st->frame_size;i++) residual_echo[i] = (spx_int32_t)MULT16_32_Q15(leak2,residual_echo[i]); } int speex_echo_ctl(SpeexEchoState *st, int request, void *ptr) { switch(request) { case SPEEX_ECHO_GET_FRAME_SIZE: (*(int*)ptr) = st->frame_size; break; case SPEEX_ECHO_SET_SAMPLING_RATE: st->sampling_rate = (*(int*)ptr); st->spec_average = DIV32_16(SHL32(EXTEND32(st->frame_size), 15), st->sampling_rate); #ifdef FIXED_POINT st->beta0 = DIV32_16(SHL32(EXTEND32(st->frame_size), 16), st->sampling_rate); st->beta_max = DIV32_16(SHL32(EXTEND32(st->frame_size), 14), st->sampling_rate); #else st->beta0 = (2.0f*st->frame_size)/st->sampling_rate; st->beta_max = (.5f*st->frame_size)/st->sampling_rate; #endif if (st->sampling_rate<12000) st->notch_radius = QCONST16(.9, 15); else if (st->sampling_rate<24000) st->notch_radius = QCONST16(.982, 15); else st->notch_radius = QCONST16(.992, 15); break; case SPEEX_ECHO_GET_SAMPLING_RATE: (*(int*)ptr) = st->sampling_rate; break; default: speex_warning_int("Unknown speex_echo_ctl request: ", request); return -1; } return 0; }