[Mplayer-cvslog] CVS: main/liba52 a52.h,NONE,1.1 a52_internal.h,NONE,1.1 bit_allocate.c,NONE,1.1 bitstream.c,NONE,1.1 bitstream.h,NONE,1.1 downmix.c,NONE,1.1 imdct.c,NONE,1.1 imdct_mlib.c,NONE,1.1 liba52.txt,NONE,1.1 mm_accel.h,NONE,1.1 parse.c,NONE,1.1 tables.h,NONE,1.1

Arpi of Ize arpi at mplayer.dev.hu
Sun Dec 9 16:28:47 CET 2001


Update of /cvsroot/mplayer/main/liba52
In directory mplayer:/var/tmp.root/cvs-serv27819

Added Files:
	a52.h a52_internal.h bit_allocate.c bitstream.c bitstream.h 
	downmix.c imdct.c imdct_mlib.c liba52.txt mm_accel.h parse.c 
	tables.h 
Log Message:
imported from liba52 CVS

--- NEW FILE ---
/*
 * a52.h
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#ifndef A52_H
#define A52_H

#ifndef LIBA52_DOUBLE
typedef float sample_t;
#else
typedef double sample_t;
#endif

typedef struct a52_ba_s {
    uint16_t fsnroffst;		/* fine SNR offset */
    uint16_t fgaincod;		/* fast gain */
    uint16_t deltbae;		/* delta bit allocation exists */
    int8_t deltba[50];		/* per-band delta bit allocation */
} a52_ba_t;

typedef struct a52_state_s {
    uint8_t fscod;		/* sample rate */
    uint8_t halfrate;		/* halfrate factor */
    uint8_t acmod;		/* coded channels */
    sample_t clev;		/* centre channel mix level */
    sample_t slev;		/* surround channels mix level */
    uint8_t lfeon;		/* coded lfe channel */

    int output;			/* type of output */
    sample_t level;		/* output level */
    sample_t bias;		/* output bias */

    int dynrnge;		/* apply dynamic range */
    sample_t dynrng;		/* dynamic range */
    void * dynrngdata;		/* dynamic range callback funtion and data */
    sample_t (* dynrngcall) (sample_t range, void * dynrngdata);

    uint16_t cplinu;		/* coupling in use */
    uint16_t chincpl[5];	/* channel coupled */
    uint16_t phsflginu;		/* phase flags in use (stereo only) */
    uint16_t cplbndstrc[18];	/* coupling band structure */
    uint16_t cplstrtmant;	/* coupling channel start mantissa */
    uint16_t cplendmant;	/* coupling channel end mantissa */
    sample_t cplco[5][18];	/* coupling coordinates */

    /* derived information */
    uint16_t cplstrtbnd;	/* coupling start band (for bit allocation) */
    uint16_t ncplbnd;		/* number of coupling bands */

    uint16_t rematflg[4];	/* stereo rematrixing */

    uint16_t endmant[5];	/* channel end mantissa */

    uint8_t cpl_exp[256];	/* decoded coupling channel exponents */
    uint8_t fbw_exp[5][256];	/* decoded channel exponents */
    uint8_t lfe_exp[7];		/* decoded lfe channel exponents */

    uint16_t sdcycod;		/* slow decay */
    uint16_t fdcycod;		/* fast decay */
    uint16_t sgaincod;		/* slow gain */
    uint16_t dbpbcod;		/* dB per bit - encodes the dbknee value */
    uint16_t floorcod;		/* masking floor */

    uint16_t csnroffst;		/* coarse SNR offset */
    a52_ba_t cplba;		/* coupling bit allocation parameters */
    a52_ba_t ba[5];		/* channel bit allocation parameters */
    a52_ba_t lfeba;		/* lfe bit allocation parameters */

    uint16_t cplfleak;		/* coupling fast leak init */
    uint16_t cplsleak;		/* coupling slow leak init */

    /* derived bit allocation information */
    int8_t fbw_bap[5][256];
    int8_t cpl_bap[256];
    int8_t lfe_bap[7];
} a52_state_t;

#define A52_CHANNEL 0
#define A52_MONO 1
#define A52_STEREO 2
#define A52_3F 3
#define A52_2F1R 4
#define A52_3F1R 5
#define A52_2F2R 6
#define A52_3F2R 7
#define A52_CHANNEL1 8
#define A52_CHANNEL2 9
#define A52_DOLBY 10
#define A52_CHANNEL_MASK 15

#define A52_LFE 16
#define A52_ADJUST_LEVEL 32

sample_t * a52_init (uint32_t mm_accel);
int a52_syncinfo (uint8_t * buf, int * flags,
		  int * sample_rate, int * bit_rate);
int a52_frame (a52_state_t * state, uint8_t * buf, int * flags,
	       sample_t * level, sample_t bias);
void a52_dynrng (a52_state_t * state,
		 sample_t (* call) (sample_t, void *), void * data);
int a52_block (a52_state_t * state, sample_t * samples);

#endif /* A52_H */

--- NEW FILE ---
/*
 * a52_internal.h
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#define LEVEL_PLUS6DB 2.0
#define LEVEL_PLUS3DB 1.4142135623730951
#define LEVEL_3DB 0.7071067811865476
#define LEVEL_45DB 0.5946035575013605
#define LEVEL_6DB 0.5

#define EXP_REUSE (0)
#define EXP_D15   (1)
#define EXP_D25   (2)
#define EXP_D45   (3)

#define DELTA_BIT_REUSE (0)
#define DELTA_BIT_NEW (1)
#define DELTA_BIT_NONE (2)
#define DELTA_BIT_RESERVED (3)

void bit_allocate (a52_state_t * state, a52_ba_t * ba, int bndstart,
		   int start, int end, int fastleak, int slowleak,
		   uint8_t * exp, int8_t * bap);

int downmix_init (int input, int flags, sample_t * level,
		  sample_t clev, sample_t slev);
int downmix_coeff (sample_t * coeff, int acmod, int output, sample_t level,
		   sample_t clev, sample_t slev);
void downmix (sample_t * samples, int acmod, int output, sample_t bias,
	      sample_t clev, sample_t slev);
void upmix (sample_t * samples, int acmod, int output);

void imdct_init (uint32_t mm_accel);
extern void (* imdct_256) (sample_t * data, sample_t * delay, sample_t bias);
extern void (* imdct_512) (sample_t * data, sample_t * delay, sample_t bias);
void imdct_do_256_mlib (sample_t * data, sample_t * delay, sample_t bias);
void imdct_do_512_mlib (sample_t * data, sample_t * delay, sample_t bias);

--- NEW FILE ---
/*
 * bit_allocate.c
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#include <inttypes.h>

#include "a52.h"
#include "a52_internal.h"

static int hthtab[3][50] = {
    {0x730, 0x730, 0x7c0, 0x800, 0x820, 0x840, 0x850, 0x850, 0x860, 0x860,
     0x860, 0x860, 0x860, 0x870, 0x870, 0x870, 0x880, 0x880, 0x890, 0x890,
     0x8a0, 0x8a0, 0x8b0, 0x8b0, 0x8c0, 0x8c0, 0x8d0, 0x8e0, 0x8f0, 0x900,
     0x910, 0x910, 0x910, 0x910, 0x900, 0x8f0, 0x8c0, 0x870, 0x820, 0x7e0,
     0x7a0, 0x770, 0x760, 0x7a0, 0x7c0, 0x7c0, 0x6e0, 0x400, 0x3c0, 0x3c0},
    {0x710, 0x710, 0x7a0, 0x7f0, 0x820, 0x830, 0x840, 0x850, 0x850, 0x860,
     0x860, 0x860, 0x860, 0x860, 0x870, 0x870, 0x870, 0x880, 0x880, 0x880,
     0x890, 0x890, 0x8a0, 0x8a0, 0x8b0, 0x8b0, 0x8c0, 0x8c0, 0x8e0, 0x8f0,
     0x900, 0x910, 0x910, 0x910, 0x910, 0x900, 0x8e0, 0x8b0, 0x870, 0x820,
     0x7e0, 0x7b0, 0x760, 0x770, 0x7a0, 0x7c0, 0x780, 0x5d0, 0x3c0, 0x3c0},
    {0x680, 0x680, 0x750, 0x7b0, 0x7e0, 0x810, 0x820, 0x830, 0x840, 0x850,
     0x850, 0x850, 0x860, 0x860, 0x860, 0x860, 0x860, 0x860, 0x860, 0x860,
     0x870, 0x870, 0x870, 0x870, 0x880, 0x880, 0x880, 0x890, 0x8a0, 0x8b0,
     0x8c0, 0x8d0, 0x8e0, 0x8f0, 0x900, 0x910, 0x910, 0x910, 0x900, 0x8f0,
     0x8d0, 0x8b0, 0x840, 0x7f0, 0x790, 0x760, 0x7a0, 0x7c0, 0x7b0, 0x720}
};

static int8_t baptab[305] = {
    16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
    16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
    16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
    16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
    16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
    16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,	/* 93 padding elems */

    16, 16, 16, 16, 16, 16, 16, 16, 16, 14, 14, 14, 14, 14, 14, 14,
    14, 12, 12, 12, 12, 11, 11, 11, 11, 10, 10, 10, 10,  9,  9,  9,
     9,  8,  8,  8,  8,  7,  7,  7,  7,  6,  6,  6,  6,  5,  5,  5,
     5,  4,  4, -3, -3,  3,  3,  3, -2, -2, -1, -1, -1, -1, -1,  0,

     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
     0,  0,  0,  0					/* 148 padding elems */
};

static int bndtab[30] = {21, 22,  23,  24,  25,  26,  27,  28,  31,  34,
			 37, 40,  43,  46,  49,  55,  61,  67,  73,  79,
			 85, 97, 109, 121, 133, 157, 181, 205, 229, 253};

static int8_t latab[256] = {
    -64, -63, -62, -61, -60, -59, -58, -57, -56, -55, -54, -53,
    -52, -52, -51, -50, -49, -48, -47, -47, -46, -45, -44, -44,
    -43, -42, -41, -41, -40, -39, -38, -38, -37, -36, -36, -35,
    -35, -34, -33, -33, -32, -32, -31, -30, -30, -29, -29, -28,
    -28, -27, -27, -26, -26, -25, -25, -24, -24, -23, -23, -22,
    -22, -21, -21, -21, -20, -20, -19, -19, -19, -18, -18, -18,
    -17, -17, -17, -16, -16, -16, -15, -15, -15, -14, -14, -14,
    -13, -13, -13, -13, -12, -12, -12, -12, -11, -11, -11, -11,
    -10, -10, -10, -10, -10,  -9,  -9,  -9,  -9,  -9,  -8,  -8,
     -8,  -8,  -8,  -8,  -7,  -7,  -7,  -7,  -7,  -7,  -6,  -6,
     -6,  -6,  -6,  -6,  -6,  -6,  -5,  -5,  -5,  -5,  -5,  -5,
     -5,  -5,  -4,  -4,  -4,  -4,  -4,  -4,  -4,  -4,  -4,  -4,
     -4,  -3,  -3,  -3,  -3,  -3,  -3,  -3,  -3,  -3,  -3,  -3,
     -3,  -3,  -3,  -2,  -2,  -2,  -2,  -2,  -2,  -2,  -2,  -2,
     -2,  -2,  -2,  -2,  -2,  -2,  -2,  -2,  -2,  -2,  -1,  -1,
     -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,
     -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,  -1,
     -1,  -1,  -1,  -1,  -1,  -1,   0,   0,   0,   0,   0,   0,
      0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
      0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
      0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
      0,   0,   0,   0
};

#define UPDATE_LEAK() 		\
do {				\
    fastleak += fdecay;		\
    if (fastleak > psd + fgain)	\
	fastleak = psd + fgain;	\
    slowleak += sdecay;		\
    if (slowleak > psd + sgain)	\
	slowleak = psd + sgain;	\
} while (0)

