Subversion Repositories Kolibri OS

Rev

Blame | Last modification | View Log | RSS feed

  1. /*
  2.  * Digital Speech Standard (DSS) demuxer
  3.  * Copyright (c) 2014 Oleksij Rempel <linux@rempel-privat.de>
  4.  *
  5.  * This file is part of FFmpeg.
  6.  *
  7.  * FFmpeg is free software; you can redistribute it and/or
  8.  * modify it under the terms of the GNU Lesser General Public
  9.  * License as published by the Free Software Foundation; either
  10.  * version 2.1 of the License, or (at your option) any later version.
  11.  *
  12.  * FFmpeg is distributed in the hope that it will be useful,
  13.  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14.  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15.  * Lesser General Public License for more details.
  16.  *
  17.  * You should have received a copy of the GNU Lesser General Public
  18.  * License along with FFmpeg; if not, write to the Free Software
  19.  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  20.  */
  21.  
  22. #include "libavutil/attributes.h"
  23. #include "libavutil/bswap.h"
  24. #include "libavutil/channel_layout.h"
  25. #include "libavutil/intreadwrite.h"
  26.  
  27. #include "avformat.h"
  28. #include "internal.h"
  29.  
  30. #define DSS_HEAD_OFFSET_AUTHOR        0xc
  31. #define DSS_AUTHOR_SIZE               16
  32.  
  33. #define DSS_HEAD_OFFSET_START_TIME    0x26
  34. #define DSS_HEAD_OFFSET_END_TIME      0x32
  35. #define DSS_TIME_SIZE                 12
  36.  
  37. #define DSS_HEAD_OFFSET_ACODEC        0x2a4
  38. #define DSS_ACODEC_DSS_SP             0x0    /* SP mode */
  39. #define DSS_ACODEC_G723_1             0x2    /* LP mode */
  40.  
  41. #define DSS_HEAD_OFFSET_COMMENT       0x31e
  42. #define DSS_COMMENT_SIZE              64
  43.  
  44. #define DSS_BLOCK_SIZE                512
  45. #define DSS_HEADER_SIZE              (DSS_BLOCK_SIZE * 2)
  46. #define DSS_AUDIO_BLOCK_HEADER_SIZE   6
  47. #define DSS_FRAME_SIZE                42
  48.  
  49. static const uint8_t frame_size[4] = { 24, 20, 4, 1 };
  50.  
  51. typedef struct DSSDemuxContext {
  52.     unsigned int audio_codec;
  53.     int counter;
  54.     int swap;
  55.     int dss_sp_swap_byte;
  56.     int8_t *dss_sp_buf;
  57.  
  58.     int packet_size;
  59. } DSSDemuxContext;
  60.  
  61. static int dss_probe(AVProbeData *p)
  62. {
  63.     if (AV_RL32(p->buf) != MKTAG(0x2, 'd', 's', 's'))
  64.         return 0;
  65.  
  66.     return AVPROBE_SCORE_MAX;
  67. }
  68.  
  69. static int dss_read_metadata_date(AVFormatContext *s, unsigned int offset,
  70.                                   const char *key)
  71. {
  72.     AVIOContext *pb = s->pb;
  73.     char datetime[64], string[DSS_TIME_SIZE + 1] = { 0 };
  74.     int y, month, d, h, minute, sec;
  75.     int ret;
  76.  
  77.     avio_seek(pb, offset, SEEK_SET);
  78.  
  79.     ret = avio_read(s->pb, string, DSS_TIME_SIZE);
  80.     if (ret < DSS_TIME_SIZE)
  81.         return ret < 0 ? ret : AVERROR_EOF;
  82.  
  83.     if (sscanf(string, "%2d%2d%2d%2d%2d%2d", &y, &month, &d, &h, &minute, &sec) != 6)
  84.         return AVERROR_INVALIDDATA;
  85.     /* We deal with a two-digit year here, so set the default date to 2000
  86.      * and hope it will never be used in the next century. */
  87.     snprintf(datetime, sizeof(datetime), "%.4d-%.2d-%.2dT%.2d:%.2d:%.2d",
  88.              y + 2000, month, d, h, minute, sec);
  89.     return av_dict_set(&s->metadata, key, datetime, 0);
  90. }
  91.  
  92. static int dss_read_metadata_string(AVFormatContext *s, unsigned int offset,
  93.                                     unsigned int size, const char *key)
  94. {
  95.     AVIOContext *pb = s->pb;
  96.     char *value;
  97.     int ret;
  98.  
  99.     avio_seek(pb, offset, SEEK_SET);
  100.  
  101.     value = av_mallocz(size + 1);
  102.     if (!value)
  103.         return AVERROR(ENOMEM);
  104.  
  105.     ret = avio_read(s->pb, value, size);
  106.     if (ret < size) {
  107.         ret = ret < 0 ? ret : AVERROR_EOF;
  108.         goto exit;
  109.     }
  110.  
  111.     ret = av_dict_set(&s->metadata, key, value, 0);
  112.  
  113. exit:
  114.     av_free(value);
