mirror of
https://github.com/bluekitchen/btstack.git
synced 2025-02-04 06:39:53 +00:00
add static declaration, move functions
This commit is contained in:
parent
7e6b1e83d2
commit
c5e169ec5e
@ -56,9 +56,52 @@ static float rcos[CVSD_OLAL] = {
|
||||
0.45386582,0.36316850,0.27713082,0.19868268,
|
||||
0.13049554,0.07489143,0.03376389,0.00851345};
|
||||
|
||||
static float CrossCorrelation(int8_t *x, int8_t *y);
|
||||
static int PatternMatch(int8_t *y);
|
||||
static float AmplitudeMatch(int8_t *y, int8_t bestmatch);
|
||||
static float CrossCorrelation(int8_t *x, int8_t *y){
|
||||
float num = 0;
|
||||
float den = 0;
|
||||
float x2 = 0;
|
||||
float y2 = 0;
|
||||
int m;
|
||||
for (m=0;m<CVSD_M;m++){
|
||||
num+=((float)x[m])*y[m];
|
||||
x2+=((float)x[m])*x[m];
|
||||
y2+=((float)y[m])*y[m];
|
||||
}
|
||||
den = (float)sqrt(x2*y2);
|
||||
return num/den;
|
||||
}
|
||||
|
||||
static int PatternMatch(int8_t *y){
|
||||
float maxCn = -999999.0; /* large negative number */
|
||||
int bestmatch = 0;
|
||||
float Cn;
|
||||
int n;
|
||||
for (n=0;n<CVSD_N;n++){
|
||||
Cn = CrossCorrelation(&y[CVSD_LHIST-CVSD_M] /* x */, &y[n]);
|
||||
if (Cn>maxCn){
|
||||
bestmatch=n;
|
||||
maxCn = Cn;
|
||||
}
|
||||
}
|
||||
return bestmatch;
|
||||
}
|
||||
|
||||
static float AmplitudeMatch(int8_t *y, int8_t bestmatch) {
|
||||
int i;
|
||||
float sumx = 0;
|
||||
float sumy = 0.000001f;
|
||||
float sf;
|
||||
|
||||
for (i=0;i<CVSD_FS;i++){
|
||||
sumx += abs(y[CVSD_LHIST-CVSD_FS+i]);
|
||||
sumy += abs(y[bestmatch+i]);
|
||||
}
|
||||
sf = sumx/sumy;
|
||||
/* This is not in the paper, but limit the scaling factor to something reasonable to avoid creating artifacts */
|
||||
if (sf<0.75f) sf=0.75f;
|
||||
if (sf>1.2f) sf=1.2f;
|
||||
return sf;
|
||||
}
|
||||
|
||||
static int8_t crop_to_int8(float val){
|
||||
float croped_val = 0;
|
||||
@ -109,12 +152,14 @@ void btstack_cvsd_plc_bad_frame(btstack_cvsd_plc_state_t *plc_state, int8_t *out
|
||||
plc_state->hist[CVSD_LHIST+i] = crop_to_int8(val);
|
||||
}
|
||||
|
||||
for (i=CVSD_FS+CVSD_OLAL;i<CVSD_FS+CVSD_OLAL+CVSD_RT;i++)
|
||||
for (i=CVSD_FS+CVSD_OLAL;i<CVSD_FS+CVSD_OLAL+CVSD_RT;i++){
|
||||
plc_state->hist[CVSD_LHIST+i] = plc_state->hist[plc_state->bestlag+i];
|
||||
}
|
||||
|
||||
} else {
|
||||
for (i=0;i<CVSD_FS+CVSD_RT+CVSD_OLAL;i++)
|
||||
for (i=0;i<CVSD_FS+CVSD_RT+CVSD_OLAL;i++){
|
||||
plc_state->hist[CVSD_LHIST+i] = plc_state->hist[plc_state->bestlag+i];
|
||||
}
|
||||
}
|
||||