#define COMPUTE_MASK()				\
do {						\
    if (psd > dbknee)				\
	mask -= (psd - dbknee) >> 2;		\
    if (mask > hth [i >> halfrate])		\
	mask = hth [i >> halfrate];		\
    mask -= snroffset + 128 * deltba[i];	\
    mask = (mask > 0) ? 0 : ((-mask) >> 5);	\
    mask -= floor;				\
} while (0)

void bit_allocate (a52_state_t * state, a52_ba_t * ba, int bndstart,
		   int start, int end, int fastleak, int slowleak,
		   uint8_t * exp, int8_t * bap)
{
    static int slowgain[4] = {0x540, 0x4d8, 0x478, 0x410};
    static int dbpbtab[4]  = {0xc00, 0x500, 0x300, 0x100};
    static int floortab[8] = {0x910, 0x950, 0x990, 0x9d0,
			      0xa10, 0xa90, 0xb10, 0x1400};

    int i, j;
    int fdecay, fgain, sdecay, sgain, dbknee, floor, snroffset;
    int psd, mask;
    int8_t * deltba;
    int * hth;
    int halfrate;

    halfrate = state->halfrate;
    fdecay = (63 + 20 * state->fdcycod) >> halfrate;
    fgain = 128 + 128 * ba->fgaincod;
    sdecay = (15 + 2 * state->sdcycod) >> halfrate;
    sgain = slowgain[state->sgaincod];
    dbknee = dbpbtab[state->dbpbcod];
    hth = hthtab[state->fscod];
    /*
     * if there is no delta bit allocation, make deltba point to an area
     * known to contain zeroes. baptab+156 here.
     */
    deltba = (ba->deltbae == DELTA_BIT_NONE) ? baptab + 156 : ba->deltba;
    floor = floortab[state->floorcod];
    snroffset = 960 - 64 * state->csnroffst - 4 * ba->fsnroffst + floor;
    floor >>= 5;

    i = bndstart;
    j = start;
    if (start == 0) {	/* not the coupling channel */
	int lowcomp;

	lowcomp = 0;
	j = end - 1;
	do {
	    if (i < j) {
		if (exp[i+1] == exp[i] - 2)
		    lowcomp = 384;
		else if (lowcomp && (exp[i+1] > exp[i]))
		    lowcomp -= 64;
	    }
	    psd = 128 * exp[i];
	    mask = psd + fgain + lowcomp;
	    COMPUTE_MASK ();
	    bap[i] = (baptab+156)[mask + 4 * exp[i]];
	    i++;
	} while ((i < 3) || ((i < 7) && (exp[i] > exp[i-1])));
	fastleak = psd + fgain;
	slowleak = psd + sgain;

	while (i < 7) {
	    if (i < j) {
		if (exp[i+1] == exp[i] - 2)
		    lowcomp = 384;
		else if (lowcomp && (exp[i+1] > exp[i]))
		    lowcomp -= 64;
	    }
	    psd = 128 * exp[i];
	    UPDATE_LEAK ();
	    mask = ((fastleak + lowcomp < slowleak) ?
		    fastleak + lowcomp : slowleak);
	    COMPUTE_MASK ();
	    bap[i] = (baptab+156)[mask + 4 * exp[i]];
	    i++;
	}

	if (end == 7)	/* lfe channel */
	    return;

	do {
	    if (exp[i+1] == exp[i] - 2)
		lowcomp = 320;
	    else if (lowcomp && (exp[i+1] > exp[i]))
		lowcomp -= 64;
	    psd = 128 * exp[i];
	    UPDATE_LEAK ();
	    mask = ((fastleak + lowcomp < slowleak) ?
		    fastleak + lowcomp : slowleak);
	    COMPUTE_MASK ();
	    bap[i] = (baptab+156)[mask + 4 * exp[i]];
	    i++;
	} while (i < 20);

	while (lowcomp > 128) {		/* two iterations maximum */
	    lowcomp -= 128;
	    psd = 128 * exp[i];
	    UPDATE_LEAK ();
	    mask = ((fastleak + lowcomp < slowleak) ?
		    fastleak + lowcomp : slowleak);
	    COMPUTE_MASK ();
	    bap[i] = (baptab+156)[mask + 4 * exp[i]];
	    i++;
	}
	j = i;
    }

    do {
	int startband, endband;

	startband = j;
	endband = ((bndtab-20)[i] < end) ? (bndtab-20)[i] : end;
	psd = 128 * exp[j++];
	while (j < endband) {
	    int next, delta;

	    next = 128 * exp[j++];
	    delta = next - psd;
	    switch (delta >> 9) {
	    case -6: case -5: case -4: case -3: case -2:
		psd = next;
		break;
	    case -1:
		psd = next + latab[(-delta) >> 1];
		break;
	    case 0:
		psd += latab[delta >> 1];
		break;
	    }
	}
	/* minpsd = -289 */
	UPDATE_LEAK ();
	mask = (fastleak < slowleak) ? fastleak : slowleak;
	COMPUTE_MASK ();
	i++;
	j = startband;
	do {
	    /* max(mask+4*exp)=147=-(minpsd+fgain-deltba-snroffset)>>5+4*exp */
	    /* min(mask+4*exp)=-156=-(sgain-deltba-snroffset)>>5 */
	    bap[j] = (baptab+156)[mask + 4 * exp[j]];
	} while (++j < endband);
    } while (j < end);
}

--- NEW FILE ---
/*
 * bitstream.c
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#include <inttypes.h>

#include "a52.h"
#include "a52_internal.h"
#include "bitstream.h"

#define BUFFER_SIZE 4096

static uint32_t * buffer_start;

uint32_t bits_left;
uint32_t current_word;

void bitstream_set_ptr (uint8_t * buf)
{
    int align;

    align = (int)buf & 3;
    buffer_start = (uint32_t *) (buf - align);
    bits_left = 0;
    bitstream_get (align * 8);
}

static inline void
bitstream_fill_current()
{
    uint32_t tmp;

    tmp = *(buffer_start++);
    current_word = swab32 (tmp);
}

/*
 * The fast paths for _get is in the
 * bitstream.h header file so it can be inlined.
 *
 * The "bottom half" of this routine is suffixed _bh
 *
 * -ah
 */

uint32_t
bitstream_get_bh(uint32_t num_bits)
{
    uint32_t result;

    num_bits -= bits_left;
    result = (current_word << (32 - bits_left)) >> (32 - bits_left);

    bitstream_fill_current();

    if(num_bits != 0)
	result = (result << num_bits) | (current_word >> (32 - num_bits));
	
    bits_left = 32 - num_bits;

    return result;
}

int32_t
bitstream_get_bh_2(uint32_t num_bits)
{
    int32_t result;

    num_bits -= bits_left;
    result = (((int32_t)current_word) << (32 - bits_left)) >> (32 - bits_left);

    bitstream_fill_current();

    if(num_bits != 0)
	result = (result << num_bits) | (current_word >> (32 - num_bits));
	
    bits_left = 32 - num_bits;

    return result;
}

--- NEW FILE ---
/*
 * bitstream.h
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

/* (stolen from the kernel) */
#ifdef WORDS_BIGENDIAN

#	define swab32(x) (x)

#else

#	if defined (__i386__)

#	define swab32(x) __i386_swab32(x)
	static inline const uint32_t __i386_swab32(uint32_t x)
	{
		__asm__("bswap %0" : "=r" (x) : "0" (x));
		return x;
	}

#	else

#	define swab32(x)\
((((uint8_t*)&x)[0] << 24) | (((uint8_t*)&x)[1] << 16) |  \
 (((uint8_t*)&x)[2] << 8)  | (((uint8_t*)&x)[3]))

#	endif
#endif

extern uint32_t bits_left;
extern uint32_t current_word;

void bitstream_set_ptr (uint8_t * buf);
uint32_t bitstream_get_bh(uint32_t num_bits);
int32_t bitstream_get_bh_2(uint32_t num_bits);

static inline uint32_t 
bitstream_get(uint32_t num_bits)
{
    uint32_t result;
	
    if(num_bits < bits_left) {
	result = (current_word << (32 - bits_left)) >> (32 - num_bits);
	bits_left -= num_bits;
	return result;
    }

    return bitstream_get_bh(num_bits);
}

static inline int32_t 
bitstream_get_2(uint32_t num_bits)
{
    int32_t result;
	
    if(num_bits < bits_left) {
	result = (((int32_t)current_word) << (32 - bits_left)) >> (32 - num_bits);
	bits_left -= num_bits;
	return result;
    }

    return bitstream_get_bh_2(num_bits);
}

--- NEW FILE ---
/*
 * downmix.c
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#include <string.h>
#include <inttypes.h>

#include "a52.h"
#include "a52_internal.h"

#define CONVERT(acmod,output) (((output) << 3) + (acmod))

int downmix_init (int input, int flags, sample_t * level,
		  sample_t clev, sample_t slev)
{
    static uint8_t table[11][8] = {
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_STEREO,
	 A52_STEREO,	A52_STEREO,	A52_STEREO,	A52_STEREO},
	{A52_MONO,	A52_MONO,	A52_MONO,	A52_MONO,
	 A52_MONO,	A52_MONO,	A52_MONO,	A52_MONO},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_STEREO,
	 A52_STEREO,	A52_STEREO,	A52_STEREO,	A52_STEREO},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_3F,
	 A52_STEREO,	A52_3F,		A52_STEREO,	A52_3F},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_STEREO,
	 A52_2F1R,	A52_2F1R,	A52_2F1R,	A52_2F1R},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_STEREO,
	 A52_2F1R,	A52_3F1R,	A52_2F1R,	A52_3F1R},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_3F,
	 A52_2F2R,	A52_2F2R,	A52_2F2R,	A52_2F2R},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_3F,
	 A52_2F2R,	A52_3F2R,	A52_2F2R,	A52_3F2R},
	{A52_CHANNEL1,	A52_MONO,	A52_MONO,	A52_MONO,
	 A52_MONO,	A52_MONO,	A52_MONO,	A52_MONO},
	{A52_CHANNEL2,	A52_MONO,	A52_MONO,	A52_MONO,
	 A52_MONO,	A52_MONO,	A52_MONO,	A52_MONO},
	{A52_CHANNEL,	A52_DOLBY,	A52_STEREO,	A52_DOLBY,
	 A52_DOLBY,	A52_DOLBY,	A52_DOLBY,	A52_DOLBY}
    };
    int output;

    output = flags & A52_CHANNEL_MASK;
    if (output > A52_DOLBY)
	return -1;

    output = table[output][input & 7];

    if ((output == A52_STEREO) &&
	((input == A52_DOLBY) || ((input == A52_3F) && (clev == LEVEL_3DB))))
	output = A52_DOLBY;

    if (flags & A52_ADJUST_LEVEL)
	switch (CONVERT (input & 7, output)) {

	case CONVERT (A52_3F, A52_MONO):
	    *level *= LEVEL_3DB / (1 + clev);
	    break;

	case CONVERT (A52_STEREO, A52_MONO):
	case CONVERT (A52_2F2R, A52_2F1R):
	case CONVERT (A52_3F2R, A52_3F1R):
	level_3db:
	    *level *= LEVEL_3DB;
	    break;

	case CONVERT (A52_3F2R, A52_2F1R):
	    if (clev < LEVEL_PLUS3DB - 1)
		goto level_3db;
	    /* break thru */
	case CONVERT (A52_3F, A52_STEREO):
	case CONVERT (A52_3F1R, A52_2F1R):
	case CONVERT (A52_3F1R, A52_2F2R):
	case CONVERT (A52_3F2R, A52_2F2R):
	    *level /= 1 + clev;
	    break;

	case CONVERT (A52_2F1R, A52_MONO):
	    *level *= LEVEL_PLUS3DB / (2 + slev);
	    break;

	case CONVERT (A52_2F1R, A52_STEREO):
	case CONVERT (A52_3F1R, A52_3F):
	    *level /= 1 + slev * LEVEL_3DB;
	    break;

	case CONVERT (A52_3F1R, A52_MONO):
	    *level *= LEVEL_3DB / (1 + clev + 0.5 * slev);
	    break;

	case CONVERT (A52_3F1R, A52_STEREO):
	    *level /= 1 + clev + slev * LEVEL_3DB;
	    break;

	case CONVERT (A52_2F2R, A52_MONO):
	    *level *= LEVEL_3DB / (1 + slev);
	    break;

	case CONVERT (A52_2F2R, A52_STEREO):
	case CONVERT (A52_3F2R, A52_3F):
	    *level /= 1 + slev;
	    break;

	case CONVERT (A52_3F2R, A52_MONO):
	    *level *= LEVEL_3DB / (1 + clev + slev);
	    break;

	case CONVERT (A52_3F2R, A52_STEREO):
	    *level /= 1 + clev + slev;
	    break;

	case CONVERT (A52_MONO, A52_DOLBY):
	    *level *= LEVEL_PLUS3DB;
	    break;

	case CONVERT (A52_3F, A52_DOLBY):
	case CONVERT (A52_2F1R, A52_DOLBY):
	    *level *= 1 / (1 + LEVEL_3DB);
	    break;

	case CONVERT (A52_3F1R, A52_DOLBY):
	case CONVERT (A52_2F2R, A52_DOLBY):
	    *level *= 1 / (1 + 2 * LEVEL_3DB);
	    break;

	case CONVERT (A52_3F2R, A52_DOLBY):
	    *level *= 1 / (1 + 3 * LEVEL_3DB);
	    break;
	}

    return output;
}

int downmix_coeff (sample_t * coeff, int acmod, int output, sample_t level,
		   sample_t clev, sample_t slev)
{
    switch (CONVERT (acmod, output & A52_CHANNEL_MASK)) {

    case CONVERT (A52_CHANNEL, A52_CHANNEL):
    case CONVERT (A52_MONO, A52_MONO):
    case CONVERT (A52_STEREO, A52_STEREO):
    case CONVERT (A52_3F, A52_3F):
    case CONVERT (A52_2F1R, A52_2F1R):
    case CONVERT (A52_3F1R, A52_3F1R):
    case CONVERT (A52_2F2R, A52_2F2R):
    case CONVERT (A52_3F2R, A52_3F2R):
    case CONVERT (A52_STEREO, A52_DOLBY):
	coeff[0] = coeff[1] = coeff[2] = coeff[3] = coeff[4] = level;
	return 0;

    case CONVERT (A52_CHANNEL, A52_MONO):
	coeff[0] = coeff[1] = level * LEVEL_6DB;
	return 3;

    case CONVERT (A52_STEREO, A52_MONO):
	coeff[0] = coeff[1] = level * LEVEL_3DB;
	return 3;

    case CONVERT (A52_3F, A52_MONO):
	coeff[0] = coeff[2] = level * LEVEL_3DB;
	coeff[1] = level * clev * LEVEL_PLUS3DB;
	return 7;

    case CONVERT (A52_2F1R, A52_MONO):
	coeff[0] = coeff[1] = level * LEVEL_3DB;
	coeff[2] = level * slev * LEVEL_3DB;
	return 7;

    case CONVERT (A52_2F2R, A52_MONO):
	coeff[0] = coeff[1] = level * LEVEL_3DB;
	coeff[2] = coeff[3] = level * slev * LEVEL_3DB;
	return 15;

    case CONVERT (A52_3F1R, A52_MONO):
	coeff[0] = coeff[2] = level * LEVEL_3DB;
	coeff[1] = level * clev * LEVEL_PLUS3DB;
	coeff[3] = level * slev * LEVEL_3DB;
	return 15;

    case CONVERT (A52_3F2R, A52_MONO):
	coeff[0] = coeff[2] = level * LEVEL_3DB;
	coeff[1] = level * clev * LEVEL_PLUS3DB;
	coeff[3] = coeff[4] = level * slev * LEVEL_3DB;
	return 31;

