Commit dc7e07ac authored by Justin Ruggles's avatar Justin Ruggles

Add stereo rematrixing support to the AC-3 encoders.

This improves the audio quality significantly for stereo source with both the
fixed-point and floating-point AC-3 encoders.
Update acodec-ac3_fixed and seek-ac3_rm test references.

Originally committed as revision 26271 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent d267b339
......@@ -43,12 +43,6 @@ const uint8_t ff_ac3_ungroup_3_in_5_bits_tab[32][3] = {
{ 3, 0, 1 }, { 3, 0, 2 }, { 3, 1, 0 }, { 3, 1, 1 }
};
/**
* Table of bin locations for rematrixing bands
* reference: Section 7.5.2 Rematrixing : Frequency Band Definitions
*/
const uint8_t ff_ac3_rematrix_band_tab[5] = { 13, 25, 37, 61, 253 };
const uint8_t ff_eac3_hebap_tab[64] = {
0, 1, 2, 3, 4, 5, 6, 7, 8, 8,
8, 8, 9, 9, 9, 10, 10, 10, 10, 11,
......
......@@ -25,7 +25,6 @@
#include <stdint.h>
extern const uint8_t ff_ac3_ungroup_3_in_5_bits_tab[32][3];
extern const uint8_t ff_ac3_rematrix_band_tab[5];
extern const uint8_t ff_eac3_hebap_tab[64];
extern const uint8_t ff_eac3_default_cpl_band_struct[18];
......
......@@ -45,6 +45,12 @@
/** Maximum number of exponent groups. +1 for separate DC exponent. */
#define AC3_MAX_EXP_GROUPS 85
/* stereo rematrixing algorithms */
#define AC3_REMATRIXING_IS_STATIC 0x1
#define AC3_REMATRIXING_SUMS 0
#define AC3_REMATRIXING_NONE 1
#define AC3_REMATRIXING_ALWAYS 3
/** Scale a float value by 2^bits and convert to an integer. */
#define SCALE_FLOAT(a, bits) lrintf((a) * (float)(1 << (bits)))
......@@ -71,6 +77,8 @@ typedef struct AC3Block {
uint16_t **qmant; ///< quantized mantissas
uint8_t exp_strategy[AC3_MAX_CHANNELS]; ///< exponent strategies
int8_t exp_shift[AC3_MAX_CHANNELS]; ///< exponent shift values
uint8_t new_rematrixing_strategy; ///< send new rematrixing flags in this block
uint8_t rematrixing_flags[4]; ///< rematrixing flags
} AC3Block;
/**
......@@ -107,6 +115,8 @@ typedef struct AC3EncodeContext {
int bandwidth_code[AC3_MAX_CHANNELS]; ///< bandwidth code (0 to 60) (chbwcod)
int nb_coefs[AC3_MAX_CHANNELS];
int rematrixing; ///< determines how rematrixing strategy is calculated
/* bitrate allocation control */
int slow_gain_code; ///< slow gain code (sgaincod)
int slow_decay_code; ///< slow decay code (sdcycod)
......@@ -261,6 +271,114 @@ static void apply_mdct(AC3EncodeContext *s)
}
/**
* Initialize stereo rematrixing.
* If the strategy does not change for each frame, set the rematrixing flags.
*/
static void rematrixing_init(AC3EncodeContext *s)
{
if (s->channel_mode == AC3_CHMODE_STEREO)
s->rematrixing = AC3_REMATRIXING_SUMS;
else
s->rematrixing = AC3_REMATRIXING_NONE;
/* NOTE: AC3_REMATRIXING_ALWAYS might be used in
the future in conjunction with channel coupling. */
if (s->rematrixing & AC3_REMATRIXING_IS_STATIC) {
int flag = (s->rematrixing == AC3_REMATRIXING_ALWAYS);
s->blocks[0].new_rematrixing_strategy = 1;
memset(s->blocks[0].rematrixing_flags, flag,
sizeof(s->blocks[0].rematrixing_flags));
}
}
/**
* Determine rematrixing flags for each block and band.
