diff --git a/libavcodec/Makefile b/libavcodec/Makefile index 7d3b423462..2aa5120b81 100644 --- a/libavcodec/Makefile +++ b/libavcodec/Makefile @@ -189,7 +189,7 @@ OBJS-$(CONFIG_CPIA_DECODER) += cpia.o OBJS-$(CONFIG_CSCD_DECODER) += cscd.o OBJS-$(CONFIG_CYUV_DECODER) += cyuv.o OBJS-$(CONFIG_DCA_DECODER) += dcadec.o dca.o dcadsp.o \ - synth_filter.o + dca_exss.o synth_filter.o OBJS-$(CONFIG_DCA_ENCODER) += dcaenc.o dca.o OBJS-$(CONFIG_DIRAC_DECODER) += diracdec.o dirac.o diracdsp.o \ dirac_arith.o mpeg12data.o dirac_dwt.o diff --git a/libavcodec/dca.h b/libavcodec/dca.h index 70dade5cd0..b752bc3fdb 100644 --- a/libavcodec/dca.h +++ b/libavcodec/dca.h @@ -26,8 +26,15 @@ #define AVCODEC_DCA_H #include + +#include "libavutil/float_dsp.h" #include "libavutil/internal.h" +#include "avcodec.h" +#include "dcadsp.h" +#include "fmtconvert.h" +#include "get_bits.h" + /** DCA syncwords, also used for bitstream type detection */ #define DCA_MARKER_RAW_BE 0x7FFE8001 #define DCA_MARKER_RAW_LE 0xFE7F0180 @@ -37,6 +44,164 @@ /** DCA-HD specific block starts with this marker. */ #define DCA_HD_MARKER 0x64582025 +#define DCA_PRIM_CHANNELS_MAX (7) +#define DCA_ABITS_MAX (32) /* Should be 28 */ +#define DCA_SUBSUBFRAMES_MAX (4) +#define DCA_SUBFRAMES_MAX (16) +#define DCA_BLOCKS_MAX (16) +#define DCA_LFE_MAX (3) +#define DCA_CHSETS_MAX (4) +#define DCA_CHSET_CHANS_MAX (8) + +#define DCA_MAX_FRAME_SIZE 16384 +#define DCA_MAX_EXSS_HEADER_SIZE 4096 + +#define DCA_BUFFER_PADDING_SIZE 1024 + +enum DCAExtensionMask { + DCA_EXT_CORE = 0x001, ///< core in core substream + DCA_EXT_XXCH = 0x002, ///< XXCh channels extension in core substream + DCA_EXT_X96 = 0x004, ///< 96/24 extension in core substream + DCA_EXT_XCH = 0x008, ///< XCh channel extension in core substream + DCA_EXT_EXSS_CORE = 0x010, ///< core in ExSS (extension substream) + DCA_EXT_EXSS_XBR = 0x020, ///< extended bitrate extension in ExSS + DCA_EXT_EXSS_XXCH = 0x040, ///< XXCh channels extension in ExSS + DCA_EXT_EXSS_X96 = 0x080, ///< 96/24 extension in ExSS + DCA_EXT_EXSS_LBR = 0x100, ///< low bitrate component in ExSS + DCA_EXT_EXSS_XLL = 0x200, ///< lossless extension in ExSS +}; + +typedef struct DCAContext { + const AVClass *class; ///< class for AVOptions + AVCodecContext *avctx; + /* 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 bit_rate_index; ///< transmission bit rate index + + 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 total_channels; ///< number of channels including extensions + int prim_channels; ///< number of primary audio channels + int subband_activity[DCA_PRIM_CHANNELS_MAX]; ///< subband activity count + int vq_start_subband[DCA_PRIM_CHANNELS_MAX]; ///< high frequency vq start subband + int joint_intensity[DCA_PRIM_CHANNELS_MAX]; ///< joint intensity coding index + int transient_huffman[DCA_PRIM_CHANNELS_MAX]; ///< transient mode code book + int scalefactor_huffman[DCA_PRIM_CHANNELS_MAX]; ///< scale factor code book + int bitalloc_huffman[DCA_PRIM_CHANNELS_MAX]; ///< bit allocation quantizer select + int quant_index_huffman[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< quantization index codebook select + float scalefactor_adj[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< scale factor adjustment + + /* Primary audio coding side information */ + int subsubframes[DCA_SUBFRAMES_MAX]; ///< number of subsubframes + int partial_samples[DCA_SUBFRAMES_MAX]; ///< partial subsubframe samples count + int prediction_mode[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< prediction mode (ADPCM used or not) + int prediction_vq[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< prediction VQ coefs + int bitalloc[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< bit allocation index + int transition_mode[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< transition mode (transients) + int32_t scale_factor[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][2];///< scale factors (2 if transient) + int joint_huff[DCA_PRIM_CHANNELS_MAX]; ///< joint subband scale factors codebook + int joint_scale_factor[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< joint subband scale factors + float downmix_coef[DCA_PRIM_CHANNELS_MAX + 1][2]; ///< stereo downmix coefficients + int dynrange_coef; ///< dynamic range coefficient + + /* Core substream's embedded downmix coefficients (cf. ETSI TS 102 114 V1.4.1) + * Input: primary audio channels (incl. LFE if present) + * Output: downmix audio channels (up to 4, no LFE) */ + uint8_t core_downmix; ///< embedded downmix coefficients available + uint8_t core_downmix_amode; ///< audio channel arrangement of embedded downmix + uint16_t core_downmix_codes[DCA_PRIM_CHANNELS_MAX + 1][4]; ///< embedded downmix coefficients (9-bit codes) + + int32_t high_freq_vq[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< VQ encoded high frequency subbands + + float lfe_data[2 * DCA_LFE_MAX * (DCA_BLOCKS_MAX + 4)]; ///< Low frequency effect data + int lfe_scale_factor; + + /* Subband samples history (for ADPCM) */ + DECLARE_ALIGNED(16, float, subband_samples_hist)[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][4]; + DECLARE_ALIGNED(32, float, subband_fir_hist)[DCA_PRIM_CHANNELS_MAX][512]; + DECLARE_ALIGNED(32, float, subband_fir_noidea)[DCA_PRIM_CHANNELS_MAX][32]; + int hist_index[DCA_PRIM_CHANNELS_MAX]; + DECLARE_ALIGNED(32, float, raXin)[32]; + + int output; ///< type of output + + DECLARE_ALIGNED(32, float, subband_samples)[DCA_BLOCKS_MAX][DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][8]; + float *samples_chanptr[DCA_PRIM_CHANNELS_MAX + 1]; + float *extra_channels[DCA_PRIM_CHANNELS_MAX + 1]; + uint8_t *extra_channels_buffer; + unsigned int extra_channels_buffer_size; + + uint8_t dca_buffer[DCA_MAX_FRAME_SIZE + DCA_MAX_EXSS_HEADER_SIZE + DCA_BUFFER_PADDING_SIZE]; + int dca_buffer_size; ///< how much data is in the dca_buffer + + const int8_t *channel_order_tab; ///< channel reordering table, lfe and non lfe + GetBitContext gb; + /* Current position in DCA frame */ + int current_subframe; + int current_subsubframe; + + int core_ext_mask; ///< present extensions in the core substream + + /* XCh extension information */ + int xch_present; ///< XCh extension present and valid + int xch_base_channel; ///< index of first (only) channel containing XCH data + int xch_disable; ///< whether the XCh extension should be decoded or not + + /* XXCH extension information */ + int xxch_chset; + int xxch_nbits_spk_mask; + uint32_t xxch_core_spkmask; + uint32_t xxch_spk_masks[4]; /* speaker masks, last element is core mask */ + int xxch_chset_nch[4]; + float xxch_dmix_sf[DCA_CHSETS_MAX]; + + uint32_t xxch_dmix_embedded; /* lower layer has mix pre-embedded, per chset */ + float xxch_dmix_coeff[DCA_PRIM_CHANNELS_MAX][32]; /* worst case sizing */ + + int8_t xxch_order_tab[32]; + int8_t lfe_index; + + /* ExSS header parser */ + int static_fields; ///< static fields present + int mix_metadata; ///< mixing metadata present + int num_mix_configs; ///< number of mix out configurations + int mix_config_num_ch[4]; ///< number of channels in each mix out configuration + + int profile; + + int debug_flag; ///< used for suppressing repeated error messages output + AVFloatDSPContext *fdsp; + FFTContext imdct; + SynthFilterContext synth; + DCADSPContext dcadsp; + FmtConvertContext fmt_conv; +} DCAContext; + extern av_export const uint32_t avpriv_dca_sample_rates[16]; /** @@ -45,4 +210,7 @@ extern av_export const uint32_t avpriv_dca_sample_rates[16]; int avpriv_dca_convert_bitstream(const uint8_t *src, int src_size, uint8_t *dst, int max_size); +int ff_dca_xbr_parse_frame(DCAContext *s); +int ff_dca_xxch_decode_frame(DCAContext *s); + #endif /* AVCODEC_DCA_H */ diff --git a/libavcodec/dca_exss.c b/libavcodec/dca_exss.c new file mode 100644 index 0000000000..c90e9950af --- /dev/null +++ b/libavcodec/dca_exss.c @@ -0,0 +1,355 @@ +/* + * DCA ExSS extension + * + * This file is part of FFmpeg. + * + * FFmpeg 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 2.1 of the License, or (at your option) any later version. + * + * FFmpeg 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 FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "libavutil/common.h" +#include "libavutil/log.h" + +#include "dca.h" +#include "dca_exss.h" +#include "get_bits.h" + +/* extensions that reside in core substream */ +#define DCA_CORE_EXTS (DCA_EXT_XCH | DCA_EXT_XXCH | DCA_EXT_X96) + +/* these are unconfirmed but should be mostly correct */ +enum DCAExSSSpeakerMask { + DCA_EXSS_FRONT_CENTER = 0x0001, + DCA_EXSS_FRONT_LEFT_RIGHT = 0x0002, + DCA_EXSS_SIDE_REAR_LEFT_RIGHT = 0x0004, + DCA_EXSS_LFE = 0x0008, + DCA_EXSS_REAR_CENTER = 0x0010, + DCA_EXSS_FRONT_HIGH_LEFT_RIGHT = 0x0020, + DCA_EXSS_REAR_LEFT_RIGHT = 0x0040, + DCA_EXSS_FRONT_HIGH_CENTER = 0x0080, + DCA_EXSS_OVERHEAD = 0x0100, + DCA_EXSS_CENTER_LEFT_RIGHT = 0x0200, + DCA_EXSS_WIDE_LEFT_RIGHT = 0x0400, + DCA_EXSS_SIDE_LEFT_RIGHT = 0x0800, + DCA_EXSS_LFE2 = 0x1000, + DCA_EXSS_SIDE_HIGH_LEFT_RIGHT = 0x2000, + DCA_EXSS_REAR_HIGH_CENTER = 0x4000, + DCA_EXSS_REAR_HIGH_LEFT_RIGHT = 0x8000, +}; + +/** + * Return the number of channels in an ExSS speaker mask (HD) + */ +static int dca_exss_mask2count(int mask) +{ + /* count bits that mean speaker pairs twice */ + return av_popcount(mask) + + av_popcount(mask & (DCA_EXSS_CENTER_LEFT_RIGHT | + DCA_EXSS_FRONT_LEFT_RIGHT | + DCA_EXSS_FRONT_HIGH_LEFT_RIGHT | + DCA_EXSS_WIDE_LEFT_RIGHT | + DCA_EXSS_SIDE_LEFT_RIGHT | + DCA_EXSS_SIDE_HIGH_LEFT_RIGHT | + DCA_EXSS_SIDE_REAR_LEFT_RIGHT | + DCA_EXSS_REAR_LEFT_RIGHT | + DCA_EXSS_REAR_HIGH_LEFT_RIGHT)); +} + +/** + * Skip mixing coefficients of a single mix out configuration (HD) + */ +static void dca_exss_skip_mix_coeffs(GetBitContext *gb, int channels, int out_ch) +{ + int i; + + for (i = 0; i < channels; i++) { + int mix_map_mask = get_bits(gb, out_ch); + int num_coeffs = av_popcount(mix_map_mask); + skip_bits_long(gb, num_coeffs * 6); + } +} + +/** + * Parse extension substream asset header (HD) + */ +static int dca_exss_parse_asset_header(DCAContext *s) +{ + int header_pos = get_bits_count(&s->gb); + int header_size; + int channels = 0; + int embedded_stereo = 0; + int embedded_6ch = 0; + int drc_code_present; + int extensions_mask = 0; + int i, j; + + if (get_bits_left(&s->gb) < 16) + return -1; + + /* We will parse just enough to get to the extensions bitmask with which + * we can set the profile value. */ + + header_size = get_bits(&s->gb, 9) + 1; + skip_bits(&s->gb, 3); // asset index + + if (s->static_fields) { + if (get_bits1(&s->gb)) + skip_bits(&s->gb, 4); // asset type descriptor + if (get_bits1(&s->gb)) + skip_bits_long(&s->gb, 24); // language descriptor + + if (get_bits1(&s->gb)) { + /* How can one fit 1024 bytes of text here if the maximum value + * for the asset header size field above was 512 bytes? */ + int text_length = get_bits(&s->gb, 10) + 1; + if (get_bits_left(&s->gb) < text_length * 8) + return -1; + skip_bits_long(&s->gb, text_length * 8); // info text + } + + skip_bits(&s->gb, 5); // bit resolution - 1 + skip_bits(&s->gb, 4); // max sample rate code + channels = get_bits(&s->gb, 8) + 1; + + if (get_bits1(&s->gb)) { // 1-to-1 channels to speakers + int spkr_remap_sets; + int spkr_mask_size = 16; + int num_spkrs[7]; + + if (channels > 2) + embedded_stereo = get_bits1(&s->gb); + if (channels > 6) + embedded_6ch = get_bits1(&s->gb); + + if (get_bits1(&s->gb)) { + spkr_mask_size = (get_bits(&s->gb, 2) + 1) << 2; + skip_bits(&s->gb, spkr_mask_size); // spkr activity mask + } + + spkr_remap_sets = get_bits(&s->gb, 3); + + for (i = 0; i < spkr_remap_sets; i++) { + /* std layout mask for each remap set */ + num_spkrs[i] = dca_exss_mask2count(get_bits(&s->gb, spkr_mask_size)); + } + + for (i = 0; i < spkr_remap_sets; i++) { + int num_dec_ch_remaps = get_bits(&s->gb, 5) + 1; + if (get_bits_left(&s->gb) < 0) + return -1; + + for (j = 0; j < num_spkrs[i]; j++) { + int remap_dec_ch_mask = get_bits_long(&s->gb, num_dec_ch_remaps); + int num_dec_ch = av_popcount(remap_dec_ch_mask); + skip_bits_long(&s->gb, num_dec_ch * 5); // remap codes + } + } + } else { + skip_bits(&s->gb, 3); // representation type + } + } + + drc_code_present = get_bits1(&s->gb); + if (drc_code_present) + get_bits(&s->gb, 8); // drc code + + if (get_bits1(&s->gb)) + skip_bits(&s->gb, 5); // dialog normalization code + + if (drc_code_present && embedded_stereo) + get_bits(&s->gb, 8); // drc stereo code + + if (s->mix_metadata && get_bits1(&s->gb)) { + skip_bits(&s->gb, 1); // external mix + skip_bits(&s->gb, 6); // post mix gain code + + if (get_bits(&s->gb, 2) != 3) // mixer drc code + skip_bits(&s->gb, 3); // drc limit + else + skip_bits(&s->gb, 8); // custom drc code + + if (get_bits1(&s->gb)) // channel specific scaling + for (i = 0; i < s->num_mix_configs; i++) + skip_bits_long(&s->gb, s->mix_config_num_ch[i] * 6); // scale codes + else + skip_bits_long(&s->gb, s->num_mix_configs * 6); // scale codes + + for (i = 0; i < s->num_mix_configs; i++) { + if (get_bits_left(&s->gb) < 0) + return -1; + dca_exss_skip_mix_coeffs(&s->gb, channels, s->mix_config_num_ch[i]); + if (embedded_6ch) + dca_exss_skip_mix_coeffs(&s->gb, 6, s->mix_config_num_ch[i]); + if (embedded_stereo) + dca_exss_skip_mix_coeffs(&s->gb, 2, s->mix_config_num_ch[i]); + } + } + + switch (get_bits(&s->gb, 2)) { + case 0: + extensions_mask = get_bits(&s->gb, 12); + break; + case 1: + extensions_mask = DCA_EXT_EXSS_XLL; + break; + case 2: + extensions_mask = DCA_EXT_EXSS_LBR; + break; + case 3: + extensions_mask = 0; /* aux coding */ + break; + } + + /* not parsed further, we were only interested in the extensions mask */ + + if (get_bits_left(&s->gb) < 0) + return -1; + + if (get_bits_count(&s->gb) - header_pos > header_size * 8) { + av_log(s->avctx, AV_LOG_WARNING, "Asset header size mismatch.\n"); + return -1; + } + skip_bits_long(&s->gb, header_pos + header_size * 8 - get_bits_count(&s->gb)); + + if (extensions_mask & DCA_EXT_EXSS_XLL) + s->profile = FF_PROFILE_DTS_HD_MA; + else if (extensions_mask & (DCA_EXT_EXSS_XBR | DCA_EXT_EXSS_X96 | + DCA_EXT_EXSS_XXCH)) + s->profile = FF_PROFILE_DTS_HD_HRA; + + if (!(extensions_mask & DCA_EXT_CORE)) + av_log(s->avctx, AV_LOG_WARNING, "DTS core detection mismatch.\n"); + if ((extensions_mask & DCA_CORE_EXTS) != s->core_ext_mask) + av_log(s->avctx, AV_LOG_WARNING, + "DTS extensions detection mismatch (%d, %d)\n", + extensions_mask & DCA_CORE_EXTS, s->core_ext_mask); + + return 0; +} + +/** + * Parse extension substream header (HD) + */ +void ff_dca_exss_parse_header(DCAContext *s) +{ + int asset_size[8]; + int ss_index; + int blownup; + int num_audiop = 1; + int num_assets = 1; + int active_ss_mask[8]; + int i, j; + int start_posn; + int hdrsize; + uint32_t mkr; + + if (get_bits_left(&s->gb) < 52) + return; + + start_posn = get_bits_count(&s->gb) - 32; + + skip_bits(&s->gb, 8); // user data + ss_index = get_bits(&s->gb, 2); + + blownup = get_bits1(&s->gb); + hdrsize = get_bits(&s->gb, 8 + 4 * blownup) + 1; // header_size + skip_bits(&s->gb, 16 + 4 * blownup); // hd_size + + s->static_fields = get_bits1(&s->gb); + if (s->static_fields) { + skip_bits(&s->gb, 2); // reference clock code + skip_bits(&s->gb, 3); // frame duration code + + if (get_bits1(&s->gb)) + skip_bits_long(&s->gb, 36); // timestamp + + /* a single stream can contain multiple audio assets that can be + * combined to form multiple audio presentations */ + + num_audiop = get_bits(&s->gb, 3) + 1; + if (num_audiop > 1) { + avpriv_request_sample(s->avctx, + "Multiple DTS-HD audio presentations"); + /* ignore such streams for now */ + return; + } + + num_assets = get_bits(&s->gb, 3) + 1; + if (num_assets > 1) { + avpriv_request_sample(s->avctx, "Multiple DTS-HD audio assets"); + /* ignore such streams for now */ + return; + } + + for (i = 0; i < num_audiop; i++) + active_ss_mask[i] = get_bits(&s->gb, ss_index + 1); + + for (i = 0; i < num_audiop; i++) + for (j = 0; j <= ss_index; j++) + if (active_ss_mask[i] & (1 << j)) + skip_bits(&s->gb, 8); // active asset mask + + s->mix_metadata = get_bits1(&s->gb); + if (s->mix_metadata) { + int mix_out_mask_size; + + skip_bits(&s->gb, 2); // adjustment level + mix_out_mask_size = (get_bits(&s->gb, 2) + 1) << 2; + s->num_mix_configs = get_bits(&s->gb, 2) + 1; + + for (i = 0; i < s->num_mix_configs; i++) { + int mix_out_mask = get_bits(&s->gb, mix_out_mask_size); + s->mix_config_num_ch[i] = dca_exss_mask2count(mix_out_mask); + } + } + } + + av_assert0(num_assets > 0); // silence a warning + + for (i = 0; i < num_assets; i++) + asset_size[i] = get_bits_long(&s->gb, 16 + 4 * blownup); + + for (i = 0; i < num_assets; i++) { + if (dca_exss_parse_asset_header(s)) + return; + } + + /* not parsed further, we were only interested in the extensions mask + * from the asset header */ + + j = get_bits_count(&s->gb); + if (start_posn + hdrsize * 8 > j) + skip_bits_long(&s->gb, start_posn + hdrsize * 8 - j); + + for (i = 0; i < num_assets; i++) { + start_posn = get_bits_count(&s->gb); + mkr = get_bits_long(&s->gb, 32); + + /* parse extensions that we know about */ + if (mkr == 0x655e315e) { + ff_dca_xbr_parse_frame(s); + } else if (mkr == 0x47004a03) { + ff_dca_xxch_decode_frame(s); + s->core_ext_mask |= DCA_EXT_XXCH; /* xxx use for chan reordering */ + } else { + av_log(s->avctx, AV_LOG_DEBUG, + "DTS-ExSS: unknown marker = 0x%08x\n", mkr); + } + + /* skip to end of block */ + j = get_bits_count(&s->gb); + if (start_posn + asset_size[i] * 8 > j) + skip_bits_long(&s->gb, start_posn + asset_size[i] * 8 - j); + } +} diff --git a/libavcodec/dca_exss.h b/libavcodec/dca_exss.h new file mode 100644 index 0000000000..6689bbb3fb --- /dev/null +++ b/libavcodec/dca_exss.h @@ -0,0 +1,28 @@ +/* + * DCA ExSS extension + * + * This file is part of FFmpeg. + * + * FFmpeg 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 2.1 of the License, or (at your option) any later version. + * + * FFmpeg 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 FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVCODEC_DCA_EXSS_H +#define AVCODEC_DCA_EXSS_H + +#include "dca.h" + +void ff_dca_exss_parse_header(DCAContext *s); + +#endif /* AVCODEC_DCA_EXSS_H */ diff --git a/libavcodec/dcadec.c b/libavcodec/dcadec.c index 4e38511aa0..c6190a58a0 100644 --- a/libavcodec/dcadec.c +++ b/libavcodec/dcadec.c @@ -40,6 +40,7 @@ #include "dcadata.h" #include "dcadsp.h" #include "dcahuff.h" +#include "dca_exss.h" #include "fft.h" #include "fmtconvert.h" #include "get_bits.h" @@ -53,15 +54,6 @@ //#define TRACE -#define DCA_PRIM_CHANNELS_MAX (7) -#define DCA_ABITS_MAX (32) /* Should be 28 */ -#define DCA_SUBSUBFRAMES_MAX (4) -#define DCA_SUBFRAMES_MAX (16) -#define DCA_BLOCKS_MAX (16) -#define DCA_LFE_MAX (3) -#define DCA_CHSETS_MAX (4) -#define DCA_CHSET_CHANS_MAX (8) - enum DCAMode { DCA_MONO = 0, DCA_CHANNEL, @@ -76,25 +68,6 @@ enum DCAMode { DCA_4F2R }; -/* these are unconfirmed but should be mostly correct */ -enum DCAExSSSpeakerMask { - DCA_EXSS_FRONT_CENTER = 0x0001, - DCA_EXSS_FRONT_LEFT_RIGHT = 0x0002, - DCA_EXSS_SIDE_REAR_LEFT_RIGHT = 0x0004, - DCA_EXSS_LFE = 0x0008, - DCA_EXSS_REAR_CENTER = 0x0010, - DCA_EXSS_FRONT_HIGH_LEFT_RIGHT = 0x0020, - DCA_EXSS_REAR_LEFT_RIGHT = 0x0040, - DCA_EXSS_FRONT_HIGH_CENTER = 0x0080, - DCA_EXSS_OVERHEAD = 0x0100, - DCA_EXSS_CENTER_LEFT_RIGHT = 0x0200, - DCA_EXSS_WIDE_LEFT_RIGHT = 0x0400, - DCA_EXSS_SIDE_LEFT_RIGHT = 0x0800, - DCA_EXSS_LFE2 = 0x1000, - DCA_EXSS_SIDE_HIGH_LEFT_RIGHT = 0x2000, - DCA_EXSS_REAR_HIGH_CENTER = 0x4000, - DCA_EXSS_REAR_HIGH_LEFT_RIGHT = 0x8000, -}; enum DCAXxchSpeakerMask { DCA_XXCH_FRONT_CENTER = 0x0000001, @@ -158,19 +131,6 @@ static const uint32_t map_xxch_to_native[28] = { AV_CH_BACK_RIGHT /* read low right -- dup */ }; -enum DCAExtensionMask { - DCA_EXT_CORE = 0x001, ///< core in core substream - DCA_EXT_XXCH = 0x002, ///< XXCh channels extension in core substream - DCA_EXT_X96 = 0x004, ///< 96/24 extension in core substream - DCA_EXT_XCH = 0x008, ///< XCh channel extension in core substream - DCA_EXT_EXSS_CORE = 0x010, ///< core in ExSS (extension substream) - DCA_EXT_EXSS_XBR = 0x020, ///< extended bitrate extension in ExSS - DCA_EXT_EXSS_XXCH = 0x040, ///< XXCh channels extension in ExSS - DCA_EXT_EXSS_X96 = 0x080, ///< 96/24 extension in ExSS - DCA_EXT_EXSS_LBR = 0x100, ///< low bitrate component in ExSS - DCA_EXT_EXSS_XLL = 0x200, ///< lossless extension in ExSS -}; - /* -1 are reserved or unknown */ static const int dca_ext_audio_descr_mask[] = { DCA_EXT_XCH, @@ -183,9 +143,6 @@ static const int dca_ext_audio_descr_mask[] = { -1, }; -/* extensions that reside in core substream */ -#define DCA_CORE_EXTS (DCA_EXT_XCH | DCA_EXT_XXCH | DCA_EXT_X96) - /* Tables for mapping dts channel configurations to libavcodec multichannel api. * Some compromises have been made for special configurations. Most configurations * are never used so complete accuracy is not needed. @@ -321,11 +278,6 @@ static const int8_t dca_channel_reorder_nolfe_xch[][9] = { #define HEADER_SIZE 14 -#define DCA_MAX_FRAME_SIZE 16384 -#define DCA_MAX_EXSS_HEADER_SIZE 4096 - -#define DCA_BUFFER_PADDING_SIZE 1024 - #define DCA_NSYNCAUX 0x9A1105A0 /** Bit allocation */ @@ -348,137 +300,6 @@ static av_always_inline int get_bitalloc(GetBitContext *gb, BitAlloc *ba, ba->offset; } -typedef struct DCAContext { - const AVClass *class; ///< class for AVOptions - AVCodecContext *avctx; - /* 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 bit_rate_index; ///< transmission bit rate index - - 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 total_channels; ///< number of channels including extensions - int prim_channels; ///< number of primary audio channels - int subband_activity[DCA_PRIM_CHANNELS_MAX]; ///< subband activity count - int vq_start_subband[DCA_PRIM_CHANNELS_MAX]; ///< high frequency vq start subband - int joint_intensity[DCA_PRIM_CHANNELS_MAX]; ///< joint intensity coding index - int transient_huffman[DCA_PRIM_CHANNELS_MAX]; ///< transient mode code book - int scalefactor_huffman[DCA_PRIM_CHANNELS_MAX]; ///< scale factor code book - int bitalloc_huffman[DCA_PRIM_CHANNELS_MAX]; ///< bit allocation quantizer select - int quant_index_huffman[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< quantization index codebook select - float scalefactor_adj[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< scale factor adjustment - - /* Primary audio coding side information */ - int subsubframes[DCA_SUBFRAMES_MAX]; ///< number of subsubframes - int partial_samples[DCA_SUBFRAMES_MAX]; ///< partial subsubframe samples count - int prediction_mode[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< prediction mode (ADPCM used or not) - int prediction_vq[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< prediction VQ coefs - int bitalloc[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< bit allocation index - int transition_mode[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< transition mode (transients) - int32_t scale_factor[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][2];///< scale factors (2 if transient) - int joint_huff[DCA_PRIM_CHANNELS_MAX]; ///< joint subband scale factors codebook - int joint_scale_factor[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< joint subband scale factors - float downmix_coef[DCA_PRIM_CHANNELS_MAX + 1][2]; ///< stereo downmix coefficients - int dynrange_coef; ///< dynamic range coefficient - - /* Core substream's embedded downmix coefficients (cf. ETSI TS 102 114 V1.4.1) - * Input: primary audio channels (incl. LFE if present) - * Output: downmix audio channels (up to 4, no LFE) */ - uint8_t core_downmix; ///< embedded downmix coefficients available - uint8_t core_downmix_amode; ///< audio channel arrangement of embedded downmix - uint16_t core_downmix_codes[DCA_PRIM_CHANNELS_MAX + 1][4]; ///< embedded downmix coefficients (9-bit codes) - - int32_t high_freq_vq[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< VQ encoded high frequency subbands - - float lfe_data[2 * DCA_LFE_MAX * (DCA_BLOCKS_MAX + 4)]; ///< Low frequency effect data - int lfe_scale_factor; - - /* Subband samples history (for ADPCM) */ - DECLARE_ALIGNED(16, float, subband_samples_hist)[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][4]; - DECLARE_ALIGNED(32, float, subband_fir_hist)[DCA_PRIM_CHANNELS_MAX][512]; - DECLARE_ALIGNED(32, float, subband_fir_noidea)[DCA_PRIM_CHANNELS_MAX][32]; - int hist_index[DCA_PRIM_CHANNELS_MAX]; - DECLARE_ALIGNED(32, float, raXin)[32]; - - int output; ///< type of output - - DECLARE_ALIGNED(32, float, subband_samples)[DCA_BLOCKS_MAX][DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][8]; - float *samples_chanptr[DCA_PRIM_CHANNELS_MAX + 1]; - float *extra_channels[DCA_PRIM_CHANNELS_MAX + 1]; - uint8_t *extra_channels_buffer; - unsigned int extra_channels_buffer_size; - - uint8_t dca_buffer[DCA_MAX_FRAME_SIZE + DCA_MAX_EXSS_HEADER_SIZE + DCA_BUFFER_PADDING_SIZE]; - int dca_buffer_size; ///< how much data is in the dca_buffer - - const int8_t *channel_order_tab; ///< channel reordering table, lfe and non lfe - GetBitContext gb; - /* Current position in DCA frame */ - int current_subframe; - int current_subsubframe; - - int core_ext_mask; ///< present extensions in the core substream - - /* XCh extension information */ - int xch_present; ///< XCh extension present and valid - int xch_base_channel; ///< index of first (only) channel containing XCH data - int xch_disable; ///< whether the XCh extension should be decoded or not - - /* XXCH extension information */ - int xxch_chset; - int xxch_nbits_spk_mask; - uint32_t xxch_core_spkmask; - uint32_t xxch_spk_masks[4]; /* speaker masks, last element is core mask */ - int xxch_chset_nch[4]; - float xxch_dmix_sf[DCA_CHSETS_MAX]; - - uint32_t xxch_dmix_embedded; /* lower layer has mix pre-embedded, per chset */ - float xxch_dmix_coeff[DCA_PRIM_CHANNELS_MAX][32]; /* worst case sizing */ - - int8_t xxch_order_tab[32]; - int8_t lfe_index; - - /* ExSS header parser */ - int static_fields; ///< static fields present - int mix_metadata; ///< mixing metadata present - int num_mix_configs; ///< number of mix out configurations - int mix_config_num_ch[4]; ///< number of channels in each mix out configuration - - int profile; - - int debug_flag; ///< used for suppressing repeated error messages output - AVFloatDSPContext *fdsp; - FFTContext imdct; - SynthFilterContext synth; - DCADSPContext dcadsp; - FmtConvertContext fmt_conv; -} DCAContext; - static float dca_dmix_code(unsigned code); static const uint16_t dca_vlc_offs[] = { @@ -1594,197 +1415,7 @@ static int dca_decode_block(DCAContext *s, int base_channel, int block_index) return 0; } -/** - * Return the number of channels in an ExSS speaker mask (HD) - */ -static int dca_exss_mask2count(int mask) -{ - /* count bits that mean speaker pairs twice */ - return av_popcount(mask) + - av_popcount(mask & (DCA_EXSS_CENTER_LEFT_RIGHT | - DCA_EXSS_FRONT_LEFT_RIGHT | - DCA_EXSS_FRONT_HIGH_LEFT_RIGHT | - DCA_EXSS_WIDE_LEFT_RIGHT | - DCA_EXSS_SIDE_LEFT_RIGHT | - DCA_EXSS_SIDE_HIGH_LEFT_RIGHT | - DCA_EXSS_SIDE_REAR_LEFT_RIGHT | - DCA_EXSS_REAR_LEFT_RIGHT | - DCA_EXSS_REAR_HIGH_LEFT_RIGHT)); -} - -/** - * Skip mixing coefficients of a single mix out configuration (HD) - */ -static void dca_exss_skip_mix_coeffs(GetBitContext *gb, int channels, int out_ch) -{ - int i; - - for (i = 0; i < channels; i++) { - int mix_map_mask = get_bits(gb, out_ch); - int num_coeffs = av_popcount(mix_map_mask); - skip_bits_long(gb, num_coeffs * 6); - } -} - -/** - * Parse extension substream asset header (HD) - */ -static int dca_exss_parse_asset_header(DCAContext *s) -{ - int header_pos = get_bits_count(&s->gb); - int header_size; - int channels = 0; - int embedded_stereo = 0; - int embedded_6ch = 0; - int drc_code_present; - int extensions_mask = 0; - int i, j; - - if (get_bits_left(&s->gb) < 16) - return -1; - - /* We will parse just enough to get to the extensions bitmask with which - * we can set the profile value. */ - - header_size = get_bits(&s->gb, 9) + 1; - skip_bits(&s->gb, 3); // asset index - - if (s->static_fields) { - if (get_bits1(&s->gb)) - skip_bits(&s->gb, 4); // asset type descriptor - if (get_bits1(&s->gb)) - skip_bits_long(&s->gb, 24); // language descriptor - - if (get_bits1(&s->gb)) { - /* How can one fit 1024 bytes of text here if the maximum value - * for the asset header size field above was 512 bytes? */ - int text_length = get_bits(&s->gb, 10) + 1; - if (get_bits_left(&s->gb) < text_length * 8) - return -1; - skip_bits_long(&s->gb, text_length * 8); // info text - } - - skip_bits(&s->gb, 5); // bit resolution - 1 - skip_bits(&s->gb, 4); // max sample rate code - channels = get_bits(&s->gb, 8) + 1; - - if (get_bits1(&s->gb)) { // 1-to-1 channels to speakers - int spkr_remap_sets; - int spkr_mask_size = 16; - int num_spkrs[7]; - - if (channels > 2) - embedded_stereo = get_bits1(&s->gb); - if (channels > 6) - embedded_6ch = get_bits1(&s->gb); - - if (get_bits1(&s->gb)) { - spkr_mask_size = (get_bits(&s->gb, 2) + 1) << 2; - skip_bits(&s->gb, spkr_mask_size); // spkr activity mask - } - - spkr_remap_sets = get_bits(&s->gb, 3); - - for (i = 0; i < spkr_remap_sets; i++) { - /* std layout mask for each remap set */ - num_spkrs[i] = dca_exss_mask2count(get_bits(&s->gb, spkr_mask_size)); - } - - for (i = 0; i < spkr_remap_sets; i++) { - int num_dec_ch_remaps = get_bits(&s->gb, 5) + 1; - if (get_bits_left(&s->gb) < 0) - return -1; - - for (j = 0; j < num_spkrs[i]; j++) { - int remap_dec_ch_mask = get_bits_long(&s->gb, num_dec_ch_remaps); - int num_dec_ch = av_popcount(remap_dec_ch_mask); - skip_bits_long(&s->gb, num_dec_ch * 5); // remap codes - } - } - } else { - skip_bits(&s->gb, 3); // representation type - } - } - - drc_code_present = get_bits1(&s->gb); - if (drc_code_present) - get_bits(&s->gb, 8); // drc code - - if (get_bits1(&s->gb)) - skip_bits(&s->gb, 5); // dialog normalization code - - if (drc_code_present && embedded_stereo) - get_bits(&s->gb, 8); // drc stereo code - - if (s->mix_metadata && get_bits1(&s->gb)) { - skip_bits(&s->gb, 1); // external mix - skip_bits(&s->gb, 6); // post mix gain code - - if (get_bits(&s->gb, 2) != 3) // mixer drc code - skip_bits(&s->gb, 3); // drc limit - else - skip_bits(&s->gb, 8); // custom drc code - - if (get_bits1(&s->gb)) // channel specific scaling - for (i = 0; i < s->num_mix_configs; i++) - skip_bits_long(&s->gb, s->mix_config_num_ch[i] * 6); // scale codes - else - skip_bits_long(&s->gb, s->num_mix_configs * 6); // scale codes - - for (i = 0; i < s->num_mix_configs; i++) { - if (get_bits_left(&s->gb) < 0) - return -1; - dca_exss_skip_mix_coeffs(&s->gb, channels, s->mix_config_num_ch[i]); - if (embedded_6ch) - dca_exss_skip_mix_coeffs(&s->gb, 6, s->mix_config_num_ch[i]); - if (embedded_stereo) - dca_exss_skip_mix_coeffs(&s->gb, 2, s->mix_config_num_ch[i]); - } - } - - switch (get_bits(&s->gb, 2)) { - case 0: - extensions_mask = get_bits(&s->gb, 12); - break; - case 1: - extensions_mask = DCA_EXT_EXSS_XLL; - break; - case 2: - extensions_mask = DCA_EXT_EXSS_LBR; - break; - case 3: - extensions_mask = 0; /* aux coding */ - break; - } - - /* not parsed further, we were only interested in the extensions mask */ - - if (get_bits_left(&s->gb) < 0) - return -1; - - if (get_bits_count(&s->gb) - header_pos > header_size * 8) { - av_log(s->avctx, AV_LOG_WARNING, "Asset header size mismatch.\n"); - return -1; - } - skip_bits_long(&s->gb, header_pos + header_size * 8 - get_bits_count(&s->gb)); - - if (extensions_mask & DCA_EXT_EXSS_XLL) - s->profile = FF_PROFILE_DTS_HD_MA; - else if (extensions_mask & (DCA_EXT_EXSS_XBR | DCA_EXT_EXSS_X96 | - DCA_EXT_EXSS_XXCH)) - s->profile = FF_PROFILE_DTS_HD_HRA; - - if (!(extensions_mask & DCA_EXT_CORE)) - av_log(s->avctx, AV_LOG_WARNING, "DTS core detection mismatch.\n"); - if ((extensions_mask & DCA_CORE_EXTS) != s->core_ext_mask) - av_log(s->avctx, AV_LOG_WARNING, - "DTS extensions detection mismatch (%d, %d)\n", - extensions_mask & DCA_CORE_EXTS, s->core_ext_mask); - - return 0; -} - -static int dca_xbr_parse_frame(DCAContext *s) +int ff_dca_xbr_parse_frame(DCAContext *s) { int scale_table_high[DCA_CHSET_CHANS_MAX][DCA_SUBBANDS][2]; int active_bands[DCA_CHSETS_MAX][DCA_CHSET_CHANS_MAX]; @@ -1944,8 +1575,9 @@ static int dca_xbr_parse_frame(DCAContext *s) return 0; } + /* parse initial header for XXCH and dump details */ -static int dca_xxch_decode_frame(DCAContext *s) +int ff_dca_xxch_decode_frame(DCAContext *s) { int hdr_size, spkmsk_bits, num_chsets, core_spk, hdr_pos; int i, chset, base_channel, chstart, fsize[8]; @@ -1998,122 +1630,6 @@ static int dca_xxch_decode_frame(DCAContext *s) return 0; } -/** - * Parse extension substream header (HD) - */ -static void dca_exss_parse_header(DCAContext *s) -{ - int asset_size[8]; - int ss_index; - int blownup; - int num_audiop = 1; - int num_assets = 1; - int active_ss_mask[8]; - int i, j; - int start_posn; - int hdrsize; - uint32_t mkr; - - if (get_bits_left(&s->gb) < 52) - return; - - start_posn = get_bits_count(&s->gb) - 32; - - skip_bits(&s->gb, 8); // user data - ss_index = get_bits(&s->gb, 2); - - blownup = get_bits1(&s->gb); - hdrsize = get_bits(&s->gb, 8 + 4 * blownup) + 1; // header_size - skip_bits(&s->gb, 16 + 4 * blownup); // hd_size - - s->static_fields = get_bits1(&s->gb); - if (s->static_fields) { - skip_bits(&s->gb, 2); // reference clock code - skip_bits(&s->gb, 3); // frame duration code - - if (get_bits1(&s->gb)) - skip_bits_long(&s->gb, 36); // timestamp - - /* a single stream can contain multiple audio assets that can be - * combined to form multiple audio presentations */ - - num_audiop = get_bits(&s->gb, 3) + 1; - if (num_audiop > 1) { - avpriv_request_sample(s->avctx, - "Multiple DTS-HD audio presentations"); - /* ignore such streams for now */ - return; - } - - num_assets = get_bits(&s->gb, 3) + 1; - if (num_assets > 1) { - avpriv_request_sample(s->avctx, "Multiple DTS-HD audio assets"); - /* ignore such streams for now */ - return; - } - - for (i = 0; i < num_audiop; i++) - active_ss_mask[i] = get_bits(&s->gb, ss_index + 1); - - for (i = 0; i < num_audiop; i++) - for (j = 0; j <= ss_index; j++) - if (active_ss_mask[i] & (1 << j)) - skip_bits(&s->gb, 8); // active asset mask - - s->mix_metadata = get_bits1(&s->gb); - if (s->mix_metadata) { - int mix_out_mask_size; - - skip_bits(&s->gb, 2); // adjustment level - mix_out_mask_size = (get_bits(&s->gb, 2) + 1) << 2; - s->num_mix_configs = get_bits(&s->gb, 2) + 1; - - for (i = 0; i < s->num_mix_configs; i++) { - int mix_out_mask = get_bits(&s->gb, mix_out_mask_size); - s->mix_config_num_ch[i] = dca_exss_mask2count(mix_out_mask); - } - } - } - - av_assert0(num_assets > 0); // silence a warning - - for (i = 0; i < num_assets; i++) - asset_size[i] = get_bits_long(&s->gb, 16 + 4 * blownup); - - for (i = 0; i < num_assets; i++) { - if (dca_exss_parse_asset_header(s)) - return; - } - - /* not parsed further, we were only interested in the extensions mask - * from the asset header */ - - j = get_bits_count(&s->gb); - if (start_posn + hdrsize * 8 > j) - skip_bits_long(&s->gb, start_posn + hdrsize * 8 - j); - - for (i = 0; i < num_assets; i++) { - start_posn = get_bits_count(&s->gb); - mkr = get_bits_long(&s->gb, 32); - - /* parse extensions that we know about */ - if (mkr == 0x655e315e) { - dca_xbr_parse_frame(s); - } else if (mkr == 0x47004a03) { - dca_xxch_decode_frame(s); - s->core_ext_mask |= DCA_EXT_XXCH; /* xxx use for chan reordering */ - } else { - av_log(s->avctx, AV_LOG_DEBUG, - "DTS-ExSS: unknown marker = 0x%08x\n", mkr); - } - - /* skip to end of block */ - j = get_bits_count(&s->gb); - if (start_posn + asset_size[i] * 8 > j) - skip_bits_long(&s->gb, start_posn + asset_size[i] * 8 - j); - } -} - static float dca_dmix_code(unsigned code) { int sign = (code >> 8) - 1; @@ -2291,7 +1807,7 @@ static int dca_decode_frame(AVCodecContext *avctx, void *data, /* usually found either in core or HD part in DTS-HD HRA streams, * but not in DTS-ES which contains XCh extensions instead */ s->core_ext_mask |= DCA_EXT_XXCH; - dca_xxch_decode_frame(s); + ff_dca_xxch_decode_frame(s); break; case 0x1d95f262: { @@ -2325,7 +1841,7 @@ static int dca_decode_frame(AVCodecContext *avctx, void *data, /* check for ExSS (HD part) */ if (s->dca_buffer_size - s->frame_size > 32 && get_bits_long(&s->gb, 32) == DCA_HD_MARKER) - dca_exss_parse_header(s); + ff_dca_exss_parse_header(s); avctx->profile = s->profile; diff --git a/libavcodec/dcaenc.c b/libavcodec/dcaenc.c index 905cdc9ade..2325b0b0ec 100644 --- a/libavcodec/dcaenc.c +++ b/libavcodec/dcaenc.c @@ -43,7 +43,7 @@ #define SUBBAND_SAMPLES (SUBFRAMES * SUBSUBFRAMES * 8) #define AUBANDS 25 -typedef struct DCAContext { +typedef struct DCAEncContext { PutBitContext pb; int frame_size; int frame_bits; @@ -73,7 +73,7 @@ typedef struct DCAContext { int32_t worst_quantization_noise; int32_t worst_noise_ever; int consumed_bits; -} DCAContext; +} DCAEncContext; static int32_t cos_table[2048]; static int32_t band_interpolation[2][512]; @@ -105,7 +105,7 @@ static double gammafilter(int i, double f) static int encode_init(AVCodecContext *avctx) { - DCAContext *c = avctx->priv_data; + DCAEncContext *c = avctx->priv_data; uint64_t layout = avctx->channel_layout; int i, min_frame_bits; @@ -235,7 +235,7 @@ static inline int32_t mul32(int32_t a, int32_t b) return r >> 32; } -static void subband_transform(DCAContext *c, const int32_t *input) +static void subband_transform(DCAEncContext *c, const int32_t *input) { int ch, subs, i, k, j; @@ -285,7 +285,7 @@ static void subband_transform(DCAContext *c, const int32_t *input) } } -static void lfe_downsample(DCAContext *c, const int32_t *input) +static void lfe_downsample(DCAEncContext *c, const int32_t *input) { /* FIXME: make 128x LFE downsampling possible */ int i, j, lfes; @@ -442,11 +442,11 @@ static void adjust_jnd(int samplerate_index, out_cb[j] = add_cb(out_cb[j], -out_cb_unnorm[j] - ca_cb - cs_cb); } -typedef void (*walk_band_t)(DCAContext *c, int band1, int band2, int f, +typedef void (*walk_band_t)(DCAEncContext *c, int band1, int band2, int f, int32_t spectrum1, int32_t spectrum2, int channel, int32_t * arg); -static void walk_band_low(DCAContext *c, int band, int channel, +static void walk_band_low(DCAEncContext *c, int band, int channel, walk_band_t walk, int32_t *arg) { int f; @@ -461,7 +461,7 @@ static void walk_band_low(DCAContext *c, int band, int channel, } } -static void walk_band_high(DCAContext *c, int band, int channel, +static void walk_band_high(DCAEncContext *c, int band, int channel, walk_band_t walk, int32_t *arg) { int f; @@ -476,7 +476,7 @@ static void walk_band_high(DCAContext *c, int band, int channel, } } -static void update_band_masking(DCAContext *c, int band1, int band2, +static void update_band_masking(DCAEncContext *c, int band1, int band2, int f, int32_t spectrum1, int32_t spectrum2, int channel, int32_t * arg) { @@ -486,7 +486,7 @@ static void update_band_masking(DCAContext *c, int band1, int band2, c->band_masking_cb[band1] = value; } -static void calc_masking(DCAContext *c, const int32_t *input) +static void calc_masking(DCAEncContext *c, const int32_t *input) { int i, k, band, ch, ssf; int32_t data[512]; @@ -519,7 +519,7 @@ static void calc_masking(DCAContext *c, const int32_t *input) } } -static void find_peaks(DCAContext *c) +static void find_peaks(DCAEncContext *c) { int band, ch; @@ -552,7 +552,7 @@ static const int snr_fudge = 128; #define USED_NABITS 2 #define USED_26ABITS 4 -static int init_quantization_noise(DCAContext *c, int noise) +static int init_quantization_noise(DCAEncContext *c, int noise) { int ch, band, ret = 0; @@ -589,7 +589,7 @@ static int init_quantization_noise(DCAContext *c, int noise) return ret; } -static void assign_bits(DCAContext *c) +static void assign_bits(DCAEncContext *c) { /* Find the bounds where the binary search should work */ int low, high, down; @@ -627,7 +627,7 @@ out: c->worst_noise_ever = high; } -static void shift_history(DCAContext *c, const int32_t *input) +static void shift_history(DCAEncContext *c, const int32_t *input) { int k, ch; @@ -677,7 +677,7 @@ static int calc_one_scale(int32_t peak_cb, int abits, softfloat *quant) return our_nscale; } -static void calc_scales(DCAContext *c) +static void calc_scales(DCAEncContext *c) { int band, ch; @@ -691,7 +691,7 @@ static void calc_scales(DCAContext *c) c->lfe_scale_factor = calc_one_scale(c->lfe_peak_cb, 11, &c->lfe_quant); } -static void quantize_all(DCAContext *c) +static void quantize_all(DCAEncContext *c) { int sample, band, ch; @@ -701,7 +701,7 @@ static void quantize_all(DCAContext *c) c->quantized[sample][band][ch] = quantize_value(c->subband[sample][band][ch], c->quant[band][ch]); } -static void put_frame_header(DCAContext *c) +static void put_frame_header(DCAEncContext *c) { /* SYNC */ put_bits(&c->pb, 16, 0x7ffe); @@ -784,7 +784,7 @@ static void put_frame_header(DCAContext *c) put_bits(&c->pb, 4, 0); } -static void put_primary_audio_header(DCAContext *c) +static void put_primary_audio_header(DCAEncContext *c) { static const int bitlen[11] = { 0, 1, 2, 2, 2, 2, 3, 3, 3, 3, 3 }; static const int thr[11] = { 0, 1, 3, 3, 3, 3, 7, 7, 7, 7, 7 }; @@ -830,7 +830,7 @@ static void put_primary_audio_header(DCAContext *c) /* Audio header CRC check word: not transmitted */ } -static void put_subframe_samples(DCAContext *c, int ss, int band, int ch) +static void put_subframe_samples(DCAEncContext *c, int ss, int band, int ch) { if (c->abits[band][ch] <= 7) { int sum, i, j; @@ -853,7 +853,7 @@ static void put_subframe_samples(DCAContext *c, int ss, int band, int ch) } } -static void put_subframe(DCAContext *c, int subframe) +static void put_subframe(DCAEncContext *c, int subframe) { int i, band, ss, ch; @@ -913,7 +913,7 @@ static void put_subframe(DCAContext *c, int subframe) static int encode_frame(AVCodecContext *avctx, AVPacket *avpkt, const AVFrame *frame, int *got_packet_ptr) { - DCAContext *c = avctx->priv_data; + DCAEncContext *c = avctx->priv_data; const int32_t *samples; int ret, i; @@ -958,7 +958,7 @@ AVCodec ff_dca_encoder = { .long_name = NULL_IF_CONFIG_SMALL("DCA (DTS Coherent Acoustics)"), .type = AVMEDIA_TYPE_AUDIO, .id = AV_CODEC_ID_DTS, - .priv_data_size = sizeof(DCAContext), + .priv_data_size = sizeof(DCAEncContext), .init = encode_init, .encode2 = encode_frame, .capabilities = CODEC_CAP_EXPERIMENTAL,