    case CONVERT (A52_MONO, A52_DOLBY):
	coeff[0] = level * LEVEL_3DB;
	return 0;

    case CONVERT (A52_3F, A52_DOLBY):
	clev = LEVEL_3DB;
    case CONVERT (A52_3F, A52_STEREO):
    case CONVERT (A52_3F1R, A52_2F1R):
    case CONVERT (A52_3F2R, A52_2F2R):
	coeff[0] = coeff[2] = coeff[3] = coeff[4] = level;
	coeff[1] = level * clev;
	return 7;

    case CONVERT (A52_2F1R, A52_DOLBY):
	slev = 1;
    case CONVERT (A52_2F1R, A52_STEREO):
	coeff[0] = coeff[1] = level;
	coeff[2] = level * slev * LEVEL_3DB;
	return 7;

    case CONVERT (A52_3F1R, A52_DOLBY):
	clev = LEVEL_3DB;
	slev = 1;
    case CONVERT (A52_3F1R, A52_STEREO):
	coeff[0] = coeff[2] = level;
	coeff[1] = level * clev;
	coeff[3] = level * slev * LEVEL_3DB;
	return 15;

    case CONVERT (A52_2F2R, A52_DOLBY):
	slev = LEVEL_3DB;
    case CONVERT (A52_2F2R, A52_STEREO):
	coeff[0] = coeff[1] = level;
	coeff[2] = coeff[3] = level * slev;
	return 15;

    case CONVERT (A52_3F2R, A52_DOLBY):
	clev = LEVEL_3DB;
    case CONVERT (A52_3F2R, A52_2F1R):
	slev = LEVEL_3DB;
    case CONVERT (A52_3F2R, A52_STEREO):
	coeff[0] = coeff[2] = level;
	coeff[1] = level * clev;
	coeff[3] = coeff[4] = level * slev;
	return 31;

    case CONVERT (A52_3F1R, A52_3F):
	coeff[0] = coeff[1] = coeff[2] = level;
	coeff[3] = level * slev * LEVEL_3DB;
	return 13;

    case CONVERT (A52_3F2R, A52_3F):
	coeff[0] = coeff[1] = coeff[2] = level;
	coeff[3] = coeff[4] = level * slev;
	return 29;

    case CONVERT (A52_2F2R, A52_2F1R):
	coeff[0] = coeff[1] = level;
	coeff[2] = coeff[3] = level * LEVEL_3DB;
	return 12;

    case CONVERT (A52_3F2R, A52_3F1R):
	coeff[0] = coeff[1] = coeff[2] = level;
	coeff[3] = coeff[4] = level * LEVEL_3DB;
	return 24;

    case CONVERT (A52_2F1R, A52_2F2R):
	coeff[0] = coeff[1] = level;
	coeff[2] = level * LEVEL_3DB;
	return 0;

    case CONVERT (A52_3F1R, A52_2F2R):
	coeff[0] = coeff[2] = level;
	coeff[1] = level * clev;
	coeff[3] = level * LEVEL_3DB;
	return 7;

    case CONVERT (A52_3F1R, A52_3F2R):
	coeff[0] = coeff[1] = coeff[2] = level;
	coeff[3] = level * LEVEL_3DB;
	return 0;

    case CONVERT (A52_CHANNEL, A52_CHANNEL1):
	coeff[0] = level;
	coeff[1] = 0;
	return 0;

    case CONVERT (A52_CHANNEL, A52_CHANNEL2):
	coeff[0] = 0;
	coeff[1] = level;
	return 0;
    }

    return -1;	/* NOTREACHED */
}

static void mix2to1 (sample_t * dest, sample_t * src, sample_t bias)
{
    int i;

    for (i = 0; i < 256; i++)
	dest[i] += src[i] + bias;
}

static void mix3to1 (sample_t * samples, sample_t bias)
{
    int i;

    for (i = 0; i < 256; i++)
	samples[i] += samples[i + 256] + samples[i + 512] + bias;
}

static void mix4to1 (sample_t * samples, sample_t bias)
{
    int i;

    for (i = 0; i < 256; i++)
	samples[i] += (samples[i + 256] + samples[i + 512] +
		       samples[i + 768] + bias);
}

static void mix5to1 (sample_t * samples, sample_t bias)
{
    int i;

    for (i = 0; i < 256; i++)
	samples[i] += (samples[i + 256] + samples[i + 512] +
		       samples[i + 768] + samples[i + 1024] + bias);
}

static void mix3to2 (sample_t * samples, sample_t bias)
{
    int i;
    sample_t common;

    for (i = 0; i < 256; i++) {
	common = samples[i + 256] + bias;
	samples[i] += common;
	samples[i + 256] = samples[i + 512] + common;
    }
}

static void mix21to2 (sample_t * left, sample_t * right, sample_t bias)
{
    int i;
    sample_t common;

    for (i = 0; i < 256; i++) {
	common = right[i + 256] + bias;
	left[i] += common;
	right[i] += common;
    }
}

static void mix21toS (sample_t * samples, sample_t bias)
{
    int i;
    sample_t surround;

    for (i = 0; i < 256; i++) {
	surround = samples[i + 512];
	samples[i] += bias - surround;
	samples[i + 256] += bias + surround;
    }
}

static void mix31to2 (sample_t * samples, sample_t bias)
{
    int i;
    sample_t common;

    for (i = 0; i < 256; i++) {
	common = samples[i + 256] + samples[i + 768] + bias;
	samples[i] += common;
	samples[i + 256] = samples[i + 512] + common;
    }
}

static void mix31toS (sample_t * samples, sample_t bias)
{
    int i;
    sample_t common, surround;

    for (i = 0; i < 256; i++) {
	common = samples[i + 256] + bias;
	surround = samples[i + 768];
	samples[i] += common - surround;
	samples[i + 256] = samples[i + 512] + common + surround;
    }
}

static void mix22toS (sample_t * samples, sample_t bias)
{
    int i;
    sample_t surround;

    for (i = 0; i < 256; i++) {
	surround = samples[i + 512] + samples[i + 768];
	samples[i] += bias - surround;
	samples[i + 256] += bias + surround;
    }
}

static void mix32to2 (sample_t * samples, sample_t bias)
{
    int i;
    sample_t common;

    for (i = 0; i < 256; i++) {
	common = samples[i + 256] + bias;
	samples[i] += common + samples[i + 768];
	samples[i + 256] = common + samples[i + 512] + samples[i + 1024];
    }
}

static void mix32toS (sample_t * samples, sample_t bias)
{
    int i;
    sample_t common, surround;

    for (i = 0; i < 256; i++) {
	common = samples[i + 256] + bias;
	surround = samples[i + 768] + samples[i + 1024];
	samples[i] += common - surround;
	samples[i + 256] = samples[i + 512] + common + surround;
    }
}

static void move2to1 (sample_t * src, sample_t * dest, sample_t bias)
{
    int i;

    for (i = 0; i < 256; i++)
	dest[i] = src[i] + src[i + 256] + bias;
}

static void zero (sample_t * samples)
{
    int i;

    for (i = 0; i < 256; i++)
	samples[i] = 0;
}

void downmix (sample_t * samples, int acmod, int output, sample_t bias,
	      sample_t clev, sample_t slev)
{
    switch (CONVERT (acmod, output & A52_CHANNEL_MASK)) {

    case CONVERT (A52_CHANNEL, A52_CHANNEL2):
	memcpy (samples, samples + 256, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_CHANNEL, A52_MONO):
    case CONVERT (A52_STEREO, A52_MONO):
    mix_2to1:
	mix2to1 (samples, samples + 256, bias);
	break;

    case CONVERT (A52_2F1R, A52_MONO):
	if (slev == 0)
	    goto mix_2to1;
    case CONVERT (A52_3F, A52_MONO):
    mix_3to1:
	mix3to1 (samples, bias);
	break;

    case CONVERT (A52_3F1R, A52_MONO):
	if (slev == 0)
	    goto mix_3to1;
    case CONVERT (A52_2F2R, A52_MONO):
	if (slev == 0)
	    goto mix_2to1;
	mix4to1 (samples, bias);
	break;

    case CONVERT (A52_3F2R, A52_MONO):
	if (slev == 0)
	    goto mix_3to1;
	mix5to1 (samples, bias);
	break;

    case CONVERT (A52_MONO, A52_DOLBY):
	memcpy (samples + 256, samples, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_3F, A52_STEREO):
    case CONVERT (A52_3F, A52_DOLBY):
    mix_3to2:
	mix3to2 (samples, bias);
	break;

    case CONVERT (A52_2F1R, A52_STEREO):
	if (slev == 0)
	    break;
	mix21to2 (samples, samples + 256, bias);
	break;

    case CONVERT (A52_2F1R, A52_DOLBY):
	mix21toS (samples, bias);
	break;

    case CONVERT (A52_3F1R, A52_STEREO):
	if (slev == 0)
	    goto mix_3to2;
	mix31to2 (samples, bias);
	break;

    case CONVERT (A52_3F1R, A52_DOLBY):
	mix31toS (samples, bias);
	break;

    case CONVERT (A52_2F2R, A52_STEREO):
	if (slev == 0)
	    break;
	mix2to1 (samples, samples + 512, bias);
	mix2to1 (samples + 256, samples + 768, bias);
	break;

    case CONVERT (A52_2F2R, A52_DOLBY):
	mix22toS (samples, bias);
	break;

    case CONVERT (A52_3F2R, A52_STEREO):
	if (slev == 0)
	    goto mix_3to2;
	mix32to2 (samples, bias);
	break;

    case CONVERT (A52_3F2R, A52_DOLBY):
	mix32toS (samples, bias);
	break;

    case CONVERT (A52_3F1R, A52_3F):
	if (slev == 0)
	    break;
	mix21to2 (samples, samples + 512, bias);
	break;

    case CONVERT (A52_3F2R, A52_3F):
	if (slev == 0)
	    break;
	mix2to1 (samples, samples + 768, bias);
	mix2to1 (samples + 512, samples + 1024, bias);
	break;

    case CONVERT (A52_3F1R, A52_2F1R):
	mix3to2 (samples, bias);
	memcpy (samples + 512, samples + 768, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_2F2R, A52_2F1R):
	mix2to1 (samples + 512, samples + 768, bias);
	break;

    case CONVERT (A52_3F2R, A52_2F1R):
	mix3to2 (samples, bias);
	move2to1 (samples + 768, samples + 512, bias);
	break;

    case CONVERT (A52_3F2R, A52_3F1R):
	mix2to1 (samples + 768, samples + 1024, bias);
	break;

    case CONVERT (A52_2F1R, A52_2F2R):
	memcpy (samples + 768, samples + 512, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_3F1R, A52_2F2R):
	mix3to2 (samples, bias);
	memcpy (samples + 512, samples + 768, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_3F2R, A52_2F2R):
	mix3to2 (samples, bias);
	memcpy (samples + 512, samples + 768, 256 * sizeof (sample_t));
	memcpy (samples + 768, samples + 1024, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_3F1R, A52_3F2R):
	memcpy (samples + 1027, samples + 768, 256 * sizeof (sample_t));
	break;
    }
}

void upmix (sample_t * samples, int acmod, int output)
{
    switch (CONVERT (acmod, output & A52_CHANNEL_MASK)) {

    case CONVERT (A52_CHANNEL, A52_CHANNEL2):
	memcpy (samples + 256, samples, 256 * sizeof (sample_t));
	break;

    case CONVERT (A52_3F2R, A52_MONO):
	zero (samples + 1024);
    case CONVERT (A52_3F1R, A52_MONO):
    case CONVERT (A52_2F2R, A52_MONO):
	zero (samples + 768);
    case CONVERT (A52_3F, A52_MONO):
    case CONVERT (A52_2F1R, A52_MONO):
	zero (samples + 512);
    case CONVERT (A52_CHANNEL, A52_MONO):
    case CONVERT (A52_STEREO, A52_MONO):
	zero (samples + 256);
	break;

    case CONVERT (A52_3F2R, A52_STEREO):
    case CONVERT (A52_3F2R, A52_DOLBY):
	zero (samples + 1024);
    case CONVERT (A52_3F1R, A52_STEREO):
    case CONVERT (A52_3F1R, A52_DOLBY):
	zero (samples + 768);
    case CONVERT (A52_3F, A52_STEREO):
    case CONVERT (A52_3F, A52_DOLBY):
    mix_3to2:
	memcpy (samples + 512, samples + 256, 256 * sizeof (sample_t));
	zero (samples + 256);
	break;

    case CONVERT (A52_2F2R, A52_STEREO):
    case CONVERT (A52_2F2R, A52_DOLBY):
	zero (samples + 768);
    case CONVERT (A52_2F1R, A52_STEREO):
    case CONVERT (A52_2F1R, A52_DOLBY):
	zero (samples + 512);
	break;

    case CONVERT (A52_3F2R, A52_3F):
	zero (samples + 1024);
    case CONVERT (A52_3F1R, A52_3F):
    case CONVERT (A52_2F2R, A52_2F1R):
	zero (samples + 768);
	break;

    case CONVERT (A52_3F2R, A52_3F1R):
	zero (samples + 1024);
	break;

    case CONVERT (A52_3F2R, A52_2F1R):
	zero (samples + 1024);
    case CONVERT (A52_3F1R, A52_2F1R):
    mix_31to21:
	memcpy (samples + 768, samples + 512, 256 * sizeof (sample_t));
	goto mix_3to2;

    case CONVERT (A52_3F2R, A52_2F2R):
	memcpy (samples + 1024, samples + 768, 256 * sizeof (sample_t));
	goto mix_31to21;
    }
}

--- NEW FILE ---
/*
 * imdct.c
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#include <math.h>
#include <stdio.h>
#ifndef M_PI
#define M_PI 3.1415926535897932384626433832795029
#endif
#include <inttypes.h>