*/
static void compute_rematrixing_strategy(AC3EncodeContext *s)
{
int nb_coefs;
int blk, bnd, i;
AC3Block *block, *block0;
if (s->rematrixing & AC3_REMATRIXING_IS_STATIC)
return;
nb_coefs = FFMIN(s->nb_coefs[0], s->nb_coefs[1]);
s->blocks[0].new_rematrixing_strategy = 1;
for (blk = 0; blk < AC3_MAX_BLOCKS; blk++) {
block = &s->blocks[blk];
for (bnd = 0; bnd < 4; bnd++) {
/* calculate calculate sum of squared coeffs for one band in one block */
int start = ff_ac3_rematrix_band_tab[bnd];
int end = FFMIN(nb_coefs, ff_ac3_rematrix_band_tab[bnd+1]);
CoefSumType sum[4] = {0,};
for (i = start; i < end; i++) {
CoefType lt = block->mdct_coef[0][i];
CoefType rt = block->mdct_coef[1][i];
CoefType md = lt + rt;
CoefType sd = lt - rt;
sum[0] += lt * lt;
sum[1] += rt * rt;
sum[2] += md * md;
sum[3] += sd * sd;
}
/* compare sums to determine if rematrixing will be used for this band */
if (FFMIN(sum[2], sum[3]) < FFMIN(sum[0], sum[1]))
block->rematrixing_flags[bnd] = 1;
else
block->rematrixing_flags[bnd] = 0;
/* determine if new rematrixing flags will be sent */
if (blk &&
!block->new_rematrixing_strategy &&
block->rematrixing_flags[bnd] != block0->rematrixing_flags[bnd]) {
block->new_rematrixing_strategy = 1;
}
}
block0 = block;
}
}
/**
* Apply stereo rematrixing to coefficients based on rematrixing flags.
*/
static void apply_rematrixing(AC3EncodeContext *s)
{
int nb_coefs;
int blk, bnd, i;
int start, end;
uint8_t *flags;
if (s->rematrixing == AC3_REMATRIXING_NONE)
return;
nb_coefs = FFMIN(s->nb_coefs[0], s->nb_coefs[1]);
for (blk = 0; blk < AC3_MAX_BLOCKS; blk++) {
AC3Block *block = &s->blocks[blk];
if (block->new_rematrixing_strategy)
flags = block->rematrixing_flags;
for (bnd = 0; bnd < 4; bnd++) {
if (flags[bnd]) {
start = ff_ac3_rematrix_band_tab[bnd];
end = FFMIN(nb_coefs, ff_ac3_rematrix_band_tab[bnd+1]);
for (i = start; i < end; i++) {
int32_t lt = block->fixed_coef[0][i];
int32_t rt = block->fixed_coef[1][i];
block->fixed_coef[0][i] = (lt + rt) >> 1;
block->fixed_coef[1][i] = (lt - rt) >> 1;
}
}
}
}
}
/**
* Initialize exponent tables.
*/
......@@ -592,7 +710,6 @@ static void count_frame_bits_fixed(AC3EncodeContext *s)
/* assumptions:
* no dynamic range codes
* no channel coupling
* no rematrixing
* bit allocation parameters do not change between blocks
* SNR offsets do not change between blocks
* no delta bit allocation
......@@ -609,8 +726,6 @@ static void count_frame_bits_fixed(AC3EncodeContext *s)
frame_bits += s->fbw_channels * 2 + 2; /* blksw * c, dithflag * c, dynrnge, cplstre */
if (s->channel_mode == AC3_CHMODE_STEREO) {
frame_bits++; /* rematstr */
if (!blk)
frame_bits += 4;
}
frame_bits += 2 * s->fbw_channels; /* chexpstr[2] * c */
if (s->lfe_on)
......@@ -681,6 +796,13 @@ static void count_frame_bits(AC3EncodeContext *s)
for (blk = 0; blk < AC3_MAX_BLOCKS; blk++) {
uint8_t *exp_strategy = s->blocks[blk].exp_strategy;
/* stereo rematrixing */
if (s->channel_mode == AC3_CHMODE_STEREO &&
s->blocks[blk].new_rematrixing_strategy) {
frame_bits += 4;
}
for (ch = 0; ch < s->fbw_channels; ch++) {
if (exp_strategy[ch] != EXP_REUSE)
frame_bits += 6 + 2; /* chbwcod[6], gainrng[2] */
......@@ -1194,16 +1316,11 @@ static void output_audio_block(AC3EncodeContext *s, int block_num)
/* stereo rematrixing */
if (s->channel_mode == AC3_CHMODE_STEREO) {
if (!block_num) {
/* first block must define rematrixing (rematstr) */
put_bits(&s->pb, 1, 1);
/* dummy rematrixing rematflg(1:4)=0 */
put_bits(&s->pb, 1, block->new_rematrixing_strategy);
if (block->new_rematrixing_strategy) {
/* rematrixing flags */
for (rbnd = 0; rbnd < 4; rbnd++)
put_bits(&s->pb, 1, 0);
} else {
/* no matrixing (but should be used in the future) */
put_bits(&s->pb, 1, 0);
put_bits(&s->pb, 1, block->rematrixing_flags[rbnd]);
}
}
......@@ -1394,8 +1511,12 @@ static int ac3_encode_frame(AVCodecContext *avctx, unsigned char *frame,
apply_mdct(s);
compute_rematrixing_strategy(s);
scale_coefficients(s);
apply_rematrixing(s);
process_exponents(s);
ret = compute_bit_allocation(s);
......@@ -1707,6 +1828,8 @@ static av_cold int ac3_encode_init(AVCodecContext *avctx)
set_bandwidth(s);
rematrixing_init(s);
exponent_init(s);
bit_alloc_init(s);
......
