acelp_pitch_delay.c File Reference

#include "avcodec.h"
#include "acelp_pitch_delay.h"
#include "acelp_math.h"

Go to the source code of this file.

Functions
int	ff_acelp_decode_8bit_to_1st_delay3 (int ac_index)
	Decode pitch delay of the first subframe encoded by 8 bits with 1/3 resolution.
int	ff_acelp_decode_4bit_to_2nd_delay3 (int ac_index, int pitch_delay_min)
	Decode pitch delay with 1/3 precision.
int	ff_acelp_decode_5_6_bit_to_2nd_delay3 (int ac_index, int pitch_delay_min)
	Decode pitch delay of the second subframe encoded by 5 or 6 bits with 1/3 precision.
int	ff_acelp_decode_9bit_to_1st_delay6 (int ac_index)
	Decode pitch delay of the first subframe encoded by 9 bits with 1/6 precision.
int	ff_acelp_decode_6bit_to_2nd_delay6 (int ac_index, int pitch_delay_min)
	Decode pitch delay of the second subframe encoded by 6 bits with 1/6 precision.
void	ff_acelp_update_past_gain (int16_t *quant_energy, int gain_corr_factor, int log2_ma_pred_order, int erasure)
	Update past quantized energies.
int16_t	ff_acelp_decode_gain_code (int gain_corr_factor, const int16_t *fc_v, int mr_energy, const int16_t *quant_energy, const int16_t *ma_prediction_coeff, int subframe_size, int ma_pred_order)
	Decode the adaptive codebook gain and add correction (4.1.5 and 3.9.1 of G.729).

---

Function Documentation

int ff_acelp_decode_4bit_to_2nd_delay3	(	int	ac_index,
		int	pitch_delay_min
	)

Decode pitch delay with 1/3 precision.

Parameters:

	ac_index	adaptive codebook index (4 bits)
	pitch_delay_min	lower bound (integer) of pitch delay interval for second subframe

Returns:

pitch delay in 1/3 units

Pitch delay is coded: integers only, -6 < pitch_delay - int(prev_pitch_delay) <= -2 with 1/3 resolution, -2 < pitch_delay - int(prev_pitch_delay) < 1 integers only, 1 <= pitch_delay - int(prev_pitch_delay) < 5

Remarks:

The routine is used in G.729 .4k, AMR .7k, AMR .9k, AMR .15k, AMR .75k for the second subframe.

Definition at line 35 of file acelp_pitch_delay.c.

{
    if(ac_index < 4)
        return 3 * (ac_index + pitch_delay_min);
    else if(ac_index < 12)
        return 3 * pitch_delay_min + ac_index + 6;
    else
        return 3 * (ac_index + pitch_delay_min) - 18;
}

int ff_acelp_decode_5_6_bit_to_2nd_delay3	(	int	ac_index,
		int	pitch_delay_min
	)

Decode pitch delay of the second subframe encoded by 5 or 6 bits with 1/3 precision.

Parameters:

	ac_index	adaptive codebook index (5 or 6 bits)
	pitch_delay_min	lower bound (integer) of pitch delay interval for second subframe

Returns:

pitch delay in 1/3 units

Pitch delay is coded: with 1/3 resolution, -6 < pitch_delay - int(prev_pitch_delay) < 5

Remarks:

The routine is used in G.729 , AMR .2k, AMR .95k, AMR .4k for the second subframe.

Definition at line 47 of file acelp_pitch_delay.c.

{
        return 3 * pitch_delay_min + ac_index - 2;
}

int ff_acelp_decode_6bit_to_2nd_delay6	(	int	ac_index,
		int	pitch_delay_min
	)

Decode pitch delay of the second subframe encoded by 6 bits with 1/6 precision.

Parameters:

	ac_index	adaptive codebook index (6 bits)
	pitch_delay_min	lower bound (integer) of pitch delay interval for second subframe

Returns:

pitch delay in 1/6 units

Pitch delay is coded: with 1/6 resolution, -6 < pitch_delay - int(prev_pitch_delay) < 5

Remarks:

The routine is used in AMR .2k for the second and fourth subframes.

Definition at line 61 of file acelp_pitch_delay.c.

{
    return 6 * pitch_delay_min + ac_index - 3;
}

int ff_acelp_decode_8bit_to_1st_delay3

(

int

ac_index

)

Decode pitch delay of the first subframe encoded by 8 bits with 1/3 resolution.

Parameters:

ac_index

adaptive codebook index (8 bits)

Returns:

pitch delay in 1/3 units

Pitch delay is coded: with 1/3 resolution, 19 < pitch_delay < 85 integers only, 85 <= pitch_delay <= 143

Definition at line 27 of file acelp_pitch_delay.c.

{
    ac_index += 58;
    if(ac_index > 254)
        ac_index = 3 * ac_index - 510;
    return ac_index;
}

int ff_acelp_decode_9bit_to_1st_delay6

(

int

ac_index

)

Decode pitch delay of the first subframe encoded by 9 bits with 1/6 precision.

Parameters:

	ac_index	adaptive codebook index (9 bits)
	pitch_delay_min	lower bound (integer) of pitch delay interval for second subframe

Returns:

pitch delay in 1/6 units

Pitch delay is coded: with 1/6 resolution, 17 < pitch_delay < 95 integers only, 95 <= pitch_delay <= 143

Remarks:

The routine is used in AMR .2k for the first and third subframes.

Definition at line 54 of file acelp_pitch_delay.c.

{
    if(ac_index < 463)
        return ac_index + 105;
    else
        return 6 * (ac_index - 368);
}

int16_t ff_acelp_decode_gain_code	(	int	gain_corr_factor,
		const int16_t *	fc_v,
		int	mr_energy,
		const int16_t *	quant_energy,
		const int16_t *	ma_prediction_coeff,
		int	subframe_size,
		int	max_pred_order
	)

Decode the adaptive codebook gain and add correction (4.1.5 and 3.9.1 of G.729).

Parameters:

	gain_corr_factor	gain correction factor (2.13)
	fc_v	fixed-codebook vector (2.13)
	mr_energy	mean innovation energy and fixed-point correction (7.13)
	quant_energy	[in/out] past quantized energies (5.10)
	subframe_size	length of subframe
	ma_pred_order	MA prediction order

Returns:

quantized fixed-codebook gain (14.1)

The routine implements equations 69, 66 and 71 of the G.729 specification (3.9.1)

Em - mean innovation energy (dB, constant, depends on decoding algorithm) Ep - mean-removed predicted energy (dB) Er - mean-removed innovation energy (dB) Ei - mean energy of the fixed-codebook contribution (dB) N - subframe_size M - MA (Moving Average) prediction order gc - fixed-codebook gain gc_p - predicted fixed-codebook gain

Fixed codebook gain is computed using predicted gain gc_p and correction factor gain_corr_factor as shown below:

gc = gc_p * gain_corr_factor

The predicted fixed codebook gain gc_p is found by predicting the energy of the fixed-codebook contribution from the energy of previous fixed-codebook contributions.

mean = 1/N * sum(i,0,N){ fc_v[i] * fc_v[i] }

Ei = 10log(mean)

Er = 10log(1/N * gc^2 * mean) - Em = 20log(gc) + Ei - Em

Replacing Er with Ep and gc with gc_p we will receive:

Ep = 10log(1/N * gc_p^2 * mean) - Em = 20log(gc_p) + Ei - Em

and from above:

gc_p = 10^((Ep - Ei + Em) / 20)

Ep is predicted using past energies and prediction coefficients:

Ep = sum(i,0,M){ ma_prediction_coeff[i] * quant_energy[i] }

gc_p in fixed-point arithmetic is calculated as following:

mean = 1/N * sum(i,0,N){ (fc_v[i] / 2^13) * (fc_v[i] / 2^13) } = = 1/N * sum(i,0,N) { fc_v[i] * fc_v[i] } / 2^26

Ei = 10log(mean) = -10log(N) - 10log(2^26) + + 10log(sum(i,0,N) { fc_v[i] * fc_v[i] })

Ep - Ei + Em = Ep + Em + 10log(N) + 10log(2^26) -

• 10log(sum(i,0,N) { fc_v[i] * fc_v[i] }) = = Ep + mr_energy - 10log(sum(i,0,N) { fc_v[i] * fc_v[i] })

gc_p = 10 ^ ((Ep - Ei + Em) / 20) = = 2 ^ (3.3219 * (Ep - Ei + Em) / 20) = 2 ^ (0.166 * (Ep - Ei + Em))

where

mr_energy = Em + 10log(N) + 10log(2^26)

Remarks:

The routine is used in G.729 and AMR (all modes).

Definition at line 89 of file acelp_pitch_delay.c.

References bidir_sal(), dot_product(), ff_exp2(), and ff_log2().

{
    int i;

    mr_energy <<= 10;

    for(i=0; i<ma_pred_order; i++)
        mr_energy += quant_energy[i] * ma_prediction_coeff[i];

#ifdef G729_BITEXACT
    mr_energy += (((-6165LL * ff_log2(dot_product(fc_v, fc_v, subframe_size, 0))) >> 3) & ~0x3ff);

    mr_energy = (5439 * (mr_energy >> 15)) >> 8;           // (0.15) = (0.15) * (7.23)

    return bidir_sal(
               ((ff_exp2(mr_energy & 0x7fff) + 16) >> 5) * (gain_corr_factor >> 1),
               (mr_energy >> 15) - 25
           );
#else
    mr_energy = gain_corr_factor * exp(M_LN10 / (20 << 23) * mr_energy) /
                sqrt(dot_product(fc_v, fc_v, subframe_size, 0));
    return mr_energy >> 12;
#endif
}

void ff_acelp_update_past_gain	(	int16_t *	quant_energy,
		int	gain_corr_factor,
		int	log2_ma_pred_order,
		int	erasure
	)

Update past quantized energies.

Parameters:

	quant_energy	[in/out] past quantized energies (5.10)
	gain_corr_factor	gain correction factor
	log2_ma_pred_order	log2() of MA prediction order
	erasure	frame erasure flag

If frame erasure flag is not equal to zero, memory is updated with averaged energy, attenuated by 4dB: max(avg(quant_energy[i])-4, -14), i=0,ma_pred_order

In normal mode memory is updated with Er - Ep = 20 * log10(gain_corr_factor)

Remarks:

The routine is used in G.729 and AMR (all modes).

Definition at line 68 of file acelp_pitch_delay.c.

References ff_log2(), and FFMAX.

{
    int i;
    int avg_gain=quant_energy[(1 << log2_ma_pred_order) - 1]; // (5.10)

    for(i=(1 << log2_ma_pred_order) - 1; i>0; i--)
    {
        avg_gain       += quant_energy[i-1];
        quant_energy[i] = quant_energy[i-1];
    }

    if(erasure)
        quant_energy[0] = FFMAX(avg_gain >> log2_ma_pred_order, -10240) - 4096; // -10 and -4 in (5.10)
    else
        quant_energy[0] = (6165 * ((ff_log2(gain_corr_factor) >> 2) - (13 << 13))) >> 13;
}

---