Mercurial > libavcodec.hg
changeset 2123:ef54decf5624 libavcodec
libdts support by (Benjamin Zores <ben at geexbox dot org>)
| author | michael |
|---|---|
| date | Wed, 14 Jul 2004 01:32:14 +0000 |
| parents | 503496800167 |
| children | 6f44ab5ad03c |
| files | Makefile allcodecs.c avcodec.c avcodec.h dts_internal.h dtsdec.c |
| diffstat | 6 files changed, 564 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- a/Makefile Sat Jul 10 23:22:47 2004 +0000 +++ b/Makefile Wed Jul 14 01:32:14 2004 +0000 @@ -73,6 +73,11 @@ endif endif +# currently using libdts for dts decoding +ifeq ($(CONFIG_DTS),yes) +OBJS+= dtsdec.o +endif + ifeq ($(CONFIG_FAAD),yes) OBJS+= faad.o ifeq ($(CONFIG_FAADBIN),yes)
--- a/allcodecs.c Sat Jul 10 23:22:47 2004 +0000 +++ b/allcodecs.c Wed Jul 14 01:32:14 2004 +0000 @@ -151,6 +151,9 @@ #ifdef CONFIG_AC3 register_avcodec(&ac3_decoder); #endif +#ifdef CONFIG_DTS + register_avcodec(&dts_decoder); +#endif register_avcodec(&ra_144_decoder); register_avcodec(&ra_288_decoder); register_avcodec(&roq_dpcm_decoder);
--- a/avcodec.c Sat Jul 10 23:22:47 2004 +0000 +++ b/avcodec.c Wed Jul 14 01:32:14 2004 +0000 @@ -38,6 +38,7 @@ { CODEC_ID_MJPEG, { MKTAG('M', 'J', 'P', 'G'), 0 } }, { CODEC_ID_MPEG1VIDEO, { MKTAG('P', 'I', 'M', '1'), 0 } }, { CODEC_ID_AC3, { 0x2000, 0 } }, + { CODEC_ID_DTS, { 0x10, 0 } }, { CODEC_ID_MP2, { 0x50, 0x55, 0 } }, { CODEC_ID_FLV1, { MKTAG('F', 'L', 'V', '1'), 0 } },
--- a/avcodec.h Sat Jul 10 23:22:47 2004 +0000 +++ b/avcodec.h Wed Jul 14 01:32:14 2004 +0000 @@ -140,6 +140,8 @@ CODEC_ID_MPEG2TS, /* _FAKE_ codec to indicate a raw MPEG2 transport stream (only used by libavformat) */ + + CODEC_ID_DTS, }; /* CODEC_ID_MP3LAME is absolete */ @@ -1858,6 +1860,7 @@ /* the following codecs use external GPL libs */ extern AVCodec ac3_decoder; +extern AVCodec dts_decoder; /* resample.c */
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/dts_internal.h Wed Jul 14 01:32:14 2004 +0000 @@ -0,0 +1,203 @@ +/* + * dts_internal.h + * Copyright (C) 2004 Gildas Bazin <gbazin@videolan.org> + * Copyright (C) 2000-2003 Michel Lespinasse <walken@zoy.org> + * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma@ess.engr.uvic.ca> + * + * This file is part of dtsdec, a free DTS Coherent Acoustics stream decoder. + * See http://www.videolan.org/dtsdec.html for updates. + * + * dtsdec 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. + * + * dtsdec 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 DTS_SUBFRAMES_MAX (16) +#define DTS_PRIM_CHANNELS_MAX (5) +#define DTS_SUBBANDS (32) +#define DTS_ABITS_MAX (32) /* Should be 28 */ +#define DTS_SUBSUBFAMES_MAX (4) +#define DTS_LFE_MAX (3) + +struct dts_state_s { + + /* Frame header */ + int frame_type; /* type of the current frame */ + int samples_deficit; /* deficit sample count */ + int crc_present; /* crc is present in the bitstream */ + int sample_blocks; /* number of PCM sample blocks */ + int frame_size; /* primary frame byte size */ + int amode; /* audio channels arrangement */ + int sample_rate; /* audio sampling rate */ + int bit_rate; /* transmission bit rate */ + + int downmix; /* embedded downmix enabled */ + int dynrange; /* embedded dynamic range