......@@ -34,6 +34,7 @@
typedef int16_t SampleType;
typedef int32_t CoefType;
typedef int64_t CoefSumType;
/**
......
......@@ -34,6 +34,7 @@
typedef float SampleType;
typedef float CoefType;
typedef float CoefSumType;
typedef struct AC3MDCTContext {
......
......@@ -132,6 +132,12 @@ const uint16_t ff_ac3_bitrate_tab[19] = {
160, 192, 224, 256, 320, 384, 448, 512, 576, 640
};
/**
* Table of bin locations for rematrixing bands
* reference: Section 7.5.2 Rematrixing : Frequency Band Definitions
*/
const uint8_t ff_ac3_rematrix_band_tab[5] = { 13, 25, 37, 61, 253 };
/* AC-3 MDCT window */
/* MDCT window */
......
......@@ -32,6 +32,7 @@ extern const uint8_t ff_ac3_enc_channel_map[8][2][6];
extern const uint8_t ff_ac3_dec_channel_map[8][2][6];
extern const uint16_t ff_ac3_sample_rate_tab[3];
extern const uint16_t ff_ac3_bitrate_tab[19];
extern const uint8_t ff_ac3_rematrix_band_tab[5];
extern const int16_t ff_ac3_window[AC3_WINDOW_SIZE/2];
extern const uint8_t ff_ac3_log_add_tab[260];
extern const uint16_t ff_ac3_hearing_threshold_tab[AC3_CRITICAL_BANDS][3];
......
b315176b519a63a35cb91566e768f62b *./tests/data/acodec/ac3.rm
9823c8f74097eab5d148cf0536ae932e *./tests/data/acodec/ac3.rm
98751 ./tests/data/acodec/ac3.rm
......@@ -5,11 +5,9 @@ ret:-1 st:-1 flags:1 ts: 1.894167
ret:-1 st: 0 flags:0 ts: 0.788000
ret: 0 st: 0 flags:1 ts:-0.317000
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
ret: 0 st:-1 flags:0 ts: 2.576668
ret: 0 st: 0 flags:1 dts: 4.179000 pts: 4.179000 pos: 68585 size: 556
ret:-1 st:-1 flags:0 ts: 2.576668
ret:-1 st:-1 flags:1 ts: 1.470835
ret: 0 st: 0 flags:0 ts: 0.365000
ret: 0 st: 0 flags:1 dts: 0.383000 pts: 0.383000 pos: 6533 size: 558
ret:-1 st: 0 flags:0 ts: 0.365000
ret: 0 st: 0 flags:1 ts:-0.741000
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
ret:-1 st:-1 flags:0 ts: 2.153336
......@@ -18,25 +16,20 @@ ret: 0 st: 0 flags:0 ts:-0.058000
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
ret:-1 st: 0 flags:1 ts: 2.836000
ret:-1 st:-1 flags:0 ts: 1.730004
ret: 0 st:-1 flags:1 ts: 0.624171
ret: 0 st: 0 flags:1 dts: 0.452000 pts: 0.452000 pos: 7671 size: 558
ret:-1 st:-1 flags:1 ts: 0.624171
ret: 0 st: 0 flags:0 ts:-0.482000
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
ret: 0 st: 0 flags:1 ts: 2.413000
ret: 0 st: 0 flags:1 dts: 2.333000 pts: 2.333000 pos: 38413 size: 556
ret:-1 st: 0 flags:1 ts: 2.413000
ret:-1 st:-1 flags:0 ts: 1.306672
ret:-1 st:-1 flags:1 ts: 0.200839
ret: 0 st: 0 flags:0 ts:-0.905000
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
ret: 0 st: 0 flags:1 ts: 1.989000
ret: 0 st: 0 flags:1 dts: 1.985000 pts: 1.985000 pos: 32719 size: 558
ret:-1 st: 0 flags:1 ts: 1.989000
ret:-1 st:-1 flags:0 ts: 0.883340
ret: 0 st:-1 flags:1 ts:-0.222493
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
ret: 0 st: 0 flags:0 ts: 2.672000
ret: 0 st: 0 flags:1 dts: 4.179000 pts: 4.179000 pos: 68585 size: 556
ret:-1 st: 0 flags:0 ts: 2.672000
ret:-1 st: 0 flags:1 ts: 1.566000
ret: 0 st:-1 flags:0 ts: 0.460008
ret: 0 st: 0 flags:1 dts: 4.179000 pts: 4.179000 pos: 68585 size: 556
ret:-1 st:-1 flags:0 ts: 0.460008
ret: 0 st:-1 flags:1 ts:-0.645825
ret: 0 st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos: 271 size: 556
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment