mirror of
https://github.com/claunia/flac.git
synced 2025-12-16 18:54:26 +00:00
revamp the ordinal types
This commit is contained in:
@@ -30,11 +30,11 @@
|
||||
|
||||
#define LOCAL_FABS(x) ((x)<0.0? -(x):(x))
|
||||
|
||||
void FLAC__lpc_compute_autocorrelation(const real data[], unsigned data_len, unsigned lag, real autoc[])
|
||||
void FLAC__lpc_compute_autocorrelation(const FLAC__real data[], unsigned data_len, unsigned lag, FLAC__real autoc[])
|
||||
{
|
||||
/* a readable, but slower, version */
|
||||
#if 0
|
||||
real d;
|
||||
FLAC__real d;
|
||||
unsigned i;
|
||||
|
||||
FLAC__ASSERT(lag > 0);
|
||||
@@ -51,7 +51,7 @@ void FLAC__lpc_compute_autocorrelation(const real data[], unsigned data_len, uns
|
||||
* this version tends to run faster because of better data locality
|
||||
* ('data_len' is usually much larger than 'lag')
|
||||
*/
|
||||
real d;
|
||||
FLAC__real d;
|
||||
unsigned sample, coeff;
|
||||
const unsigned limit = data_len - lag;
|
||||
|
||||
@@ -72,10 +72,10 @@ void FLAC__lpc_compute_autocorrelation(const real data[], unsigned data_len, uns
|
||||
}
|
||||
}
|
||||
|
||||
void FLAC__lpc_compute_lp_coefficients(const real autoc[], unsigned max_order, real lp_coeff[][FLAC__MAX_LPC_ORDER], real error[])
|
||||
void FLAC__lpc_compute_lp_coefficients(const FLAC__real autoc[], unsigned max_order, FLAC__real lp_coeff[][FLAC__MAX_LPC_ORDER], FLAC__real error[])
|
||||
{
|
||||
unsigned i, j;
|
||||
real r, err, ref[FLAC__MAX_LPC_ORDER], lpc[FLAC__MAX_LPC_ORDER];
|
||||
FLAC__real r, err, ref[FLAC__MAX_LPC_ORDER], lpc[FLAC__MAX_LPC_ORDER];
|
||||
|
||||
FLAC__ASSERT(0 < max_order);
|
||||
FLAC__ASSERT(max_order <= FLAC__MAX_LPC_ORDER);
|
||||
@@ -93,7 +93,7 @@ void FLAC__lpc_compute_lp_coefficients(const real autoc[], unsigned max_order, r
|
||||
/* Update LPC coefficients and total error. */
|
||||
lpc[i]=r;
|
||||
for(j = 0; j < (i>>1); j++) {
|
||||
real tmp = lpc[j];
|
||||
FLAC__real tmp = lpc[j];
|
||||
lpc[j] += r * lpc[i-1-j];
|
||||
lpc[i-1-j] += r * tmp;
|
||||
}
|
||||
@@ -109,16 +109,16 @@ void FLAC__lpc_compute_lp_coefficients(const real autoc[], unsigned max_order, r
|
||||
}
|
||||
}
|
||||
|
||||
int FLAC__lpc_quantize_coefficients(const real lp_coeff[], unsigned order, unsigned precision, unsigned bits_per_sample, int32 qlp_coeff[], int *shift)
|
||||
int FLAC__lpc_quantize_coefficients(const FLAC__real lp_coeff[], unsigned order, unsigned precision, unsigned bits_per_sample, FLAC__int32 qlp_coeff[], int *shift)
|
||||
{
|
||||
unsigned i;
|
||||
real d, cmax = -1e32;
|
||||
FLAC__real d, cmax = -1e32;
|
||||
|
||||
FLAC__ASSERT(bits_per_sample > 0);
|
||||
FLAC__ASSERT(bits_per_sample <= sizeof(int32)*8);
|
||||
FLAC__ASSERT(bits_per_sample <= sizeof(FLAC__int32)*8);
|
||||
FLAC__ASSERT(precision > 0);
|
||||
FLAC__ASSERT(precision >= FLAC__MIN_QLP_COEFF_PRECISION);
|
||||
FLAC__ASSERT(precision + bits_per_sample < sizeof(int32)*8);
|
||||
FLAC__ASSERT(precision + bits_per_sample < sizeof(FLAC__int32)*8);
|
||||
#ifdef NDEBUG
|
||||
(void)bits_per_sample; /* silence compiler warning about unused parameter */
|
||||
#endif