flag */ + int timestamp; /* embedded time stamp flag */ + int aux_data; /* auxiliary data flag */ + int hdcd; /* source material is mastered in HDCD */ + int ext_descr; /* extension audio descriptor flag */ + int ext_coding; /* extended coding flag */ + int aspf; /* audio sync word insertion flag */ + int lfe; /* low frequency effects flag */ + int predictor_history; /* predictor history flag */ + int header_crc; /* header crc check bytes */ + int multirate_inter; /* multirate interpolator switch */ + int version; /* encoder software revision */ + int copy_history; /* copy history */ + int source_pcm_res; /* source pcm resolution */ + int front_sum; /* front sum/difference flag */ + int surround_sum; /* surround sum/difference flag */ + int dialog_norm; /* dialog normalisation parameter */ + + /* Primary audio coding header */ + int subframes; /* number of subframes */ + int prim_channels; /* number of primary audio channels */ + /* subband activity count */ + int subband_activity[DTS_PRIM_CHANNELS_MAX]; + /* high frequency vq start subband */ + int vq_start_subband[DTS_PRIM_CHANNELS_MAX]; + /* joint intensity coding index */ + int joint_intensity[DTS_PRIM_CHANNELS_MAX]; + /* transient mode code book */ + int transient_huffman[DTS_PRIM_CHANNELS_MAX]; + /* scale factor code book */ + int scalefactor_huffman[DTS_PRIM_CHANNELS_MAX]; + /* bit allocation quantizer select */ + int bitalloc_huffman[DTS_PRIM_CHANNELS_MAX]; + /* quantization index codebook select */ + int quant_index_huffman[DTS_PRIM_CHANNELS_MAX][DTS_ABITS_MAX]; + /* scale factor adjustment */ + float scalefactor_adj[DTS_PRIM_CHANNELS_MAX][DTS_ABITS_MAX]; + + /* Primary audio coding side information */ + int subsubframes; /* number of subsubframes */ + int partial_samples; /* partial subsubframe samples count */ + /* prediction mode (ADPCM used or not) */ + int prediction_mode[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS]; + /* prediction VQ coefs */ + int prediction_vq[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS]; + /* bit allocation index */ + int bitalloc[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS]; + /* transition mode (transients) */ + int transition_mode[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS]; + /* scale factors (2 if transient)*/ + int scale_factor[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS][2]; + /* joint subband scale factors codebook */ + int joint_huff[DTS_PRIM_CHANNELS_MAX]; + /* joint subband scale factors */ + int joint_scale_factor[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS]; + /* stereo downmix coefficients */ + int downmix_coef[DTS_PRIM_CHANNELS_MAX][2]; + /* dynamic range coefficient */ + int dynrange_coef; + + /* VQ encoded high frequency subbands */ + int high_freq_vq[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS]; + + /* Low frequency effect data */ + double lfe_data[2*DTS_SUBSUBFAMES_MAX*DTS_LFE_MAX * 2 /*history*/]; + int lfe_scale_factor; + + /* Subband samples history (for ADPCM) */ + double subband_samples_hist[DTS_PRIM_CHANNELS_MAX][DTS_SUBBANDS][4]; + double subband_fir_hist[DTS_PRIM_CHANNELS_MAX][512]; + double subband_fir_noidea[DTS_PRIM_CHANNELS_MAX][64]; + + /* Audio