acelp_vectors.h File Reference

#include <stdint.h>

Go to the source code of this file.

Data Structures
struct	ACELPVContext
struct	AMRFixed
	Sparse representation for the algebraic codebook (fixed) vector. More...

Functions
void	ff_acelp_vectors_init (ACELPVContext *c)
	Initialize ACELPVContext.
void	ff_acelp_vectors_init_mips (ACELPVContext *c)
void	ff_acelp_fc_pulse_per_track (int16_t *fc_v, const uint8_t *tab1, const uint8_t *tab2, int pulse_indexes, int pulse_signs, int pulse_count, int bits)
	Decode fixed-codebook vector (3.8 and D.5.8 of G.729, 5.7.1 of AMR).
void	ff_decode_10_pulses_35bits (const int16_t *fixed_index, AMRFixed *fixed_sparse, const uint8_t *gray_decode, int half_pulse_count, int bits)
	Decode the algebraic codebook index to pulse positions and signs and construct the algebraic codebook vector for MODE_12k2.
void	ff_acelp_weighted_vector_sum (int16_t *out, const int16_t *in_a, const int16_t *in_b, int16_t weight_coeff_a, int16_t weight_coeff_b, int16_t rounder, int shift, int length)
	weighted sum of two vectors with rounding.
void	ff_weighted_vector_sumf (float *out, const float *in_a, const float *in_b, float weight_coeff_a, float weight_coeff_b, int length)
	float implementation of weighted sum of two vectors.
void	ff_adaptive_gain_control (float *out, const float *in, float speech_energ, int size, float alpha, float *gain_mem)
	Adaptive gain control (as used in AMR postfiltering)
void	ff_scale_vector_to_given_sum_of_squares (float *out, const float *in, float sum_of_squares, const int n)
	Set the sum of squares of a signal by scaling.
void	ff_set_fixed_vector (float *out, const AMRFixed *in, float scale, int size)
	Add fixed vector to an array from a sparse representation.
void	ff_clear_fixed_vector (float *out, const AMRFixed *in, int size)
	Clear array values set by set_fixed_vector.

Variables
const uint8_t	ff_fc_4pulses_8bits_tracks_13 [16]
	Track|Pulse| Positions 1 | 0 | 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75 2 | 1 | 1, 6, 11, 16, 21, 26, 31, 36, 41, 46, 51, 56, 61, 66, 71, 76 3 | 2 | 2, 7, 12, 17, 22, 27, 32, 37, 42, 47, 52, 57, 62, 67, 72, 77
const uint8_t	ff_fc_4pulses_8bits_track_4 [32]
	Track|Pulse| Positions 4 | 3 | 3, 8, 13, 18, 23, 28, 33, 38, 43, 48, 53, 58, 63, 68, 73, 78 | | 4, 9, 14, 19, 24, 29, 34, 39, 44, 49, 54, 59, 64, 69, 74, 79
const uint8_t	ff_fc_2pulses_9bits_track1 [16]
	Track|Pulse| Positions 1 | 0 | 1, 6, 11, 16, 21, 26, 31, 36 | | 3, 8, 13, 18, 23, 28, 33, 38
const uint8_t	ff_fc_2pulses_9bits_track1_gray [16]
const uint8_t	ff_fc_2pulses_9bits_track2_gray [32]
	Track|Pulse| Positions 2 | 1 | 0, 7, 14, 20, 27, 34, 1, 21 | | 2, 9, 15, 22, 29, 35, 6, 26 | | 4,10, 17, 24, 30, 37, 11, 31 | | 5,12, 19, 25, 32, 39, 16, 36
const float	ff_b60_sinc [61]
	b60 hamming windowed sinc function coefficients
const float	ff_pow_0_7 [10]
	Table of pow(0.7,n)
const float	ff_pow_0_75 [10]
	Table of pow(0.75,n)
const float	ff_pow_0_55 [10]
	Table of pow(0.55,n)

Function Documentation

void ff_acelp_vectors_init

(

ACELPVContext *

c

)

Initialize ACELPVContext.

Definition at line 266 of file acelp_vectors.c.

Referenced by amrnb_decode_init(), and amrwb_decode_init().

void ff_acelp_vectors_init_mips

(

ACELPVContext *

c

)

Definition at line 97 of file acelp_vectors_mips.c.

Referenced by ff_acelp_vectors_init().

void ff_acelp_fc_pulse_per_track	(	int16_t *	fc_v,
		const uint8_t *	tab1,
		const uint8_t *	tab2,
		int	pulse_indexes,
		int	pulse_signs,
		int	pulse_count,
		int	bits
	)

Decode fixed-codebook vector (3.8 and D.5.8 of G.729, 5.7.1 of AMR).