#include "a52.h"
#include "a52_internal.h"
#include "mm_accel.h"

void (* imdct_256) (sample_t data[], sample_t delay[], sample_t bias);
void (* imdct_512) (sample_t data[], sample_t delay[], sample_t bias);

typedef struct complex_s {
    sample_t real;
    sample_t imag;
} complex_t;


/* 128 point bit-reverse LUT */
static uint8_t bit_reverse_512[] = {
	0x00, 0x40, 0x20, 0x60, 0x10, 0x50, 0x30, 0x70, 
	0x08, 0x48, 0x28, 0x68, 0x18, 0x58, 0x38, 0x78, 
	0x04, 0x44, 0x24, 0x64, 0x14, 0x54, 0x34, 0x74, 
	0x0c, 0x4c, 0x2c, 0x6c, 0x1c, 0x5c, 0x3c, 0x7c, 
	0x02, 0x42, 0x22, 0x62, 0x12, 0x52, 0x32, 0x72, 
	0x0a, 0x4a, 0x2a, 0x6a, 0x1a, 0x5a, 0x3a, 0x7a, 
	0x06, 0x46, 0x26, 0x66, 0x16, 0x56, 0x36, 0x76, 
	0x0e, 0x4e, 0x2e, 0x6e, 0x1e, 0x5e, 0x3e, 0x7e, 
	0x01, 0x41, 0x21, 0x61, 0x11, 0x51, 0x31, 0x71, 
	0x09, 0x49, 0x29, 0x69, 0x19, 0x59, 0x39, 0x79, 
	0x05, 0x45, 0x25, 0x65, 0x15, 0x55, 0x35, 0x75, 
	0x0d, 0x4d, 0x2d, 0x6d, 0x1d, 0x5d, 0x3d, 0x7d, 
	0x03, 0x43, 0x23, 0x63, 0x13, 0x53, 0x33, 0x73, 
	0x0b, 0x4b, 0x2b, 0x6b, 0x1b, 0x5b, 0x3b, 0x7b, 
	0x07, 0x47, 0x27, 0x67, 0x17, 0x57, 0x37, 0x77, 
	0x0f, 0x4f, 0x2f, 0x6f, 0x1f, 0x5f, 0x3f, 0x7f};

static uint8_t bit_reverse_256[] = {
	0x00, 0x20, 0x10, 0x30, 0x08, 0x28, 0x18, 0x38, 
	0x04, 0x24, 0x14, 0x34, 0x0c, 0x2c, 0x1c, 0x3c, 
	0x02, 0x22, 0x12, 0x32, 0x0a, 0x2a, 0x1a, 0x3a, 
	0x06, 0x26, 0x16, 0x36, 0x0e, 0x2e, 0x1e, 0x3e, 
	0x01, 0x21, 0x11, 0x31, 0x09, 0x29, 0x19, 0x39, 
	0x05, 0x25, 0x15, 0x35, 0x0d, 0x2d, 0x1d, 0x3d, 
	0x03, 0x23, 0x13, 0x33, 0x0b, 0x2b, 0x1b, 0x3b, 
	0x07, 0x27, 0x17, 0x37, 0x0f, 0x2f, 0x1f, 0x3f};

static complex_t buf[128];

/* Twiddle factor LUT */
static complex_t w_1[1];
static complex_t w_2[2];
static complex_t w_4[4];
static complex_t w_8[8];
static complex_t w_16[16];
static complex_t w_32[32];
static complex_t w_64[64];
static complex_t * w[7] = {w_1, w_2, w_4, w_8, w_16, w_32, w_64};

/* Twiddle factors for IMDCT */
static sample_t xcos1[128];
static sample_t xsin1[128];
static sample_t xcos2[64];
static sample_t xsin2[64];

/* Windowing function for Modified DCT - Thank you acroread */
sample_t imdct_window[] = {
	0.00014, 0.00024, 0.00037, 0.00051, 0.00067, 0.00086, 0.00107, 0.00130,
	0.00157, 0.00187, 0.00220, 0.00256, 0.00297, 0.00341, 0.00390, 0.00443,
	0.00501, 0.00564, 0.00632, 0.00706, 0.00785, 0.00871, 0.00962, 0.01061,
	0.01166, 0.01279, 0.01399, 0.01526, 0.01662, 0.01806, 0.01959, 0.02121,
	0.02292, 0.02472, 0.02662, 0.02863, 0.03073, 0.03294, 0.03527, 0.03770,
	0.04025, 0.04292, 0.04571, 0.04862, 0.05165, 0.05481, 0.05810, 0.06153,
	0.06508, 0.06878, 0.07261, 0.07658, 0.08069, 0.08495, 0.08935, 0.09389,
	0.09859, 0.10343, 0.10842, 0.11356, 0.11885, 0.12429, 0.12988, 0.13563,
	0.14152, 0.14757, 0.15376, 0.16011, 0.16661, 0.17325, 0.18005, 0.18699,
	0.19407, 0.20130, 0.20867, 0.21618, 0.22382, 0.23161, 0.23952, 0.24757,
	0.25574, 0.26404, 0.27246, 0.28100, 0.28965, 0.29841, 0.30729, 0.31626,
	0.32533, 0.33450, 0.34376, 0.35311, 0.36253, 0.37204, 0.38161, 0.39126,
	0.40096, 0.41072, 0.42054, 0.43040, 0.44030, 0.45023, 0.46020, 0.47019,
	0.48020, 0.49022, 0.50025, 0.51028, 0.52031, 0.53033, 0.54033, 0.55031,
	0.56026, 0.57019, 0.58007, 0.58991, 0.59970, 0.60944, 0.61912, 0.62873,
	0.63827, 0.64774, 0.65713, 0.66643, 0.67564, 0.68476, 0.69377, 0.70269,
	0.71150, 0.72019, 0.72877, 0.73723, 0.74557, 0.75378, 0.76186, 0.76981,
	0.77762, 0.78530, 0.79283, 0.80022, 0.80747, 0.81457, 0.82151, 0.82831,
	0.83496, 0.84145, 0.84779, 0.85398, 0.86001, 0.86588, 0.87160, 0.87716,
	0.88257, 0.88782, 0.89291, 0.89785, 0.90264, 0.90728, 0.91176, 0.91610,
	0.92028, 0.92432, 0.92822, 0.93197, 0.93558, 0.93906, 0.94240, 0.94560,
	0.94867, 0.95162, 0.95444, 0.95713, 0.95971, 0.96217, 0.96451, 0.96674,
	0.96887, 0.97089, 0.97281, 0.97463, 0.97635, 0.97799, 0.97953, 0.98099,
	0.98236, 0.98366, 0.98488, 0.98602, 0.98710, 0.98811, 0.98905, 0.98994,
	0.99076, 0.99153, 0.99225, 0.99291, 0.99353, 0.99411, 0.99464, 0.99513,
	0.99558, 0.99600, 0.99639, 0.99674, 0.99706, 0.99736, 0.99763, 0.99788,
	0.99811, 0.99831, 0.99850, 0.99867, 0.99882, 0.99895, 0.99908, 0.99919,
	0.99929, 0.99938, 0.99946, 0.99953, 0.99959, 0.99965, 0.99969, 0.99974,
	0.99978, 0.99981, 0.99984, 0.99986, 0.99988, 0.99990, 0.99992, 0.99993,
	0.99994, 0.99995, 0.99996, 0.99997, 0.99998, 0.99998, 0.99998, 0.99999,
	0.99999, 0.99999, 0.99999, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000,
	1.00000, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000 };


static inline void swap_cmplx(complex_t *a, complex_t *b)
{
    complex_t tmp;

    tmp = *a;
    *a = *b;
    *b = tmp;
}



static inline complex_t cmplx_mult(complex_t a, complex_t b)
{
    complex_t ret;

    ret.real = a.real * b.real - a.imag * b.imag;
    ret.imag = a.real * b.imag + a.imag * b.real;

    return ret;
}

void
imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
{
    int i,k;
    int p,q;
    int m;
    int two_m;
    int two_m_plus_one;

    sample_t tmp_a_i;
    sample_t tmp_a_r;
    sample_t tmp_b_i;
    sample_t tmp_b_r;

    sample_t *data_ptr;
    sample_t *delay_ptr;
    sample_t *window_ptr;
	
    /* 512 IMDCT with source and dest data in 'data' */
	
    /* Pre IFFT complex multiply plus IFFT cmplx conjugate */
    for( i=0; i < 128; i++) {
	/* z[i] = (X[256-2*i-1] + j * X[2*i]) * (xcos1[i] + j * xsin1[i]) ; */ 
	buf[i].real =         (data[256-2*i-1] * xcos1[i])  -  (data[2*i]       * xsin1[i]);
	buf[i].imag = -1.0 * ((data[2*i]       * xcos1[i])  +  (data[256-2*i-1] * xsin1[i]));
    }

    /* Bit reversed shuffling */
    for(i=0; i<128; i++) {
	k = bit_reverse_512[i];
	if (k < i)
	    swap_cmplx(&buf[i],&buf[k]);
    }

    /* FFT Merge */
    for (m=0; m < 7; m++) {
	if(m)
	    two_m = (1 << m);
	else
	    two_m = 1;

	two_m_plus_one = (1 << (m+1));

	for(k = 0; k < two_m; k++) {
	    for(i = 0; i < 128; i += two_m_plus_one) {
		p = k + i;
		q = p + two_m;
		tmp_a_r = buf[p].real;
		tmp_a_i = buf[p].imag;
		tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag;
		tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag;
		buf[p].real = tmp_a_r + tmp_b_r;
		buf[p].imag =  tmp_a_i + tmp_b_i;
		buf[q].real = tmp_a_r - tmp_b_r;
		buf[q].imag =  tmp_a_i - tmp_b_i;
	    }
	}
    }

    /* Post IFFT complex multiply  plus IFFT complex conjugate*/
    for( i=0; i < 128; i++) {
	/* y[n] = z[n] * (xcos1[n] + j * xsin1[n]) ; */
	tmp_a_r =        buf[i].real;
	tmp_a_i = -1.0 * buf[i].imag;
	buf[i].real =(tmp_a_r * xcos1[i])  -  (tmp_a_i  * xsin1[i]);
	buf[i].imag =(tmp_a_r * xsin1[i])  +  (tmp_a_i  * xcos1[i]);
    }
	
    data_ptr = data;
    delay_ptr = delay;
    window_ptr = imdct_window;

    /* Window and convert to real valued signal */
    for(i=0; i< 64; i++) { 
	*data_ptr++   = -buf[64+i].imag   * *window_ptr++ + *delay_ptr++ + bias; 
	*data_ptr++   =  buf[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias; 
    }

    for(i=0; i< 64; i++) { 
	*data_ptr++  = -buf[i].real       * *window_ptr++ + *delay_ptr++ + bias; 
	*data_ptr++  =  buf[128-i-1].imag * *window_ptr++ + *delay_ptr++ + bias; 
    }

    /* The trailing edge of the window goes into the delay line */
    delay_ptr = delay;

    for(i=0; i< 64; i++) { 
	*delay_ptr++  = -buf[64+i].real   * *--window_ptr; 
	*delay_ptr++  =  buf[64-i-1].imag * *--window_ptr; 
    }

    for(i=0; i<64; i++) {
	*delay_ptr++  =  buf[i].imag       * *--window_ptr; 
	*delay_ptr++  = -buf[128-i-1].real * *--window_ptr; 
    }
}

void
imdct_do_256(sample_t data[],sample_t delay[],sample_t bias)
{
    int i,k;
    int p,q;
    int m;
    int two_m;
    int two_m_plus_one;

    sample_t tmp_a_i;
    sample_t tmp_a_r;
    sample_t tmp_b_i;
    sample_t tmp_b_r;

    sample_t *data_ptr;
    sample_t *delay_ptr;
    sample_t *window_ptr;

    complex_t *buf_1, *buf_2;

    buf_1 = &buf[0];
    buf_2 = &buf[64];

    /* Pre IFFT complex multiply plus IFFT cmplx conjugate */
    for(k=0; k<64; k++) { 
	/* X1[k] = X[2*k]  */
	/* X2[k] = X[2*k+1]     */

	p = 2 * (128-2*k-1);
	q = 2 * (2 * k);

	/* Z1[k] = (X1[128-2*k-1] + j * X1[2*k]) * (xcos2[k] + j * xsin2[k]); */ 
	buf_1[k].real =         data[p] * xcos2[k] - data[q] * xsin2[k];
	buf_1[k].imag = -1.0f * (data[q] * xcos2[k] + data[p] * xsin2[k]); 
	/* Z2[k] = (X2[128-2*k-1] + j * X2[2*k]) * (xcos2[k] + j * xsin2[k]); */ 
	buf_2[k].real =          data[p + 1] * xcos2[k] - data[q + 1] * xsin2[k];
	buf_2[k].imag = -1.0f * ( data[q + 1] * xcos2[k] + data[p + 1] * xsin2[k]); 
    }

    /* IFFT Bit reversed shuffling */
    for(i=0; i<64; i++) { 
	k = bit_reverse_256[i];
	if (k < i) {
	    swap_cmplx(&buf_1[i],&buf_1[k]);
	    swap_cmplx(&buf_2[i],&buf_2[k]);
	}
    }

    /* FFT Merge */
    for (m=0; m < 6; m++) {
	two_m = (1 << m);
	two_m_plus_one = (1 << (m+1));

	/* FIXME */
	if(m)
	    two_m = (1 << m);
	else
	    two_m = 1;

	for(k = 0; k < two_m; k++) {
	    for(i = 0; i < 64; i += two_m_plus_one) {
		p = k + i;
		q = p + two_m;
		/* Do block 1 */
		tmp_a_r = buf_1[p].real;
		tmp_a_i = buf_1[p].imag;
		tmp_b_r = buf_1[q].real * w[m][k].real - buf_1[q].imag * w[m][k].imag;
		tmp_b_i = buf_1[q].imag * w[m][k].real + buf_1[q].real * w[m][k].imag;
		buf_1[p].real = tmp_a_r + tmp_b_r;
		buf_1[p].imag =  tmp_a_i + tmp_b_i;
		buf_1[q].real = tmp_a_r - tmp_b_r;
		buf_1[q].imag =  tmp_a_i - tmp_b_i;

