FFmpeg
|
#include <ac3dsp.h>
Data Fields | |
void(* | ac3_exponent_min )(uint8_t *exp, int num_reuse_blocks, int nb_coefs) |
Set each encoded exponent in a block to the minimum of itself and the exponents in the same frequency bin of up to 5 following blocks. More... | |
int(* | ac3_max_msb_abs_int16 )(const int16_t *src, int len) |
Calculate the maximum MSB of the absolute value of each element in an array of int16_t. More... | |
void(* | ac3_lshift_int16 )(int16_t *src, unsigned int len, unsigned int shift) |
Left-shift each value in an array of int16_t by a specified amount. More... | |
void(* | ac3_rshift_int32 )(int32_t *src, unsigned int len, unsigned int shift) |
Right-shift each value in an array of int32_t by a specified amount. More... | |
void(* | float_to_fixed24 )(int32_t *dst, const float *src, unsigned int len) |
Convert an array of float in range [-1.0,1.0] to int32_t with range [-(1<<24),(1<<24)]. More... | |
void(* | bit_alloc_calc_bap )(int16_t *mask, int16_t *psd, int start, int end, int snr_offset, int floor, const uint8_t *bap_tab, uint8_t *bap) |
Calculate bit allocation pointers. More... | |
void(* | update_bap_counts )(uint16_t mant_cnt[16], uint8_t *bap, int len) |
Update bap counts using the supplied array of bap. More... | |
int(* | compute_mantissa_size )(uint16_t mant_cnt[6][16]) |
Calculate the number of bits needed to encode a set of mantissas. More... | |
void(* | extract_exponents )(uint8_t *exp, int32_t *coef, int nb_coefs) |
void(* | sum_square_butterfly_int32 )(int64_t sum[4], const int32_t *coef0, const int32_t *coef1, int len) |
void(* | sum_square_butterfly_float )(float sum[4], const float *coef0, const float *coef1, int len) |
int | out_channels |
int | in_channels |
void(* | downmix )(float **samples, float **matrix, int len) |
void(* | downmix_fixed )(int32_t **samples, int16_t **matrix, int len) |
void(* | apply_window_int16 )(int16_t *output, const int16_t *input, const int16_t *window, unsigned int len) |
Apply symmetric window in 16-bit fixed-point. More... | |
Set each encoded exponent in a block to the minimum of itself and the exponents in the same frequency bin of up to 5 following blocks.
exp | pointer to the start of the current block of exponents. constraints: align 16 |
num_reuse_blocks | number of blocks that will reuse exponents from the current block. constraints: range 0 to 5 |
nb_coefs | number of frequency coefficients. |
Calculate the maximum MSB of the absolute value of each element in an array of int16_t.
src | input array constraints: align 16. values must be in range [-32767,32767] |
len | number of values in the array constraints: multiple of 16 greater than 0 |
Convert an array of float in range [-1.0,1.0] to int32_t with range [-(1<<24),(1<<24)].
dst | destination array of int32_t. constraints: 16-byte aligned |
src | source array of float. constraints: 16-byte aligned |
len | number of elements to convert. constraints: multiple of 32 greater than zero |
void(* AC3DSPContext::bit_alloc_calc_bap) (int16_t *mask, int16_t *psd, int start, int end, int snr_offset, int floor, const uint8_t *bap_tab, uint8_t *bap) |
Calculate bit allocation pointers.
The SNR is the difference between the masking curve and the signal. AC-3 uses this value for each frequency bin to allocate bits. The snroffset parameter is a global adjustment to the SNR for all bins.
[in] | mask | masking curve |
[in] | psd | signal power for each frequency bin |
[in] | start | starting bin location |
[in] | end | ending bin location |
[in] | snr_offset | SNR adjustment |
[in] | floor | noise floor |
[in] | bap_tab | look-up table for bit allocation pointers |
[out] | bap | bit allocation pointers |
int(* AC3DSPContext::compute_mantissa_size) (uint16_t mant_cnt[6][16]) |
void(* AC3DSPContext::apply_window_int16) (int16_t *output, const int16_t *input, const int16_t *window, unsigned int len) |
Apply symmetric window in 16-bit fixed-point.
output | destination array constraints: 16-byte aligned |
input | source array constraints: 16-byte aligned |
window | window array constraints: 16-byte aligned, at least len/2 elements |
len | full window length constraints: multiple of ? greater than zero |