output */ + level_t clev; /* centre channel mix level */ + level_t slev; /* surround channels mix level */ + + int output; /* type of output */ + level_t level; /* output level */ + sample_t bias; /* output bias */ + + sample_t * samples; /* pointer to the internal audio samples buffer */ + int downmixed; + + int dynrnge; /* apply dynamic range */ + level_t dynrng; /* dynamic range */ + void * dynrngdata; /* dynamic range callback funtion and data */ + level_t (* dynrngcall) (level_t range, void * dynrngdata); + + /* Bitstream handling */ + uint32_t * buffer_start; + uint32_t bits_left; + uint32_t current_word; + int word_mode; /* 16/14 bits word format (1 -> 16, 0 -> 14) */ + int bigendian_mode; /* endianness (1 -> be, 0 -> le) */ + + /* Current position in DTS frame */ + int current_subframe; + int current_subsubframe; + + /* Pre-calculated cosine modulation coefs for the QMF */ + double cos_mod[544]; + + /* Debug flag */ + int debug_flag; +}; + +#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 + +int dts_downmix_init (int input, int flags, level_t * level, + level_t clev, level_t slev); +int dts_downmix_coeff (level_t * coeff, int acmod, int output, level_t level, + level_t clev, level_t slev); +void dts_downmix (sample_t * samples, int acmod, int output, sample_t bias, + level_t clev, level_t slev); +void dts_upmix (sample_t * samples, int acmod, int output); + +#define ROUND(x) ((int)((x) + ((x) > 0 ? 0.5 : -0.5))) + +#ifndef LIBDTS_FIXED + +typedef sample_t quantizer_t; +#define SAMPLE(x) (x) +#define LEVEL(x) (x) +#define MUL(a,b) ((a) * (b)) +#define MUL_L(a,b) ((a) * (b)) +#define MUL_C(a,b) ((a) * (b)) +#define DIV(a,b) ((a) / (b)) +#define BIAS(x) ((x) + bias) + +#else /* LIBDTS_FIXED */ + +typedef int16_t quantizer_t; +#define SAMPLE(x) (sample_t)((x) * (1 << 30)) +#define LEVEL(x) (level_t)((x) * (1 << 26)) + +#if 0 +#define MUL(a,b) ((int)(((int64_t)(a) * (b) + (1 << 29)) >> 30)) +#define MUL_L(a,b) ((int)(((int64_t)(a) * (b) + (1 << 25)) >> 26)) +#elif 1 +#define MUL(a,b) \ +({ int32_t _ta=(a), _tb=(b), _tc; \ + _tc=(_ta & 0xffff)*(_tb >> 16)+(_ta >> 16)*(_tb & 0xffff); (int32_t)(((_tc >> 14))+ (((_ta >> 16)*(_tb >> 16)) << 2 )); }) +#define MUL_L(a,b) \ +({ int32_t _ta=(a), _tb=(b), _tc; \ + _tc=(_ta & 0xffff)*(_tb >> 16)+(_ta >> 16)*(_tb & 0xffff); (int32_t)((_tc >> 10) + (((_ta >> 16)*(_tb >> 16)) << 6)); }) +#else +#define MUL(a,b) (((a) >> 15) * ((b) >> 15)) +#define MUL_L(a,b) (((a) >> 13) * ((b) >> 13)) +#endif + +#define MUL_C(a,b) MUL_L (a, LEVEL (b)) +#define DIV(a,b) ((((int64_t)LEVEL (a)) << 26) / (b)) +#define BIAS(x) (x) + +#endif
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/dtsdec.c Wed Jul 14 01:32:14 2004 +0000 @@ -0,0 +1,349 @@ +/* + * dtsdec.c : free DTS Coherent Acoustics stream decoder. + * Copyright (C) 2004 Benjamin Zores <ben@geexbox.org> + * + * This file is part of libavcodec. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2 of the License, or (at your option) any later version. + * + * This library 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 + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#ifdef HAVE_AV_CONFIG_H +#undef HAVE_AV_CONFIG_H +#endif + +#include "avcodec.h" +#include <dts.h> +#include "dts_internal.h" + +#include <stdlib.h> +#include <string.h> +#include <malloc.h> +#include <math.h> + +#define INBUF_SIZE 4096 +#define BUFFER_SIZE 4096 +#define HEADER_SIZE 14 + +#ifdef LIBDTS_FIXED +#define CONVERT_LEVEL (1 << 26) +#define CONVERT_BIAS 0 +#else +#define CONVERT_LEVEL 1 +#define CONVERT_BIAS 384 +#endif + +static void +pre_calc_cosmod (dts_state_t * state) +{ + int i, j, k; + + for (j=0,k=0;k<16;k++) + for (i=0;i<16;i++) + state->cos_mod[j++] = cos((2*i+1)*(2*k+1)*M_PI/64); + + for (k=0;k<16;k++) + for (i=0;i<16;i++) + state->cos_mod[j++] = cos((i)*(2*k+1)*M_PI/32); + + for (k=0;k<16;k++) + state->cos_mod[j++] = 0.25/(2*cos((2*k+1)*M_PI/128)); + + for (k=0;k<16;k++) + state->cos_mod[j++] = -0.25/(2.0*sin((2*k+1)*M_PI/128)); +} + +static inline +int16_t convert (int32_t i) +{ +#ifdef LIBDTS_FIXED + i >>= 15; +#else + i -= 0x43c00000; +#endif + return (i > 32767) ? 32767 : ((i < -32768) ? -32768 : i); +} + +void +convert2s16_2 (sample_t * _f, int16_t * s16) +{ + int i; + int32_t * f = (int32_t *) _f; + + for (i = 0; i < 256; i++) + { + s16[2*i] = convert (f[i]); + s16[2*i+1] = convert (f[i+256]); + } +} + +void +convert2s16_4 (sample_t * _f, int16_t * s16) +{ + int i; + int32_t * f = (int32_t *) _f; + + for (i = 0; i < 256; i++) + { + s16[4*i] = convert (f[i]); + s16[4*i+1] = convert (f[i+256]); + s16[4*i+2] = convert (f[i+512]); + s16[4*i+3] = convert (f[i+768]); + } +} + +void +convert2s16_5 (sample_t * _f, int16_t * s16) +{ + int i; + int32_t * f = (int32_t *) _f; + + for (i = 0; i < 256; i++) + { + s16[5*i] = convert (f[i]); + s16[5*i+1] = convert (f[i+256]); + s16[5*i+2] = convert (f[i+512]); + s16[5*i+3] = convert (f[i+768]); + s16[5*i+4] = convert (f[i+1024]); + } +} + +static void +convert2s16_multi (sample_t * _f, int16_t * s16, int flags) +{ + int i; + int32_t * f = (int32_t *) _f; + + switch (flags) + { + case DTS_MONO: + for (i = 0; i < 256; i++) + { + s16[5*i] = s16[5*i+1] = s16[5*i+2] = s16[5*i+3] = 0; + s16[5*i+4] = convert (f[i]); + } + break; + case DTS_CHANNEL: + case DTS_STEREO: + case DTS_DOLBY: + convert2s16_2 (_f, s16); + break; + case DTS_3F: + for (i = 0; i < 256; i++) + { + s16[5*i] = convert (f[i]); + s16[5*i+1] = convert (f[i+512]); + s16[5*i+2] = s16[5*i+3] = 0; + s16[5*i+4] = convert (f[i+256]); + } + break; + case DTS_2F2R: + convert2s16_4 (_f, s16); + break; + case DTS_3F2R: + convert2s16_5 (_f, s16); + break; + case DTS_MONO | DTS_LFE: + for (i = 0; i < 256; i++) + { + s16[6*i] = s16[6*i+1] = s16[6*i+2] = s16[6*i+3] = 0; + s16[6*i+4] = convert (f[i+256]); + s16[6*i+5] = convert (f[i]); + } + break; + case DTS_CHANNEL | DTS_LFE: + case DTS_STEREO | DTS_LFE: + case DTS_DOLBY | DTS_LFE: + for (i = 0; i < 256; i++) + { + s16[6*i] = convert (f[i+256]); + s16[6*i+1] = convert (f[i+512]); + s16[6*i+2] = s16[6*i+3] = s16[6*i+4] = 0; + s16[6*i+5] = convert (f[i]); + } + break; + case DTS_3F | DTS_LFE: + for (i = 0; i < 256; i++) + { + s16[6*i] = convert (f[i+256]); + s16[6*i+1] = convert (f[i+768]); + s16[6*i+2] = s16[6*i+3] = 0; + s16[6*i+4] = convert (f[i+512]); + s16[6*i+5] = convert (f[i]); + } + break; + case DTS_2F2R | DTS_LFE: + for (i = 0; i < 256; i++) + { + s16[6*i] = convert (f[i+256]); + s16[6*i+1] = convert (f[i+512]); + s16[6*i+2] = convert (f[i+768]); + s16[6*i+3] = convert (f[i+1024]); + s16[6*i+4] = 0; + s16[6*i+5] = convert (f[i]); + } + break; + case DTS_3F2R | DTS_LFE: + for (i = 0; i < 256; i++) + { + s16[6*i] = convert (f[i+256]); + s16[6*i+1] = convert (f[i+768]); + s16[6*i+2] = convert (f[i+1024]); + s16[6*i+3] = convert (f[i+1280]); + s16[6*i+4] = convert (f[i+512]); + s16[6*i+5] = convert (f[i]); + } + break; + } +} + +static int +channels_multi (int flags) +{ + if (flags & DTS_LFE) + return 6; + else if (flags & 1) /* center channel */ + return 5; + else if ((flags & DTS_CHANNEL_MASK) == DTS_2F2R) + return 4; + else + return 2; +} + +static int +dts_decode_frame (AVCodecContext *avctx, void *data, int *data_size, + uint8_t *buff, int buff_size) +{ + uint8_t * start = buff; + uint8_t * end = buff + buff_size; + *data_size = 0; + + static uint8_t buf[BUFFER_SIZE]; + static uint8_t * bufptr = buf; + static uint8_t * bufpos = buf + HEADER_SIZE; + + static int sample_rate; + static int frame_length; + static int flags; + int bit_rate; + int len; + dts_state_t *state = avctx->priv_data; + + while (1) + { + len = end - start; + if (!len) + break; + if (len > bufpos - bufptr) + len = bufpos - bufptr; + memcpy (bufptr, start, len); + bufptr += len; + start += len; + if (bufptr == bufpos) + { + if (bufpos == buf + HEADER_SIZE) + { + int length; + + length = dts_syncinfo (state, buf, &flags, &sample_rate, + &bit_rate, &frame_length); + if (!length) + { + av_log (NULL, AV_LOG_INFO, "skip\n"); + for (bufptr = buf; bufptr < buf + HEADER_SIZE-1; bufptr++) + bufptr[0] = bufptr[1]; + continue; + } + bufpos = buf + length; + } + else + { + level_t level; + sample_t bias; + int i; + + flags = 2; /* ???????????? */ + level = CONVERT_LEVEL; + bias = CONVERT_BIAS; + + flags |= DTS_ADJUST_LEVEL; + if (dts_frame (state, buf, &flags, &level, bias)) + goto error; + for (i = 0; i < dts_blocks_num (state); i++) + { + if (dts_block (state)) + goto error; + { + int chans; + chans = channels_multi (flags); + convert2s16_multi (dts_samples (state), data, + flags & (DTS_CHANNEL_MASK | DTS_LFE)); + + data += 256 * sizeof (int16_t) * chans; + *data_size += 256 * sizeof (int16_t) * chans; + } + } + bufptr = buf; + bufpos = buf + HEADER_SIZE; + continue; + error: + av_log (NULL, AV_LOG_ERROR, "error\n"); + bufptr = buf; + bufpos = buf + HEADER_SIZE; + } + } + } + + return buff_size; +} + +static int +dts_decode_init (AVCodecContext *avctx) +{ + dts_state_t * state; + int i; + + state = avctx->priv_data; + memset (state, 0, sizeof (dts_state_t)); + + state->samples = (sample_t *) memalign (16, 256 * 12 * sizeof (sample_t)); + if (state->samples == NULL) + return 1; + + for (i = 0; i < 256 * 12; i++) + state->samples[i] = 0; + + /* Pre-calculate cosine modulation coefficients */ + pre_calc_cosmod (state); + state->downmixed = 1; + + return 0; +} + +static int +dts_decode_end (AVCodecContext *s) +{ + return 0; +} + +AVCodec dts_decoder = { + "dts", + CODEC_TYPE_AUDIO, + CODEC_ID_DTS, + sizeof (dts_state_t), + dts_decode_init, + NULL, + dts_decode_end, + dts_decode_frame, +};