|
||||
for (i=0;i<CVSD_FS;i++){
|
||||
@ -160,51 +205,3 @@ void btstack_cvsd_plc_good_frame(btstack_cvsd_plc_state_t *plc_state, int8_t *in
|
||||
plc_state->nbf=0;
|
||||
}
|
||||
|
||||
|
||||
float CrossCorrelation(int8_t *x, int8_t *y){
|
||||
float num = 0;
|
||||
float den = 0;
|
||||
float x2 = 0;
|
||||
float y2 = 0;
|
||||
int m;
|
||||
for (m=0;m<CVSD_M;m++){
|
||||
num+=((float)x[m])*y[m];
|
||||
x2+=((float)x[m])*x[m];
|
||||
y2+=((float)y[m])*y[m];
|
||||
}
|
||||
den = (float)sqrt(x2*y2);
|
||||
return num/den;
|
||||
}
|
||||
|
||||
int PatternMatch(int8_t *y){
|
||||
float maxCn = -999999.0; /* large negative number */
|
||||
int bestmatch = 0;
|
||||
float Cn;
|
||||
int n;
|
||||
for (n=0;n<CVSD_N;n++){
|
||||
Cn = CrossCorrelation(&y[CVSD_LHIST-CVSD_M] /* x */, &y[n]);
|
||||
if (Cn>maxCn){
|
||||
bestmatch=n;
|
||||
maxCn = Cn;
|
||||
}
|
||||
}
|
||||
return bestmatch;
|
||||
}
|
||||
|
||||
|
||||
float AmplitudeMatch(int8_t *y, int8_t bestmatch) {
|
||||
int i;
|
||||
float sumx = 0;
|
||||
float sumy = 0.000001f;
|
||||
float sf;
|
||||
|
||||
for (i=0;i<CVSD_FS;i++){
|
||||
sumx += abs(y[CVSD_LHIST-CVSD_FS+i]);
|
||||
sumy += abs(y[bestmatch+i]);
|
||||
}
|
||||
sf = sumx/sumy;
|
||||
/* This is not in the paper, but limit the scaling factor to something reasonable to avoid creating artifacts */
|
||||
if (sf<0.75f) sf=0.75f;
|
||||
if (sf>1.2f) sf=1.2f;
|
||||
return sf;
|
||||
}
|
@ -63,9 +63,53 @@ static float rcos[SBC_OLAL] = {
|
||||
0.45386582f,0.36316850f,0.27713082f,0.19868268f,
|
||||
0.13049554f,0.07489143f,0.03376389f,0.00851345f};
|
||||
|
||||
static float CrossCorrelation(int16_t *x, int16_t *y);
|
||||
static int PatternMatch(int16_t *y);
|
||||
static float AmplitudeMatch(int16_t *y, int16_t bestmatch);
|
||||
static float CrossCorrelation(int16_t *x, int16_t *y){
|
||||
float num = 0;
|
||||
float den = 0;
|
||||
float x2 = 0;
|
||||
float y2 = 0;
|
||||
int m;
|
||||
for (m=0;m<SBC_M;m++){
|
||||
num+=((float)x[m])*y[m];
|
||||
x2+=((float)x[m])*x[m];
|
||||
y2+=((float)y[m])*y[m];
|
||||
}
|
||||
den = (float)sqrt(x2*y2);
|
||||
return num/den;
|
||||
}
|
||||
|
||||
static int PatternMatch(int16_t *y){
|
||||
float maxCn = -999999.0; /* large negative number */
|
||||
int bestmatch = 0;
|
||||
float Cn;
|
||||
int n;
|
||||
for (n=0;n<SBC_N;n++){
|
||||
Cn = CrossCorrelation(&y[SBC_LHIST-SBC_M] /* x */, &y[n]);
|
||||
if (Cn>maxCn){
|
||||
bestmatch=n;
|
||||
maxCn = Cn;
|
||||
}
|
||||
}
|
||||