		/* Do block 2 */
		tmp_a_r = buf_2[p].real;
		tmp_a_i = buf_2[p].imag;
		tmp_b_r = buf_2[q].real * w[m][k].real - buf_2[q].imag * w[m][k].imag;
		tmp_b_i = buf_2[q].imag * w[m][k].real + buf_2[q].real * w[m][k].imag;
		buf_2[p].real = tmp_a_r + tmp_b_r;
		buf_2[p].imag =  tmp_a_i + tmp_b_i;
		buf_2[q].real = tmp_a_r - tmp_b_r;
		buf_2[q].imag =  tmp_a_i - tmp_b_i;
	    }
	}
    }

    /* Post IFFT complex multiply */
    for( i=0; i < 64; i++) {
	/* y1[n] = z1[n] * (xcos2[n] + j * xs in2[n]) ; */ 
	tmp_a_r =  buf_1[i].real;
	tmp_a_i = -buf_1[i].imag;
	buf_1[i].real =(tmp_a_r * xcos2[i])  -  (tmp_a_i  * xsin2[i]);
	buf_1[i].imag =(tmp_a_r * xsin2[i])  +  (tmp_a_i  * xcos2[i]);
	/* y2[n] = z2[n] * (xcos2[n] + j * xsin2[n]) ; */ 
	tmp_a_r =  buf_2[i].real;
	tmp_a_i = -buf_2[i].imag;
	buf_2[i].real =(tmp_a_r * xcos2[i])  -  (tmp_a_i  * xsin2[i]);
	buf_2[i].imag =(tmp_a_r * xsin2[i])  +  (tmp_a_i  * xcos2[i]);
    }
	
    data_ptr = data;
    delay_ptr = delay;
    window_ptr = imdct_window;

    /* Window and convert to real valued signal */
    for(i=0; i< 64; i++) { 
	*data_ptr++  = -buf_1[i].imag      * *window_ptr++ + *delay_ptr++ + bias;
	*data_ptr++  =  buf_1[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias;
    }

    for(i=0; i< 64; i++) {
	*data_ptr++  = -buf_1[i].real      * *window_ptr++ + *delay_ptr++ + bias;
	*data_ptr++  =  buf_1[64-i-1].imag * *window_ptr++ + *delay_ptr++ + bias;
    }
	
    delay_ptr = delay;

    for(i=0; i< 64; i++) {
	*delay_ptr++ = -buf_2[i].real      * *--window_ptr;
	*delay_ptr++ =  buf_2[64-i-1].imag * *--window_ptr;
    }

    for(i=0; i< 64; i++) {
	*delay_ptr++ =  buf_2[i].imag      * *--window_ptr;
	*delay_ptr++ = -buf_2[64-i-1].real * *--window_ptr;
    }
}

void imdct_init (uint32_t mm_accel)
{
#ifdef LIBA52_MLIB
    if (mm_accel & MM_ACCEL_MLIB) {
        fprintf (stderr, "Using mlib for IMDCT transform\n");
	imdct_512 = imdct_do_512_mlib;
	imdct_256 = imdct_do_256_mlib;
    } else
#endif
    {
	int i, j, k;

	fprintf (stderr, "No accelerated IMDCT transform found\n");

	/* Twiddle factors to turn IFFT into IMDCT */
	for (i = 0; i < 128; i++) {
	    xcos1[i] = -cos ((M_PI / 2048) * (8 * i + 1));
	    xsin1[i] = -sin ((M_PI / 2048) * (8 * i + 1));
	}

	/* More twiddle factors to turn IFFT into IMDCT */
	for (i = 0; i < 64; i++) {
	    xcos2[i] = -cos ((M_PI / 1024) * (8 * i + 1));
	    xsin2[i] = -sin ((M_PI / 1024) * (8 * i + 1));
	}

	for (i = 0; i < 7; i++) {
	    j = 1 << i;
	    for (k = 0; k < j; k++) {
		w[i][k].real = cos (-M_PI * k / j);
		w[i][k].imag = sin (-M_PI * k / j);
	    }
	}
	imdct_512 = imdct_do_512;
	imdct_256 = imdct_do_256;
    }
}

--- NEW FILE ---
/*
 * imdct_mlib.c
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#ifdef LIBA52_MLIB

#include <mlib_types.h>
#include <mlib_status.h>
#include <mlib_signal.h>
#include <string.h>
#include <inttypes.h>

#include "a52.h"
#include "a52_internal.h"
#include "attributes.h"

extern sample_t imdct_window[];

void
imdct_do_512_mlib(sample_t data[], sample_t delay[], sample_t bias)
{
	sample_t *buf_real;
	sample_t *buf_imag;
	sample_t *data_ptr;
	sample_t *delay_ptr;
	sample_t *window_ptr;
	sample_t tmp[256] ATTR_ALIGN (16);
	int i;
	
	memcpy(tmp, data, 256 * sizeof(sample_t));
	mlib_SignalIMDCT_F32(tmp);
  
	buf_real = tmp;
	buf_imag = tmp + 128;
	data_ptr = data;
	delay_ptr = delay;
	window_ptr = imdct_window;

	/* Window and convert to real valued signal */
	for(i=0; i< 64; i++) 
	{ 
		*data_ptr++ = -buf_imag[64+i]   * *window_ptr++ + *delay_ptr++ + bias; 
		*data_ptr++ =  buf_real[64-i-1] * *window_ptr++ + *delay_ptr++ + bias; 
	}

	for(i=0; i< 64; i++) 
	{ 
		*data_ptr++ = -buf_real[i]       * *window_ptr++ + *delay_ptr++ + bias; 
		*data_ptr++ =  buf_imag[128-i-1] * *window_ptr++ + *delay_ptr++ + bias; 
	}
	
	/* The trailing edge of the window goes into the delay line */
	delay_ptr = delay;

	for(i=0; i< 64; i++) 
	{ 
		*delay_ptr++ = -buf_real[64+i]   * *--window_ptr; 
		*delay_ptr++ =  buf_imag[64-i-1] * *--window_ptr; 
	}

	for(i=0; i<64; i++) 
	{
		*delay_ptr++ =  buf_imag[i]       * *--window_ptr; 
		*delay_ptr++ = -buf_real[128-i-1] * *--window_ptr; 
	}
}

void
imdct_do_256_mlib(sample_t data[], sample_t delay[], sample_t bias)
{
	sample_t *buf1_real, *buf1_imag;
	sample_t *buf2_real, *buf2_imag;
	sample_t *data_ptr;
	sample_t *delay_ptr;
	sample_t *window_ptr;
	sample_t tmp[256] ATTR_ALIGN (16);
	int i;
	
	memcpy(tmp, data, 256 * sizeof(sample_t));
	mlib_SignalIMDCTSplit_F32(tmp);
  
	buf1_real = tmp;
	buf1_imag = tmp + 128 + 64;
	buf2_real = tmp + 64;
	buf2_imag = tmp + 128;
	data_ptr = data;
	delay_ptr = delay;
	window_ptr = imdct_window;

	/* Window and convert to real valued signal */
	for(i=0; i< 64; i++) 
	{
		*data_ptr++ = -buf1_imag[i]      * *window_ptr++ + *delay_ptr++ + bias;
		*data_ptr++ =  buf1_real[64-i-1] * *window_ptr++ + *delay_ptr++ + bias;
	}

	for(i=0; i< 64; i++) 
	{
		*data_ptr++ = -buf1_real[i]      * *window_ptr++ + *delay_ptr++ + bias;
		*data_ptr++ =  buf1_imag[64-i-1] * *window_ptr++ + *delay_ptr++ + bias;
	}
	
	delay_ptr = delay;

	for(i=0; i< 64; i++) 
	{
		*delay_ptr++ = -buf2_real[i]      * *--window_ptr;
		*delay_ptr++ =  buf2_imag[64-i-1] * *--window_ptr;
	}

	for(i=0; i< 64; i++) 
	{
		*delay_ptr++ =  buf2_imag[i]      * *--window_ptr;
		*delay_ptr++ = -buf2_real[64-i-1] * *--window_ptr;
	}
}

#endif

--- NEW FILE ---
Using the liba52 API
--------------------

liba52 provides a low-level interface to decoding audio frames encoded
using ATSC standard A/52 aka AC-3. liba52 provides downmixing and
dynamic range compression for the following output configurations:

A52_CHANNEL  : Dual mono. Two independant mono channels.
A52_CHANNEL1 : First of the two mono channels above.
A52_CHANNEL2 : Second of the two mono channels above.
A52_MONO     : Mono.
A52_STEREO   : Stereo.
A52_DOLBY    : Dolby surround compatible stereo.
A52_3F       : 3 front channels (left, center, right)
A52_2F1R     : 2 front, 1 rear surround channel (L, R, S)
A52_3F1R     : 3 front, 1 rear surround channel (L, C, R, S)
A52_2F2R     : 2 front, 2 rear surround channels (L, R, LS, RS)
A52_3F2R     : 3 front, 2 rear surround channels (L, C, R, LS, RS)

A52_LFE      : Low frequency effects channel. Normally used to connect a
               subwoofer. Can be combined with any of the above channels.
               For example: A52_3F2R | A52_LFE -> 3 front, 2 rear, 1 LFE (5.1)


Initialization
--------------

sample_t * a52_init (uint32_t mm_accel);

Initializes the A/52 library. Takes as a parameter the acceptable
optimizations which may be used, such as MMX. These are found in the
included header file 'mm_accel', along with an autodetection function
(mm_accel()). Currently, the only accelleration implemented is
MM_ACCEL_MLIB, which uses the 'mlib' library if installed. mlib is
only available on some Sun Microsystems platforms.

The return value is a pointer to a properly-aligned sample buffer used
for output samples.


Probing the bitstream
---------------------

int a52_syncinfo (uint8_t * buf, int * flags,
                  int * sample_rate, int * bit_rate);

The A/52 bitstream is composed of several a52 frames concatenated one
after each other. An a52 frame is the smallest independantly decodable
unit in the stream.

buf must contain at least 7 bytes from the input stream. If these look
like the start of a valid a52 frame, a52_syncinfo() returns the size
of the coded frame in bytes, and fills flags, sample_rate and bit_rate
with the information encoded in the stream. The returned size is
guaranteed to be an even number between 128 and 3840. sample_rate will
be the sampling frequency in Hz, bit_rate is for the compressed stream
and is in bits per second, and flags is a description of the coded
channels: the A52_LFE bit is set if there is an LFE channel coded in
this stream, and by masking flags with A52_CHANNEL_MASK you will get a
value that describes the full-bandwidth channels, as one of the
A52_CHANNEL...A52_3F2R flags.

If this can not possibly be a valid frame, then the function returns
0. You should then try to re-synchronize with the a52 stream - one way
to try this would be to advance buf by one byte until its contents
looks like a valid frame, but there might be better
application-specific ways to synchronize.

It is recommended to call this function for each frame, for several
reasons: this function detects errors that the other functions will
not double-check, consecutive frames might have different lengths, and
it helps you re-sync with the stream if you get de-synchronized.


Starting to decode a frame
--------------------------

int a52_frame (a52_state_t * state, uint8_t * buf, int * flags,
	       sample_t * level, sample_t bias);

This starts the work of decoding the A/52 frame (to be completed using
a52_block()). buf should point to the beginning of the complete frame
of the full size returned by a52_syncinfo().

You should pass in the flags the speaker configuration that you
support, and liba52 will return the speaker configuration it will use
for its output, based on what is coded in the stream and what you
asked for. For example, if the stream contains 2+2 channels
(a52_syncinfo() returned A52_2F2R in the flags), and you have 3+1
speakers (you passed A52_3F1R), then liba52 will choose do downmix to
2+1 speakers, since there is no center channel to send to your center
speaker. So in that case the left and right channels will be
essentially unmodified by the downmix, and the two surround channels
will be added together and sent to your surround speaker. liba52 will
return A52_2F1R to indicate this.

The good news is that when you downmix to stereo you dont have to
worry about this, you will ALWAYS get a stereo output no matter what
was coded in the stream. For more complex output configurations you
will have to handle the case where liba52 couldnt give you what you
wanted because some of the channels were not encoded in the stream
though.

Level, bias, and A52_ADJUST_LEVEL:

Before downmixing, samples are floating point values with a range of
[-1,1]. Most types of downmixing will combine channels together, which
will potentially result in a larger range for the output
samples. liba52 provides two methods of controlling the range of the
output, either before or after the downmix stage.

If you do not set A52_ADJUST_LEVEL, liba52 will multiply the samples
by your level value, so that they fit in the [-level,level]
range. Then it will apply the standardized downmix equations,
potentially making the samples go out of that interval again. The
level parameter is not modified.

Setting the A52_ADJUST_LEVEL flag will instruct liba52 to treat your
level value as the intended range interval after downmixing. It will
then figure out what level to use before the downmix (what you should
have passed if you hadnt used the A52_ADJUST_LEVEL flag), and
overwrite the level value you gave it with that new level value.

The bias represents a value which should be added to the result
regardless:

output_sample = (input_sample * level) + bias;

For example, a bias of 384 and a level of 1 tells liba52 you want
samples between 383 and 385 instead of -1 and 1. This is what the
sample program a52dec does, as it makes it faster to convert the
samples to integer format, using a trick based on the IEEE
floating-point format.

This function also initialises the state for that frame, which will be
reused next when decoding blocks.


Dynamic range compression
-------------------------

void a52_dynrng (a52_state_t * state,
                 sample_t (* call) (sample_t, void *), void * data);

This function is purely optional. If you dont call it, liba52 will
provide the default behaviour, which is to apply the full dynamic
range compression as specified in the A/52 stream. This basically
makes the loud sounds softer, and the soft sounds louder, so you can
more easily listen to the stream in a noisy environment without
disturbing anyone.

If you do call this function and set a NULL callback, this will
totally disable the dynamic range compression and provide a playback
more adapted to a movie theater or a listening room.

If you call this function and specify a callback function, this
callback might be called up to once for each block, with two
arguments: the compression factor 'c' recommended by the bitstream,
and the private data pointer you specified in a52_dynrng(). The
callback will then return the amount of compression to actually use -
typically pow(c,x) where x is somewhere between 0 and 1. More
elaborate compression functions might want to use a different value
for 'x' depending wether c>1 or c<1 - or even something more complex
if this is what you want.


Decoding blocks
---------------

int a52_block (a52_state_t * state, sample_t * samples);

Every A/52 frame is composed of 6 blocks, each with an output of 256
samples for each channel. The a52_block() function decodes the next
block in the frame, and should be called 6 times to decode all of the
audio in the frame. After each call, you should extract the audio data
from the sample buffer.

The sample pointer given should be the one a52_init() returned.

After this function returns, the samples buuffer will contain 256
samples for the first channel, followed by 256 samples for the second
channel, etc... the channel order is LFE, left, center, right, left
surround, right surround. If one of the channels is not present in the
liba52 output, as indicated by the flags returned by a52_frame(), then
this channel is skipped and the following channels are shifted so
liba52 does not leave an empty space between channels.


Pseudocode example
------------------

sample_t * samples = a52_init (mm_accel());

loop on input bytes:
  if at least 7 bytes in the buffer:

    bytes_to_get = a52_syncinfo (...)

    if bytes_to_get == 0:
      goto loop to keep looking for sync point
    else
      get rest of bytes

      a52_frame (state, buf, ...)
      [a52_dynrng (state, ...); this is only optional]
      for i = 1 ... 6:
        a52_block (state, samples)
        convert samples to integer and queue to soundcard

--- NEW FILE ---
/*
 * mm_accel.h
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#ifndef MM_ACCEL_H
#define MM_ACCEL_H

/* generic accelerations */
#define MM_ACCEL_MLIB		0x00000001

/* x86 accelerations */
#define MM_ACCEL_X86_MMX	0x80000000
#define MM_ACCEL_X86_3DNOW	0x40000000
#define MM_ACCEL_X86_MMXEXT	0x20000000

uint32_t mm_accel (void);

#endif /* MM_ACCEL_H */

--- NEW FILE ---
/*
 * parse.c
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#include <stdlib.h>
#include <string.h>
#include <inttypes.h>

#include "a52.h"
#include "a52_internal.h"
#include "bitstream.h"
#include "tables.h"

#ifdef HAVE_MEMALIGN
/* some systems have memalign() but no declaration for it */
void * memalign (size_t align, size_t size);
#else
/* assume malloc alignment is sufficient */
#define memalign(align,size) malloc (size)
#endif

typedef struct {
    sample_t q1[2];
    sample_t q2[2];
    sample_t q4;
    int q1_ptr;
    int q2_ptr;
    int q4_ptr;
} quantizer_t;

static uint8_t halfrate[12] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3};

sample_t * a52_init (uint32_t mm_accel)
{
    sample_t * samples;
    int i;

    imdct_init (mm_accel);

    samples = memalign (16, 256 * 12 * sizeof (sample_t));
    if (samples == NULL)
	return NULL;

    for (i = 0; i < 256 * 12; i++)
	samples[i] = 0;

    return samples;
}

int a52_syncinfo (uint8_t * buf, int * flags,
		  int * sample_rate, int * bit_rate)
{
    static int rate[] = { 32,  40,  48,  56,  64,  80,  96, 112,
			 128, 160, 192, 224, 256, 320, 384, 448,
			 512, 576, 640};
    static uint8_t lfeon[8] = {0x10, 0x10, 0x04, 0x04, 0x04, 0x01, 0x04, 0x01};
    int frmsizecod;
    int bitrate;
    int half;
    int acmod;

    if ((buf[0] != 0x0b) || (buf[1] != 0x77))	/* syncword */
	return 0;

    if (buf[5] >= 0x60)		/* bsid >= 12 */
	return 0;
    half = halfrate[buf[5] >> 3];

    /* acmod, dsurmod and lfeon */
    acmod = buf[6] >> 5;
    *flags = ((((buf[6] & 0xf8) == 0x50) ? A52_DOLBY : acmod) |
	      ((buf[6] & lfeon[acmod]) ? A52_LFE : 0));

    frmsizecod = buf[4] & 63;
    if (frmsizecod >= 38)
	return 0;
    bitrate = rate [frmsizecod >> 1];
    *bit_rate = (bitrate * 1000) >> half;

    switch (buf[4] & 0xc0) {
    case 0:	/* 48 KHz */
	*sample_rate = 48000 >> half;
	return 4 * bitrate;
    case 0x40:
	*sample_rate = 44100 >> half;
	return 2 * (320 * bitrate / 147 + (frmsizecod & 1));
    case 0x80:
	*sample_rate = 32000 >> half;
	return 6 * bitrate;
    default:
	return 0;
    }
}

int a52_frame (a52_state_t * state, uint8_t * buf, int * flags,
	       sample_t * level, sample_t bias)
{
    static sample_t clev[4] = {LEVEL_3DB, LEVEL_45DB, LEVEL_6DB, LEVEL_45DB};
    static sample_t slev[4] = {LEVEL_3DB, LEVEL_6DB, 0, LEVEL_6DB};
    int chaninfo;
    int acmod;

    state->fscod = buf[4] >> 6;
    state->halfrate = halfrate[buf[5] >> 3];
    state->acmod = acmod = buf[6] >> 5;

    bitstream_set_ptr (buf + 6);
    bitstream_get (3);	/* skip acmod we already parsed */

    if ((acmod == 2) && (bitstream_get (2) == 2))	/* dsurmod */
	acmod = A52_DOLBY;

    if ((acmod & 1) && (acmod != 1))
	state->clev = clev[bitstream_get (2)];	/* cmixlev */

    if (acmod & 4)
	state->slev = slev[bitstream_get (2)];	/* surmixlev */

    state->lfeon = bitstream_get (1);

    state->output = downmix_init (acmod, *flags, level,
				  state->clev, state->slev);
    if (state->output < 0)
	return 1;
    if (state->lfeon && (*flags & A52_LFE))
	state->output |= A52_LFE;
    *flags = state->output;
    /* the 2* compensates for differences in imdct */
    state->dynrng = state->level = 2 * *level;
    state->bias = bias;
    state->dynrnge = 1;
    state->dynrngcall = NULL;

    chaninfo = !acmod;
    do {
	bitstream_get (5);	/* dialnorm */
	if (bitstream_get (1))	/* compre */
	    bitstream_get (8);	/* compr */
	if (bitstream_get (1))	/* langcode */
	    bitstream_get (8);	/* langcod */
	if (bitstream_get (1))	/* audprodie */
	    bitstream_get (7);	/* mixlevel + roomtyp */
    } while (chaninfo--);

    bitstream_get (2);		/* copyrightb + origbs */

    if (bitstream_get (1))	/* timecod1e */
	bitstream_get (14);	/* timecod1 */
    if (bitstream_get (1))	/* timecod2e */
	bitstream_get (14);	/* timecod2 */

    if (bitstream_get (1)) {	/* addbsie */
	int addbsil;

	addbsil = bitstream_get (6);
	do {
	    bitstream_get (8);	/* addbsi */
	} while (addbsil--);
    }

    return 0;
}

void a52_dynrng (a52_state_t * state,
		 sample_t (* call) (sample_t, void *), void * data)
{
    state->dynrnge = 0;
    if (call) {
	state->dynrnge = 1;
	state->dynrngcall = call;
	state->dynrngdata = data;
    }
}

static int parse_exponents (int expstr, int ngrps, uint8_t exponent,
			    uint8_t * dest)
{
    int exps;

    while (ngrps--) {
	exps = bitstream_get (7);

	exponent += exp_1[exps];
	if (exponent > 24)
	    return 1;

	switch (expstr) {
	case EXP_D45:
	    *(dest++) = exponent;
	    *(dest++) = exponent;
	case EXP_D25:
	    *(dest++) = exponent;
	case EXP_D15:
	    *(dest++) = exponent;
	}

	exponent += exp_2[exps];
	if (exponent > 24)
	    return 1;

	switch (expstr) {
	case EXP_D45:
	    *(dest++) = exponent;
	    *(dest++) = exponent;
	case EXP_D25:
	    *(dest++) = exponent;
	case EXP_D15:
	    *(dest++) = exponent;
	}

	exponent += exp_3[exps];
	if (exponent > 24)
	    return 1;

	switch (expstr) {
	case EXP_D45:
	    *(dest++) = exponent;
	    *(dest++) = exponent;
	case EXP_D25:
	    *(dest++) = exponent;
	case EXP_D15:
	    *(dest++) = exponent;
	}
    }	

    return 0;
}

static int parse_deltba (int8_t * deltba)
{
    int deltnseg, deltlen, delta, j;

    memset (deltba, 0, 50);

    deltnseg = bitstream_get (3);
    j = 0;
    do {
	j += bitstream_get (5);
	deltlen = bitstream_get (4);
	delta = bitstream_get (3);
	delta -= (delta >= 4) ? 3 : 4;
	if (!deltlen)
	    continue;
	if (j + deltlen >= 50)
	    return 1;
	while (deltlen--)
	    deltba[j++] = delta;
    } while (deltnseg--);

    return 0;
}

static inline int zero_snr_offsets (int nfchans, a52_state_t * state)
{
    int i;

    if ((state->csnroffst) || (state->cplinu && state->cplba.fsnroffst) ||
	(state->lfeon && state->lfeba.fsnroffst))
	return 0;
    for (i = 0; i < nfchans; i++)
	if (state->ba[i].fsnroffst)
	    return 0;
    return 1;
}

static inline int16_t dither_gen (void)
{
    static uint16_t lfsr_state = 1;
    int16_t state;

    state = dither_lut[lfsr_state >> 8] ^ (lfsr_state << 8);
	
    lfsr_state = (uint16_t) state;

    return state;
}

static void coeff_get (sample_t * coeff, uint8_t * exp, int8_t * bap,
		       quantizer_t * quantizer, sample_t level,
		       int dither, int end)
{
    int i;
    sample_t factor[25];

    for (i = 0; i <= 24; i++)
	factor[i] = scale_factor[i] * level;

    for (i = 0; i < end; i++) {
	int bapi;

	bapi = bap[i];
	switch (bapi) {
	case 0:
	    if (dither) {
		coeff[i] = dither_gen() * LEVEL_3DB * factor[exp[i]];
		continue;
	    } else {
		coeff[i] = 0;
		continue;
	    }

	case -1:
	    if (quantizer->q1_ptr >= 0) {
		coeff[i] = quantizer->q1[quantizer->q1_ptr--] * factor[exp[i]];
		continue;
	    } else {
		int code;

		code = bitstream_get (5);

		quantizer->q1_ptr = 1;
		quantizer->q1[0] = q_1_2[code];
		quantizer->q1[1] = q_1_1[code];
		coeff[i] = q_1_0[code] * factor[exp[i]];
		continue;
	    }

	case -2:
	    if (quantizer->q2_ptr >= 0) {
		coeff[i] = quantizer->q2[quantizer->q2_ptr--] * factor[exp[i]];
		continue;
	    } else {
		int code;

		code = bitstream_get (7);

		quantizer->q2_ptr = 1;
		quantizer->q2[0] = q_2_2[code];
		quantizer->q2[1] = q_2_1[code];
		coeff[i] = q_2_0[code] * factor[exp[i]];
		continue;
	    }

	case 3:
	    coeff[i] = q_3[bitstream_get (3)] * factor[exp[i]];
	    continue;

	case -3:
	    if (quantizer->q4_ptr == 0) {
		quantizer->q4_ptr = -1;
		coeff[i] = quantizer->q4 * factor[exp[i]];
		continue;
	    } else {
		int code;

		code = bitstream_get (7);

		quantizer->q4_ptr = 0;
		quantizer->q4 = q_4_1[code];
		coeff[i] = q_4_0[code] * factor[exp[i]];
		continue;
	    }

	case 4:
	    coeff[i] = q_5[bitstream_get (4)] * factor[exp[i]];
	    continue;

	default:
	    coeff[i] = ((bitstream_get_2 (bapi) << (16 - bapi)) *
			  factor[exp[i]]);
	}
    }
}

static void coeff_get_coupling (a52_state_t * state, int nfchans,
				sample_t * coeff, sample_t (* samples)[256],
				quantizer_t * quantizer, uint8_t dithflag[5])
{
    int sub_bnd, bnd, i, i_end, ch;
    int8_t * bap;
    uint8_t * exp;
    sample_t cplco[5];

    bap = state->cpl_bap;
    exp = state->cpl_exp;
    sub_bnd = bnd = 0;
    i = state->cplstrtmant;
    while (i < state->cplendmant) {
	i_end = i + 12;
	while (state->cplbndstrc[sub_bnd++])
	    i_end += 12;
	for (ch = 0; ch < nfchans; ch++)
	    cplco[ch] = state->cplco[ch][bnd] * coeff[ch];
	bnd++;

	while (i < i_end) {
	    sample_t cplcoeff;
	    int bapi;

	    bapi = bap[i];
	    switch (bapi) {
	    case 0:
		cplcoeff = LEVEL_3DB * scale_factor[exp[i]];
		for (ch = 0; ch < nfchans; ch++)
		    if (state->chincpl[ch]) {
			if (dithflag[ch])
			    samples[ch][i] = (cplcoeff * cplco[ch] *
					      dither_gen ());
			else
			    samples[ch][i] = 0;
		    }
		i++;
		continue;

	    case -1:
		if (quantizer->q1_ptr >= 0) {
		    cplcoeff = quantizer->q1[quantizer->q1_ptr--];
		    break;
		} else {
		    int code;

		    code = bitstream_get (5);

		    quantizer->q1_ptr = 1;
		    quantizer->q1[0] = q_1_2[code];
		    quantizer->q1[1] = q_1_1[code];
		    cplcoeff = q_1_0[code];
		    break;
		}

	    case -2:
		if (quantizer->q2_ptr >= 0) {
		    cplcoeff = quantizer->q2[quantizer->q2_ptr--];
		    break;
		} else {
		    int code;

		    code = bitstream_get (7);

		    quantizer->q2_ptr = 1;
		    quantizer->q2[0] = q_2_2[code];
		    quantizer->q2[1] = q_2_1[code];
		    cplcoeff = q_2_0[code];
		    break;
		}

	    case 3:
		cplcoeff = q_3[bitstream_get (3)];
		break;

	    case -3:
		if (quantizer->q4_ptr == 0) {
		    quantizer->q4_ptr = -1;
		    cplcoeff = quantizer->q4;
		    break;
		} else {
		    int code;

		    code = bitstream_get (7);

		    quantizer->q4_ptr = 0;
		    quantizer->q4 = q_4_1[code];
		    cplcoeff = q_4_0[code];
		    break;
		}

	    case 4:
		cplcoeff = q_5[bitstream_get (4)];
		break;

	    default:
		cplcoeff = bitstream_get_2 (bapi) << (16 - bapi);
	    }

	    cplcoeff *= scale_factor[exp[i]];
	    for (ch = 0; ch < nfchans; ch++)
		if (state->chincpl[ch])
		    samples[ch][i] = cplcoeff * cplco[ch];
	    i++;
	}
    }
}

int a52_block (a52_state_t * state, sample_t * samples)
{
    static const uint8_t nfchans_tbl[] = {2, 1, 2, 3, 3, 4, 4, 5, 1, 1, 2};
    static int rematrix_band[4] = {25, 37, 61, 253};
    int i, nfchans, chaninfo;
    uint8_t cplexpstr, chexpstr[5], lfeexpstr, do_bit_alloc, done_cpl;
    uint8_t blksw[5], dithflag[5];
    sample_t coeff[5];
    int chanbias;
    quantizer_t quantizer;

    nfchans = nfchans_tbl[state->acmod];

    for (i = 0; i < nfchans; i++)
	blksw[i] = bitstream_get (1);

    for (i = 0; i < nfchans; i++)
	dithflag[i] = bitstream_get (1);

    chaninfo = !(state->acmod);
    do {
	if (bitstream_get (1)) {	/* dynrnge */
	    int dynrng;

	    dynrng = bitstream_get_2 (8);
	    if (state->dynrnge) {
		sample_t range;

		range = ((((dynrng & 0x1f) | 0x20) << 13) *
			 scale_factor[3 - (dynrng >> 5)]);
		if (state->dynrngcall)
		    range = state->dynrngcall (range, state->dynrngdata);
		state->dynrng = state->level * range;
	    }
	}
    } while (chaninfo--);

    if (bitstream_get (1)) {	/* cplstre */
	state->cplinu = bitstream_get (1);
	if (state->cplinu) {
	    static int bndtab[16] = {31, 35, 37, 39, 41, 42, 43, 44,
				     45, 45, 46, 46, 47, 47, 48, 48};
	    int cplbegf;
	    int cplendf;
	    int ncplsubnd;

	    for (i = 0; i < nfchans; i++)
		state->chincpl[i] = bitstream_get (1);
	    switch (state->acmod) {
	    case 0: case 1:
		return 1;
	    case 2:
		state->phsflginu = bitstream_get (1);
	    }
	    cplbegf = bitstream_get (4);
	    cplendf = bitstream_get (4);

	    if (cplendf + 3 - cplbegf < 0)
		return 1;
	    state->ncplbnd = ncplsubnd = cplendf + 3 - cplbegf;
	    state->cplstrtbnd = bndtab[cplbegf];
	    state->cplstrtmant = cplbegf * 12 + 37;
	    state->cplendmant = cplendf * 12 + 73;

	    for (i = 0; i < ncplsubnd - 1; i++) {
		state->cplbndstrc[i] = bitstream_get (1);
		state->ncplbnd -= state->cplbndstrc[i];
	    }
	    state->cplbndstrc[i] = 0;	/* last value is a sentinel */
	}
    }

    if (state->cplinu) {
	int j, cplcoe;

	cplcoe = 0;
	for (i = 0; i < nfchans; i++)
	    if (state->chincpl[i])
		if (bitstream_get (1)) {	/* cplcoe */
		    int mstrcplco, cplcoexp, cplcomant;

		    cplcoe = 1;
		    mstrcplco = 3 * bitstream_get (2);
		    for (j = 0; j < state->ncplbnd; j++) {
			cplcoexp = bitstream_get (4);
			cplcomant = bitstream_get (4);
			if (cplcoexp == 15)
			    cplcomant <<= 14;
			else
			    cplcomant = (cplcomant | 0x10) << 13;
			state->cplco[i][j] =
			    cplcomant * scale_factor[cplcoexp + mstrcplco];
		    }
		}
	if ((state->acmod == 2) && state->phsflginu && cplcoe)
	    for (j = 0; j < state->ncplbnd; j++)
		if (bitstream_get (1))	/* phsflg */
		    state->cplco[1][j] = -state->cplco[1][j];
    }

    if ((state->acmod == 2) && (bitstream_get (1))) {	/* rematstr */
	int end;

	end = (state->cplinu) ? state->cplstrtmant : 253;
	i = 0;
	do
	    state->rematflg[i] = bitstream_get (1);
	while (rematrix_band[i++] < end);
    }

    cplexpstr = EXP_REUSE;
    lfeexpstr = EXP_REUSE;
    if (state->cplinu)
	cplexpstr = bitstream_get (2);
    for (i = 0; i < nfchans; i++)
	chexpstr[i] = bitstream_get (2);
    if (state->lfeon) 
	lfeexpstr = bitstream_get (1);

    for (i = 0; i < nfchans; i++)
	if (chexpstr[i] != EXP_REUSE) {
	    if (state->cplinu && state->chincpl[i])
		state->endmant[i] = state->cplstrtmant;
	    else {
		int chbwcod;

		chbwcod = bitstream_get (6);
		if (chbwcod > 60)
		    return 1;
		state->endmant[i] = chbwcod * 3 + 73;
	    }
	}

    do_bit_alloc = 0;

    if (cplexpstr != EXP_REUSE) {
	int cplabsexp, ncplgrps;

	do_bit_alloc = 64;
	ncplgrps = ((state->cplendmant - state->cplstrtmant) /
		    (3 << (cplexpstr - 1)));
	cplabsexp = bitstream_get (4) << 1;
	if (parse_exponents (cplexpstr, ncplgrps, cplabsexp,
			     state->cpl_exp + state->cplstrtmant))
	    return 1;
    }
    for (i = 0; i < nfchans; i++)
	if (chexpstr[i] != EXP_REUSE) {
	    int grp_size, nchgrps;

	    do_bit_alloc |= 1 << i;
	    grp_size = 3 << (chexpstr[i] - 1);
	    nchgrps = (state->endmant[i] + grp_size - 4) / grp_size;
	    state->fbw_exp[i][0] = bitstream_get (4);
	    if (parse_exponents (chexpstr[i], nchgrps, state->fbw_exp[i][0],
				 state->fbw_exp[i] + 1))
		return 1;
	    bitstream_get (2);	/* gainrng */
	}
    if (lfeexpstr != EXP_REUSE) {
	do_bit_alloc |= 32;
	state->lfe_exp[0] = bitstream_get (4);
	if (parse_exponents (lfeexpstr, 2, state->lfe_exp[0],
			     state->lfe_exp + 1))
	    return 1;
    }

    if (bitstream_get (1)) {	/* baie */
	do_bit_alloc = -1;
	state->sdcycod = bitstream_get (2);
	state->fdcycod = bitstream_get (2);
	state->sgaincod = bitstream_get (2);
	state->dbpbcod = bitstream_get (2);
	state->floorcod = bitstream_get (3);
    }
    if (bitstream_get (1)) {	/* snroffste */
	do_bit_alloc = -1;
	state->csnroffst = bitstream_get (6);
	if (state->cplinu) {
	    state->cplba.fsnroffst = bitstream_get (4);
	    state->cplba.fgaincod = bitstream_get (3);
	}
	for (i = 0; i < nfchans; i++) {
	    state->ba[i].fsnroffst = bitstream_get (4);
	    state->ba[i].fgaincod = bitstream_get (3);
	}
	if (state->lfeon) {
	    state->lfeba.fsnroffst = bitstream_get (4);
	    state->lfeba.fgaincod = bitstream_get (3);
	}
    }
    if ((state->cplinu) && (bitstream_get (1))) {	/* cplleake */
	do_bit_alloc |= 64;
	state->cplfleak = 2304 - (bitstream_get (3) << 8);
	state->cplsleak = 2304 - (bitstream_get (3) << 8);
    }

    if (bitstream_get (1)) {	/* deltbaie */
	do_bit_alloc = -1;
	if (state->cplinu)
	    state->cplba.deltbae = bitstream_get (2);
	for (i = 0; i < nfchans; i++)
	    state->ba[i].deltbae = bitstream_get (2);
	if (state->cplinu && (state->cplba.deltbae == DELTA_BIT_NEW) &&
	    parse_deltba (state->cplba.deltba))
	    return 1;
	for (i = 0; i < nfchans; i++)
	    if ((state->ba[i].deltbae == DELTA_BIT_NEW) &&
		parse_deltba (state->ba[i].deltba))
		return 1;
    }

    if (do_bit_alloc) {
	if (zero_snr_offsets (nfchans, state)) {
	    memset (state->cpl_bap, 0, sizeof (state->cpl_bap));
	    memset (state->fbw_bap, 0, sizeof (state->fbw_bap));
	    memset (state->lfe_bap, 0, sizeof (state->lfe_bap));
	} else {
	    if (state->cplinu && (do_bit_alloc & 64))
		bit_allocate (state, &state->cplba, state->cplstrtbnd,
			      state->cplstrtmant, state->cplendmant,
			      state->cplfleak, state->cplsleak,
			      state->cpl_exp, state->cpl_bap);
	    for (i = 0; i < nfchans; i++)
		if (do_bit_alloc & (1 << i))
		    bit_allocate (state, state->ba + i, 0, 0,
				  state->endmant[i], 0, 0, state->fbw_exp[i],
				  state->fbw_bap[i]);
	    if (state->lfeon && (do_bit_alloc & 32)) {
		state->lfeba.deltbae = DELTA_BIT_NONE;
		bit_allocate (state, &state->lfeba, 0, 0, 7, 0, 0,
			      state->lfe_exp, state->lfe_bap);
	    }
	}
    }

    if (bitstream_get (1)) {	/* skiple */
	i = bitstream_get (9);	/* skipl */
	while (i--)
	    bitstream_get (8);
    }

    if (state->output & A52_LFE)
	samples += 256;	/* shift for LFE channel */

    chanbias = downmix_coeff (coeff, state->acmod, state->output,
			      state->dynrng, state->clev, state->slev);

    quantizer.q1_ptr = quantizer.q2_ptr = quantizer.q4_ptr = -1;
    done_cpl = 0;

    for (i = 0; i < nfchans; i++) {
	int j;

	coeff_get (samples + 256 * i, state->fbw_exp[i], state->fbw_bap[i],
		   &quantizer, coeff[i], dithflag[i], state->endmant[i]);

	if (state->cplinu && state->chincpl[i]) {
	    if (!done_cpl) {
		done_cpl = 1;
		coeff_get_coupling (state, nfchans, coeff,
				    (sample_t (*)[256])samples, &quantizer,
				    dithflag);
	    }
	    j = state->cplendmant;
	} else
	    j = state->endmant[i];
	do
	    (samples + 256 * i)[j] = 0;
	while (++j < 256);
    }

    if (state->acmod == 2) {
	int j, end, band;

	end = ((state->endmant[0] < state->endmant[1]) ?
	       state->endmant[0] : state->endmant[1]);

	i = 0;
	j = 13;
	do {
	    if (!state->rematflg[i]) {
		j = rematrix_band[i++];
		continue;
	    }
	    band = rematrix_band[i++];
	    if (band > end)
		band = end;
	    do {
		sample_t tmp0, tmp1;

		tmp0 = samples[j];
		tmp1 = (samples+256)[j];
		samples[j] = tmp0 + tmp1;
		(samples+256)[j] = tmp0 - tmp1;
	    } while (++j < band);
	} while (j < end);
    }

    if (state->lfeon) {
	if (state->output & A52_LFE) {
	    coeff_get (samples - 256, state->lfe_exp, state->lfe_bap,
		       &quantizer, state->dynrng, 0, 7);
	    for (i = 7; i < 256; i++)
		(samples-256)[i] = 0;
	    imdct_512 (samples - 256, samples + 1536 - 256, state->bias);
	} else {
	    /* just skip the LFE coefficients */
	    coeff_get (samples + 1280, state->lfe_exp, state->lfe_bap,
		       &quantizer, 0, 0, 7);
	}
    }

    i = 0;
    if (nfchans_tbl[state->output & A52_CHANNEL_MASK] < nfchans)
	for (i = 1; i < nfchans; i++)
	    if (blksw[i] != blksw[0])
		break;

    if (i < nfchans) {
	if (samples[2 * 1536 - 1] == (sample_t)0x776b6e21) {
	    samples[2 * 1536 - 1] = 0;
	    upmix (samples + 1536, state->acmod, state->output);
	}

	for (i = 0; i < nfchans; i++) {
	    sample_t bias;

	    bias = 0;
	    if (!(chanbias & (1 << i)))
		bias = state->bias;

	    if (coeff[i]) {
		if (blksw[i])
		    imdct_256 (samples + 256 * i, samples + 1536 + 256 * i,
			       bias);
		else 
		    imdct_512 (samples + 256 * i, samples + 1536 + 256 * i,
			       bias);
	    } else {
		int j;

		for (j = 0; j < 256; j++)
		    (samples + 256 * i)[j] = bias;
	    }
	}

	downmix (samples, state->acmod, state->output, state->bias,
		 state->clev, state->slev);
    } else {
	nfchans = nfchans_tbl[state->output & A52_CHANNEL_MASK];

	downmix (samples, state->acmod, state->output, 0,
		 state->clev, state->slev);

	if (samples[2 * 1536 - 1] != (sample_t)0x776b6e21) {
	    downmix (samples + 1536, state->acmod, state->output, 0,
		     state->clev, state->slev);
	    samples[2 * 1536 - 1] = (sample_t)0x776b6e21;
	}

	if (blksw[0])
	    for (i = 0; i < nfchans; i++)
		imdct_256 (samples + 256 * i, samples + 1536 + 256 * i,
			   state->bias);
	else 
	    for (i = 0; i < nfchans; i++)
		imdct_512 (samples + 256 * i, samples + 1536 + 256 * i,
			   state->bias);
    }

    return 0;
}

--- NEW FILE ---
/*
 * tables.h
 * Copyright (C) 2000-2001 Michel Lespinasse <walken at zoy.org>
 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma at ess.engr.uvic.ca>
 *
 * This file is part of a52dec, a free ATSC A-52 stream decoder.
 * See http://liba52.sourceforge.net/ for updates.
 *
 * a52dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * a52dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

static int8_t exp_1[128] = {
    -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,
    -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
     2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
    25,25,25
};
static int8_t exp_2[128] = {
    -2,-2,-2,-2,-2,-1,-1,-1,-1,-1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,
    -2,-2,-2,-2,-2,-1,-1,-1,-1,-1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,
    -2,-2,-2,-2,-2,-1,-1,-1,-1,-1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,
    -2,-2,-2,-2,-2,-1,-1,-1,-1,-1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,
    -2,-2,-2,-2,-2,-1,-1,-1,-1,-1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,
    25,25,25
};
static int8_t exp_3[128] = {
    -2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,
    -2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,
    -2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,
    -2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,
    -2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,-2,-1, 0, 1, 2,
    25,25,25
};

#define Q0 ((-2 << 15) / 3.0)
#define Q1 (0)
#define Q2 ((2 << 15) / 3.0)

static const sample_t q_1_0[32] = {
    Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,
    Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,
    Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,
    0,0,0,0,0
};

static const sample_t q_1_1[32] = {
    Q0,Q0,Q0,Q1,Q1,Q1,Q2,Q2,Q2,
    Q0,Q0,Q0,Q1,Q1,Q1,Q2,Q2,Q2,
    Q0,Q0,Q0,Q1,Q1,Q1,Q2,Q2,Q2,
    0,0,0,0,0
};

static const sample_t q_1_2[32] = {
    Q0,Q1,Q2,Q0,Q1,Q2,Q0,Q1,Q2,
    Q0,Q1,Q2,Q0,Q1,Q2,Q0,Q1,Q2,
    Q0,Q1,Q2,Q0,Q1,Q2,Q0,Q1,Q2,
    0,0,0,0,0
};

#undef Q0
#undef Q1
#undef Q2

#define Q0 ((-4 << 15) / 5.0)
#define Q1 ((-2 << 15) / 5.0)
#define Q2 (0)
#define Q3 ((2 << 15) / 5.0)
#define Q4 ((4 << 15) / 5.0)

static const sample_t q_2_0[128] = {
    Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,Q0,
    Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,Q1,
    Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,Q2,
    Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,Q3,
    Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,Q4,
    0,0,0
};

static const sample_t q_2_1[128] = {
    Q0,Q0,Q0,Q0,Q0,Q1,Q1,Q1,Q1,Q1,Q2,Q2,Q2,Q2,Q2,Q3,Q3,Q3,Q3,Q3,Q4,Q4,Q4,Q4,Q4,
    Q0,Q0,Q0,Q0,Q0,Q1,Q1,Q1,Q1,Q1,Q2,Q2,Q2,Q2,Q2,Q3,Q3,Q3,Q3,Q3,Q4,Q4,Q4,Q4,Q4,
    Q0,Q0,Q0,Q0,Q0,Q1,Q1,Q1,Q1,Q1,Q2,Q2,Q2,Q2,Q2,Q3,Q3,Q3,Q3,Q3,Q4,Q4,Q4,Q4,Q4,
    Q0,Q0,Q0,Q0,Q0,Q1,Q1,Q1,Q1,Q1,Q2,Q2,Q2,Q2,Q2,Q3,Q3,Q3,Q3,Q3,Q4,Q4,Q4,Q4,Q4,
    Q0,Q0,Q0,Q0,Q0,Q1,Q1,Q1,Q1,Q1,Q2,Q2,Q2,Q2,Q2,Q3,Q3,Q3,Q3,Q3,Q4,Q4,Q4,Q4,Q4,
    0,0,0
};

static const sample_t q_2_2[128] = {
    Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,
    Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,
    Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,
    Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,
    Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,Q0,Q1,Q2,Q3,Q4,
    0,0,0
};

#undef Q0
#undef Q1
#undef Q2
#undef Q3
#undef Q4

static const sample_t q_3[8] = {
    (-6 << 15)/7.0, (-4 << 15)/7.0, (-2 << 15)/7.0, 0,
    ( 2 << 15)/7.0, ( 4 << 15)/7.0, ( 6 << 15)/7.0, 0
};

#define Q0 ((-10 << 15) / 11.0)
#define Q1 ((-8 << 15) / 11.0)
#define Q2 ((-6 << 15) / 11.0)
#define Q3 ((-4 << 15) / 11.0)
#define Q4 ((-2 << 15) / 11.0)
#define Q5 (0)
#define Q6 ((2 << 15) / 11.0)
#define Q7 ((4 << 15) / 11.0)
#define Q8 ((6 << 15) / 11.0)
#define Q9 ((8 << 15) / 11.0)
#define QA ((10 << 15) / 11.0)

static const sample_t q_4_0[128] = {
    Q0, Q0, Q0, Q0, Q0, Q0, Q0, Q0, Q0, Q0, Q0,
    Q1, Q1, Q1, Q1, Q1, Q1, Q1, Q1, Q1, Q1, Q1,
    Q2, Q2, Q2, Q2, Q2, Q2, Q2, Q2, Q2, Q2, Q2,
    Q3, Q3, Q3, Q3, Q3, Q3, Q3, Q3, Q3, Q3, Q3,
    Q4, Q4, Q4, Q4, Q4, Q4, Q4, Q4, Q4, Q4, Q4,
    Q5, Q5, Q5, Q5, Q5, Q5, Q5, Q5, Q5, Q5, Q5,
    Q6, Q6, Q6, Q6, Q6, Q6, Q6, Q6, Q6, Q6, Q6,
    Q7, Q7, Q7, Q7, Q7, Q7, Q7, Q7, Q7, Q7, Q7,
    Q8, Q8, Q8, Q8, Q8, Q8, Q8, Q8, Q8, Q8, Q8,
    Q9, Q9, Q9, Q9, Q9, Q9, Q9, Q9, Q9, Q9, Q9,
    QA, QA, QA, QA, QA, QA, QA, QA, QA, QA, QA,
    0,  0,  0,  0,  0,  0,  0
};

static const sample_t q_4_1[128] = {
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, QA,
    0,  0,  0,  0,  0,  0,  0
};

#undef Q0
#undef Q1
#undef Q2
#undef Q3
#undef Q4
#undef Q5
#undef Q6
#undef Q7
#undef Q8
#undef Q9
#undef QA

static const sample_t q_5[16] = {
    (-14 << 15)/15.0,(-12 << 15)/15.0,(-10 << 15)/15.0,
    ( -8 << 15)/15.0,( -6 << 15)/15.0,( -4 << 15)/15.0,
    ( -2 << 15)/15.0,   0            ,(  2 << 15)/15.0,
    (  4 << 15)/15.0,(  6 << 15)/15.0,(  8 << 15)/15.0,
    ( 10 << 15)/15.0,( 12 << 15)/15.0,( 14 << 15)/15.0,
    0
};

static sample_t scale_factor[25] = {
    0.000030517578125,
    0.0000152587890625,
    0.00000762939453125,
    0.000003814697265625,
    0.0000019073486328125,
    0.00000095367431640625,
    0.000000476837158203125,
    0.0000002384185791015625,
    0.00000011920928955078125,
    0.000000059604644775390625,
    0.0000000298023223876953125,
    0.00000001490116119384765625,
    0.000000007450580596923828125,
    0.0000000037252902984619140625,
    0.00000000186264514923095703125,
    0.000000000931322574615478515625,
    0.0000000004656612873077392578125,
    0.00000000023283064365386962890625,
    0.000000000116415321826934814453125,
    0.0000000000582076609134674072265625,
    0.00000000002910383045673370361328125,
    0.000000000014551915228366851806640625,
    0.0000000000072759576141834259033203125,
    0.00000000000363797880709171295166015625,
    0.000000000001818989403545856475830078125
};

static const uint16_t dither_lut[256] = {
    0x0000, 0xa011, 0xe033, 0x4022, 0x6077, 0xc066, 0x8044, 0x2055,
    0xc0ee, 0x60ff, 0x20dd, 0x80cc, 0xa099, 0x0088, 0x40aa, 0xe0bb,
    0x21cd, 0x81dc, 0xc1fe, 0x61ef, 0x41ba, 0xe1ab, 0xa189, 0x0198,
    0xe123, 0x4132, 0x0110, 0xa101, 0x8154, 0x2145, 0x6167, 0xc176,
    0x439a, 0xe38b, 0xa3a9, 0x03b8, 0x23ed, 0x83fc, 0xc3de, 0x63cf,
    0x8374, 0x2365, 0x6347, 0xc356, 0xe303, 0x4312, 0x0330, 0xa321,
    0x6257, 0xc246, 0x8264, 0x2275, 0x0220, 0xa231, 0xe213, 0x4202,
    0xa2b9, 0x02a8, 0x428a, 0xe29b, 0xc2ce, 0x62df, 0x22fd, 0x82ec,
    0x8734, 0x2725, 0x6707, 0xc716, 0xe743, 0x4752, 0x0770, 0xa761,
    0x47da, 0xe7cb, 0xa7e9, 0x07f8, 0x27ad, 0x87bc, 0xc79e, 0x678f,
    0xa6f9, 0x06e8, 0x46ca, 0xe6db, 0xc68e, 0x669f, 0x26bd, 0x86ac,
    0x6617, 0xc606, 0x8624, 0x2635, 0x0660, 0xa671, 0xe653, 0x4642,
    0xc4ae, 0x64bf, 0x249d, 0x848c, 0xa4d9, 0x04c8, 0x44ea, 0xe4fb,
    0x0440, 0xa451, 0xe473, 0x4462, 0x6437, 0xc426, 0x8404, 0x2415,
    0xe563, 0x4572, 0x0550, 0xa541, 0x8514, 0x2505, 0x6527, 0xc536,
    0x258d, 0x859c, 0xc5be, 0x65af, 0x45fa, 0xe5eb, 0xa5c9, 0x05d8,
    0xae79, 0x0e68, 0x4e4a, 0xee5b, 0xce0e, 0x6e1f, 0x2e3d, 0x8e2c,
    0x6e97, 0xce86, 0x8ea4, 0x2eb5, 0x0ee0, 0xaef1, 0xeed3, 0x4ec2,
    0x8fb4, 0x2fa5, 0x6f87, 0xcf96, 0xefc3, 0x4fd2, 0x0ff0, 0xafe1,
    0x4f5a, 0xef4b, 0xaf69, 0x0f78, 0x2f2d, 0x8f3c, 0xcf1e, 0x6f0f,
    0xede3, 0x4df2, 0x0dd0, 0xadc1, 0x8d94, 0x2d85, 0x6da7, 0xcdb6,
    0x2d0d, 0x8d1c, 0xcd3e, 0x6d2f, 0x4d7a, 0xed6b, 0xad49, 0x0d58,
    0xcc2e, 0x6c3f, 0x2c1d, 0x8c0c, 0xac59, 0x0c48, 0x4c6a, 0xec7b,
    0x0cc0, 0xacd1, 0xecf3, 0x4ce2, 0x6cb7, 0xcca6, 0x8c84, 0x2c95,
    0x294d, 0x895c, 0xc97e, 0x696f, 0x493a, 0xe92b, 0xa909, 0x0918,
    0xe9a3, 0x49b2, 0x0990, 0xa981, 0x89d4, 0x29c5, 0x69e7, 0xc9f6,
    0x0880, 0xa891, 0xe8b3, 0x48a2, 0x68f7, 0xc8e6, 0x88c4, 0x28d5,
    0xc86e, 0x687f, 0x285d, 0x884c, 0xa819, 0x0808, 0x482a, 0xe83b,
    0x6ad7, 0xcac6, 0x8ae4, 0x2af5, 0x0aa0, 0xaab1, 0xea93, 0x4a82,
    0xaa39, 0x0a28, 0x4a0a, 0xea1b, 0xca4e, 0x6a5f, 0x2a7d, 0x8a6c,
    0x4b1a, 0xeb0b, 0xab29, 0x0b38, 0x2b6d, 0x8b7c, 0xcb5e, 0x6b4f,
    0x8bf4, 0x2be5, 0x6bc7, 0xcbd6, 0xeb83, 0x4b92, 0x0bb0, 0xaba1
};




More information about the MPlayer-cvslog mailing list