Parameters

[out]	fc_v	decoded fixed codebook vector (2.13)
	tab1	table used for first pulse_count pulses
	tab2	table used for last pulse
	pulse_indexes	fixed codebook indexes
	pulse_signs	signs of the excitation pulses (0 bit value means negative sign)
	bits	number of bits per one pulse index
	pulse_count	number of pulses decoded using first table
	bits	length of one pulse index in bits

Used in G.729 @8k, G.729 @4.4k, G.729 @6.4k, AMR @7.95k, AMR @7.40k

Definition at line 127 of file acelp_vectors.c.

Referenced by decode_frame().

void ff_decode_10_pulses_35bits	(	const int16_t *	fixed_index,
		AMRFixed *	fixed_sparse,
		const uint8_t *	gray_decode,
		int	half_pulse_count,
		int	bits
	)

Decode the algebraic codebook index to pulse positions and signs and construct the algebraic codebook vector for MODE_12k2.

Note

: The positions and signs are explicitly coded in MODE_12k2.

Parameters

fixed_index	positions of the ten pulses
fixed_sparse	pointer to the algebraic codebook vector
gray_decode	gray decoding table
half_pulse_count	number of couples of pulses
bits	length of one pulse index in bits

Definition at line 151 of file acelp_vectors.c.

Referenced by decode_fixed_sparse(), and ff_sipr_decode_frame_16k().

void ff_acelp_weighted_vector_sum	(	int16_t *	out,
		const int16_t *	in_a,
		const int16_t *	in_b,
		int16_t	weight_coeff_a,
		int16_t	weight_coeff_b,
		int16_t	rounder,
		int	shift,
		int	length
	)

weighted sum of two vectors with rounding.

Parameters

[out]	out	result of addition
	in_a	first vector
	in_b	second vector
	weight_coeff_a	first vector weight coefficient
	weight_coeff_a	second vector weight coefficient
	rounder	this value will be added to the sum of the two vectors
	shift	result will be shifted to right by this value
	length	vectors length

Note

It is safe to pass the same buffer for out and in_a or in_b.

out[i] = (in_a[i]*weight_a + in_b[i]*weight_b + rounder) >> shift

Definition at line 172 of file acelp_vectors.c.

Referenced by decode_frame(), g723_1_decode_frame(), long_term_filter(), and lsp_interpolate().

void ff_weighted_vector_sumf	(	float *	out,
		const float *	in_a,
		const float *	in_b,
		float	weight_coeff_a,
		float	weight_coeff_b,
		int	length
	)

float implementation of weighted sum of two vectors.

Parameters

[out]	out	result of addition
	in_a	first vector
	in_b	second vector
	weight_coeff_a	first vector weight coefficient
	weight_coeff_a	second vector weight coefficient
	length	vectors length

Note

It is safe to pass the same buffer for out and in_a or in_b.

Definition at line 192 of file acelp_vectors.c.

Referenced by decode_frame(), decode_lspf(), ff_acelp_vectors_init(), ff_sipr_decode_frame_16k(), interpolate_lpc(), interpolate_lsp(), and synth_block_fcb_acb().

void ff_adaptive_gain_control	(	float *	out,
		const float *	in,
		float	speech_energ,
		int	size,
		float	alpha,
		float *	gain_mem
	)

Adaptive gain control (as used in AMR postfiltering)

Parameters

out	output buffer for filtered speech data
in	the input speech buffer (may be the same as out)
speech_energ	input energy
size	the input buffer size
alpha	exponential filter factor
gain_mem	a pointer to the filter memory (single float of size)

Definition at line 202 of file acelp_vectors.c.

Referenced by decode_frame(), and postfilter().

void ff_scale_vector_to_given_sum_of_squares	(	float *	out,
		const float *	in,
		float	sum_of_squares,
		const int	n
	)

Set the sum of squares of a signal by scaling.

Parameters

out	output samples
in	input samples
sum_of_squares	new sum of squares
n	number of samples

Note

If the input is zero (or its energy underflows), the output is zero. This is the behavior of AGC in the AMR reference decoder. The QCELP reference decoder seems to have undefined behavior.

TIA/EIA/IS-733 2.4.8.3-2/3/4/5, 2.4.8.6 3GPP TS 26.090 6.1 (6)

Definition at line 223 of file acelp_vectors.c.

Referenced by apply_gain_ctrl(), scaled_hb_excitation(), and synthesis().

void ff_set_fixed_vector	(	float *	out,
		const AMRFixed *	in,
		float	scale,
		int	size
	)

Add fixed vector to an array from a sparse representation.