return bestmatch;
|
||||
}
|
||||
|
||||
|
||||
static float AmplitudeMatch(int16_t *y, int16_t bestmatch) {
|
||||
int i;
|
||||
float sumx = 0;
|
||||
float sumy = 0.000001f;
|
||||
float sf;
|
||||
|
||||
for (i=0;i<SBC_FS;i++){
|
||||
sumx += abs(y[SBC_LHIST-SBC_FS+i]);
|
||||
sumy += abs(y[bestmatch+i]);
|
||||
}
|
||||
sf = sumx/sumy;
|
||||
/* This is not in the paper, but limit the scaling factor to something reasonable to avoid creating artifacts */
|
||||
if (sf<0.75f) sf=0.75f;
|
||||
if (sf>1.2f) sf=1.2f;
|
||||
return sf;
|
||||
}
|
||||
|
||||
static int16_t crop_to_int16(float val){
|
||||
float croped_val = 0;
|
||||
@ -123,8 +167,9 @@ void btstack_sbc_plc_bad_frame(btstack_sbc_plc_state_t *plc_state, int16_t *ZIRb
|
||||
}
|
||||
|
||||
} else {
|
||||
for (i=0;i<SBC_FS+SBC_RT+SBC_OLAL;i++)
|
||||
for (i=0;i<SBC_FS+SBC_RT+SBC_OLAL;i++){
|
||||
plc_state->hist[SBC_LHIST+i] = plc_state->hist[plc_state->bestlag+i];
|
||||
}
|
||||
}
|
||||
for (i=0;i<SBC_FS;i++){
|
||||
out[i] = plc_state->hist[SBC_LHIST+i];
|
||||
@ -167,52 +212,3 @@ void btstack_sbc_plc_good_frame(btstack_sbc_plc_state_t *plc_state, int16_t *in,
|
||||
|
||||
plc_state->nbf=0;
|
||||
}
|
||||
|
||||
|
||||
float CrossCorrelation(int16_t *x, int16_t *y){
|
||||
float num = 0;
|
||||
float den = 0;
|
||||
float x2 = 0;
|
||||
float y2 = 0;
|
||||
int m;
|
||||
for (m=0;m<SBC_M;m++){
|
||||
num+=((float)x[m])*y[m];
|
||||
x2+=((float)x[m])*x[m];
|
||||
y2+=((float)y[m])*y[m];
|
||||
}
|
||||
den = (float)sqrt(x2*y2);
|
||||
return num/den;
|
||||
}
|
||||
|
||||
int PatternMatch(int16_t *y){
|
||||
float maxCn = -999999.0; /* large negative number */
|
||||
int bestmatch = 0;
|
||||
float Cn;
|
||||
int n;
|
||||
for (n=0;n<SBC_N;n++){
|
||||
Cn = CrossCorrelation(&y[SBC_LHIST-SBC_M] /* x */, &y[n]);
|
||||
if (Cn>maxCn){
|
||||
bestmatch=n;
|
||||
maxCn = Cn;
|
||||
}
|
||||
}
|
||||
return bestmatch;
|
||||
}
|
||||
|
||||
|
||||
float AmplitudeMatch(int16_t *y, int16_t bestmatch) {
|
||||
int i;
|
||||
float sumx = 0;
|
||||
float sumy = 0.000001f;
|
||||
float sf;
|
||||
|
||||
for (i=0;i<SBC_FS;i++){
|
||||
sumx += abs(y[SBC_LHIST-SBC_FS+i]);
|
||||
sumy += abs(y[bestmatch+i]);
|
||||
}
|
||||
sf = sumx/sumy;
|
||||
/* This is not in the paper, but limit the scaling factor to something reasonable to avoid creating artifacts */
|
||||
if (sf<0.75f) sf=0.75f;
|
||||
if (sf>1.2f) sf=1.2f;
|
||||
return sf;
|
||||
}
|
Loading…
x
Reference in New Issue
Block a user