|
||||
@@ -151,19 +151,19 @@ int FLAC__lpc_quantize_coefficients(const real lp_coeff[], unsigned order, unsig
|
||||
|
||||
if(*shift != 0) { /* just to avoid wasting time... */
|
||||
for(i = 0; i < order; i++)
|
||||
qlp_coeff[i] = (int32)floor(lp_coeff[i] * (real)(1 << *shift));
|
||||
qlp_coeff[i] = (FLAC__int32)floor(lp_coeff[i] * (FLAC__real)(1 << *shift));
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
void FLAC__lpc_compute_residual_from_qlp_coefficients(const int32 data[], unsigned data_len, const int32 qlp_coeff[], unsigned order, int lp_quantization, int32 residual[])
|
||||
void FLAC__lpc_compute_residual_from_qlp_coefficients(const FLAC__int32 data[], unsigned data_len, const FLAC__int32 qlp_coeff[], unsigned order, int lp_quantization, FLAC__int32 residual[])
|
||||
{
|
||||
#ifdef FLAC__OVERFLOW_DETECT
|
||||
int64 sumo;
|
||||
FLAC__int64 sumo;
|
||||
#endif
|
||||
unsigned i, j;
|
||||
int32 sum;
|
||||
const int32 *history;
|
||||
FLAC__int32 sum;
|
||||
const FLAC__int32 *history;
|
||||
|
||||
#ifdef FLAC__OVERFLOW_DETECT_VERBOSE
|
||||
fprintf(stderr,"FLAC__lpc_compute_residual_from_qlp_coefficients: data_len=%d, order=%u, lpq=%d",data_len,order,lp_quantization);
|
||||
@@ -182,7 +182,7 @@ void FLAC__lpc_compute_residual_from_qlp_coefficients(const int32 data[], unsign
|
||||
for(j = 0; j < order; j++) {
|
||||
sum += qlp_coeff[j] * (*(--history));
|
||||
#ifdef FLAC__OVERFLOW_DETECT
|
||||
sumo += (int64)qlp_coeff[j] * (int64)(*history);
|
||||
sumo += (FLAC__int64)qlp_coeff[j] * (FLAC__int64)(*history);
|
||||
if(sumo > 2147483647ll || sumo < -2147483648ll) {
|
||||
fprintf(stderr,"FLAC__lpc_compute_residual_from_qlp_coefficients: OVERFLOW, i=%u, j=%u, c=%d, d=%d, sumo=%lld\n",i,j,qlp_coeff[j],*history,sumo);
|
||||
}
|
||||
@@ -201,14 +201,14 @@ void FLAC__lpc_compute_residual_from_qlp_coefficients(const int32 data[], unsign
|
||||
*/
|
||||
}
|
||||
|
||||
void FLAC__lpc_restore_signal(const int32 residual[], unsigned data_len, const int32 qlp_coeff[], unsigned order, int lp_quantization, int32 data[])
|
||||
void FLAC__lpc_restore_signal(const FLAC__int32 residual[], unsigned data_len, const FLAC__int32 qlp_coeff[], unsigned order, int lp_quantization, FLAC__int32 data[])
|
||||
{
|
||||
#ifdef FLAC__OVERFLOW_DETECT
|
||||
int64 sumo;
|
||||
FLAC__int64 sumo;
|
||||
#endif
|
||||
unsigned i, j;
|
||||
int32 sum;
|
||||
const int32 *history;
|
||||
FLAC__int32 sum;
|
||||
const FLAC__int32 *history;
|
||||
|
||||
#ifdef FLAC__OVERFLOW_DETECT_VERBOSE
|
||||
fprintf(stderr,"FLAC__lpc_restore_signal: data_len=%d, order=%u, lpq=%d",data_len,order,lp_quantization);
|
||||
@@ -227,7 +227,7 @@ void FLAC__lpc_restore_signal(const int32 residual[], unsigned data_len, const i
|
||||
for(j = 0; j < order; j++) {
|
||||
sum += qlp_coeff[j] * (*(--history));
|
||||
#ifdef FLAC__OVERFLOW_DETECT
|
||||
sumo += (int64)qlp_coeff[j] * (int64)(*history);
|
||||
sumo += (FLAC__int64)qlp_coeff[j] * (FLAC__int64)(*history);
|
||||
if(sumo > 2147483647ll || sumo < -2147483648ll) {
|
||||
fprintf(stderr,"FLAC__lpc_restore_signal: OVERFLOW, i=%u, j=%u, c=%d, d=%d, sumo=%lld\n",i,j,qlp_coeff[j],*history,sumo);
|
||||
}
|
||||
@@ -246,16 +246,16 @@ void FLAC__lpc_restore_signal(const int32 residual[], unsigned data_len, const i
|
||||
*/
|
||||
}
|
||||
|
||||
real FLAC__lpc_compute_expected_bits_per_residual_sample(real lpc_error, unsigned total_samples)
|
||||
FLAC__real FLAC__lpc_compute_expected_bits_per_residual_sample(FLAC__real lpc_error, unsigned total_samples)
|
||||
{
|
||||
real error_scale;
|
||||
FLAC__real error_scale;
|
||||
|
||||
FLAC__ASSERT(total_samples > 0);
|
||||
|
||||
error_scale = 0.5 * M_LN2 * M_LN2 / (real)total_samples;
|
||||
error_scale = 0.5 * M_LN2 * M_LN2 / (FLAC__real)total_samples;
|
||||
|
||||
if(lpc_error > 0.0) {
|
||||
real bps = 0.5 * log(error_scale * lpc_error) / M_LN2;
|
||||
FLAC__real bps = 0.5 * log(error_scale * lpc_error) / M_LN2;
|
||||
if(bps >= 0.0)
|
||||
return bps;
|
||||
else
|
||||
@@ -269,10 +269,10 @@ real FLAC__lpc_compute_expected_bits_per_residual_sample(real lpc_error, unsigne
|
||||
}
|
||||
}
|
||||
|
||||
real FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(real lpc_error, real error_scale)
|
||||
FLAC__real FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(FLAC__real lpc_error, FLAC__real error_scale)
|
||||
{
|
||||
if(lpc_error > 0.0) {
|
||||
real bps = 0.5 * log(error_scale * lpc_error) / M_LN2;
|
||||
FLAC__real bps = 0.5 * log(error_scale * lpc_error) / M_LN2;
|
||||
if(bps >= 0.0)
|
||||
return bps;
|
||||
else
|
||||
@@ -286,21 +286,21 @@ real FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(real l
|
||||
}
|
||||
}
|
||||
|
||||
unsigned FLAC__lpc_compute_best_order(const real lpc_error[], unsigned max_order, unsigned total_samples, unsigned bits_per_signal_sample)
|
||||
unsigned FLAC__lpc_compute_best_order(const FLAC__real lpc_error[], unsigned max_order, unsigned total_samples, unsigned bits_per_signal_sample)
|
||||
{
|
||||
unsigned order, best_order;
|
||||
real best_bits, tmp_bits, error_scale;
|
||||
FLAC__real best_bits, tmp_bits, error_scale;
|
||||
|
||||
FLAC__ASSERT(max_order > 0);
|
||||
FLAC__ASSERT(total_samples > 0);
|
||||
|
||||
error_scale = 0.5 * M_LN2 * M_LN2 / (real)total_samples;
|
||||
error_scale = 0.5 * M_LN2 * M_LN2 / (FLAC__real)total_samples;
|
||||
|
||||
best_order = 0;
|
||||
best_bits = FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(lpc_error[0], error_scale) * (real)total_samples;
|
||||
best_bits = FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(lpc_error[0], error_scale) * (FLAC__real)total_samples;
|
||||
|
||||
for(order = 1; order < max_order; order++) {
|
||||
tmp_bits = FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(lpc_error[order], error_scale) * (real)(total_samples - order) + (real)(order * bits_per_signal_sample);
|
||||
tmp_bits = FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale(lpc_error[order], error_scale) * (FLAC__real)(total_samples - order) + (FLAC__real)(order * bits_per_signal_sample);
|
||||
if(tmp_bits < best_bits) {
|
||||
best_order = order;
|
||||
best_bits = tmp_bits;
|
||||
|
||||
Reference in New Issue
Block a user