Parameters

out	fixed vector with pitch sharpening
in	sparse fixed vector
scale	number to multiply the fixed vector by
size	the output vector size

Definition at line 234 of file acelp_vectors.c.

Referenced by amrnb_decode_frame(), ff_sipr_decode_frame_16k(), and synth_block_fcb_acb().

void ff_clear_fixed_vector	(	float *	out,
		const AMRFixed *	in,
		int	size
	)

Clear array values set by set_fixed_vector.

Parameters

out	fixed vector to be cleared
in	sparse fixed vector
size	the output vector size

Definition at line 251 of file acelp_vectors.c.

Referenced by amrnb_decode_frame().

Variable Documentation

const uint8_t ff_fc_4pulses_8bits_tracks_13[16]

Track|Pulse| Positions

1 | 0 | 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75

2 | 1 | 1, 6, 11, 16, 21, 26, 31, 36, 41, 46, 51, 56, 61, 66, 71, 76

3 | 2 | 2, 7, 12, 17, 22, 27, 32, 37, 42, 47, 52, 57, 62, 67, 72, 77

Table contains only first the pulse indexes.

Used in G.729 @8k, G.729 @4.4k, AMR @7.95k, AMR @7.40k

Definition at line 73 of file acelp_vectors.c.

Referenced by decode_frame(), and ff_sipr_decode_frame_16k().

const uint8_t ff_fc_4pulses_8bits_track_4[32]

Track|Pulse| Positions

4 | 3 | 3, 8, 13, 18, 23, 28, 33, 38, 43, 48, 53, 58, 63, 68, 73, 78

| | 4, 9, 14, 19, 24, 29, 34, 39, 44, 49, 54, 59, 64, 69, 74, 79

Remarks

Track in the table should be read top-to-bottom, left-to-right.

Used in G.729 @8k, G.729 @4.4k, AMR @7.95k, AMR @7.40k

Definition at line 78 of file acelp_vectors.c.

Referenced by decode_frame().

const uint8_t ff_fc_2pulses_9bits_track1[16]

Track|Pulse| Positions

1 | 0 | 1, 6, 11, 16, 21, 26, 31, 36

| | 3, 8, 13, 18, 23, 28, 33, 38

Remarks

Track in the table should be read top-to-bottom, left-to-right.

Note

(EE) Reference G.729D code also uses gray decoding for each pulse index before looking up the value in the table.

Used in G.729 @6.4k (with gray coding), AMR @5.9k (without gray coding)

Definition at line 30 of file acelp_vectors.c.

const uint8_t ff_fc_2pulses_9bits_track1_gray[16]

Definition at line 41 of file acelp_vectors.c.

Referenced by decode_frame().

const uint8_t ff_fc_2pulses_9bits_track2_gray[32]

Track|Pulse| Positions

2 | 1 | 0, 7, 14, 20, 27, 34, 1, 21 | | 2, 9, 15, 22, 29, 35, 6, 26 | | 4,10, 17, 24, 30, 37, 11, 31

| | 5,12, 19, 25, 32, 39, 16, 36

Remarks

Track in the table should be read top-to-bottom, left-to-right.

Note

(EE.1) This table (from the reference code) does not comply with the specification. The specification contains the following table:

Track|Pulse| Positions

2 | 1 | 0, 5, 10, 15, 20, 25, 30, 35 | | 1, 6, 11, 16, 21, 26, 31, 36 | | 2, 7, 12, 17, 22, 27, 32, 37 | | 4, 9, 14, 19, 24, 29, 34, 39

---

Note

(EE.2) Reference G.729D code also uses gray decoding for each pulse index before looking up the value in the table.

Used in G.729 @6.4k (with gray coding)

Definition at line 53 of file acelp_vectors.c.

Referenced by decode_frame().

const float ff_b60_sinc[61]

b60 hamming windowed sinc function coefficients

Definition at line 113 of file acelp_vectors.c.

Referenced by decode_frame(), and decode_pitch_vector().

const float ff_pow_0_7[10]

Table of pow(0.7,n)

Definition at line 98 of file acelp_vectors.c.

Referenced by eval_ir(), and postfilter().

const float ff_pow_0_75[10]

Table of pow(0.75,n)

Definition at line 103 of file acelp_vectors.c.

Referenced by postfilter(), and postfilter_5k0().

const float ff_pow_0_55[10]

Table of pow(0.55,n)

Definition at line 108 of file acelp_vectors.c.

Referenced by eval_ir(), and postfilter().