  115.     return ret;
  116. }
  117.  
  118. static int dss_read_header(AVFormatContext *s)
  119. {
  120.     DSSDemuxContext *ctx = s->priv_data;
  121.     AVIOContext *pb = s->pb;
  122.     AVStream *st;
  123.     int ret;
  124.  
  125.     st = avformat_new_stream(s, NULL);
  126.     if (!st)
  127.         return AVERROR(ENOMEM);
  128.  
  129.     ret = dss_read_metadata_string(s, DSS_HEAD_OFFSET_AUTHOR,
  130.                                    DSS_AUTHOR_SIZE, "author");
  131.     if (ret)
  132.         return ret;
  133.  
  134.     ret = dss_read_metadata_date(s, DSS_HEAD_OFFSET_END_TIME, "date");
  135.     if (ret)
  136.         return ret;
  137.  
  138.     ret = dss_read_metadata_string(s, DSS_HEAD_OFFSET_COMMENT,
  139.                                    DSS_COMMENT_SIZE, "comment");
  140.     if (ret)
  141.         return ret;
  142.  
  143.     avio_seek(pb, DSS_HEAD_OFFSET_ACODEC, SEEK_SET);
  144.     ctx->audio_codec = avio_r8(pb);
  145.  
  146.     if (ctx->audio_codec == DSS_ACODEC_DSS_SP) {
  147.         st->codec->codec_id    = AV_CODEC_ID_DSS_SP;
  148.         st->codec->sample_rate = 11025;
  149.     } else if (ctx->audio_codec == DSS_ACODEC_G723_1) {
  150.         st->codec->codec_id    = AV_CODEC_ID_G723_1;
  151.         st->codec->sample_rate = 8000;
  152.     } else {
  153.         avpriv_request_sample(s, "Support for codec %x in DSS",
  154.                               ctx->audio_codec);
  155.         return AVERROR_PATCHWELCOME;
  156.     }
  157.  
  158.     st->codec->codec_type     = AVMEDIA_TYPE_AUDIO;
  159.     st->codec->channel_layout = AV_CH_LAYOUT_MONO;
  160.     st->codec->channels       = 1;
  161.  
  162.     avpriv_set_pts_info(st, 64, 1, st->codec->sample_rate);
  163.     st->start_time = 0;
  164.  
  165.     /* Jump over header */
  166.  
  167.     if (avio_seek(pb, DSS_HEADER_SIZE, SEEK_SET) != DSS_HEADER_SIZE)
  168.         return AVERROR(EIO);
  169.  
  170.     ctx->counter = 0;
  171.     ctx->swap    = 0;
  172.  
  173.     ctx->dss_sp_buf = av_malloc(DSS_FRAME_SIZE + 1);
  174.     if (!ctx->dss_sp_buf)
  175.         return AVERROR(ENOMEM);
  176.  
  177.     return 0;
  178. }
  179.  
  180. static void dss_skip_audio_header(AVFormatContext *s, AVPacket *pkt)
  181. {
  182.     DSSDemuxContext *ctx = s->priv_data;
  183.     AVIOContext *pb = s->pb;
  184.  
  185.     avio_skip(pb, DSS_AUDIO_BLOCK_HEADER_SIZE);
  186.     ctx->counter += DSS_BLOCK_SIZE - DSS_AUDIO_BLOCK_HEADER_SIZE;
  187. }
  188.  
  189. static void dss_sp_byte_swap(DSSDemuxContext *ctx,
  190.                              uint8_t *dst, const uint8_t *src)
  191. {
  192.     int i;
  193.  
  194.     if (ctx->swap) {
  195.         for (i = 3; i < DSS_FRAME_SIZE; i += 2)
  196.             dst[i] = src[i];
  197.  
  198.         for (i = 0; i < DSS_FRAME_SIZE - 2; i += 2)
  199.             dst[i] = src[i + 4];
  200.  
  201.         dst[1] = ctx->dss_sp_swap_byte;
  202.     } else {
  203.         memcpy(dst, src, DSS_FRAME_SIZE);
  204.         ctx->dss_sp_swap_byte = src[DSS_FRAME_SIZE - 2];
  205.     }
  206.  
  207.     /* make sure byte 40 is always 0 */
  208.     dst[DSS_FRAME_SIZE - 2] = 0;
  209.     ctx->swap             ^= 1;
  210. }
  211.  
  212. static int dss_sp_read_packet(AVFormatContext *s, AVPacket *pkt)
  213. {
  214.     DSSDemuxContext *ctx = s->priv_data;
  215.     AVStream *st = s->streams[0];
  216.     int read_size, ret, offset = 0, buff_offset = 0;
  217.     int64_t pos = avio_tell(s->pb);
  218.  
  219.     if (ctx->counter == 0)
  220.         dss_skip_audio_header(s, pkt);
  221.  
  222.     if (ctx->swap) {
  223.         read_size   = DSS_FRAME_SIZE - 2;
  224.         buff_offset = 3;
  225.     } else
  226.         read_size = DSS_FRAME_SIZE;
  227.  
  228.     ctx->counter -= read_size;
  229.     ctx->packet_size = DSS_FRAME_SIZE - 1;
  230.  
  231.     ret = av_new_packet(pkt, DSS_FRAME_SIZE);
  232.     if (ret < 0)
  233.         return ret;
  234.  
  235.     pkt->duration     = 264;
  236.     pkt->pos = pos;
  237.     pkt->stream_index = 0;
  238.     s->bit_rate = 8LL * ctx->packet_size * st->codec->sample_rate * 512 / (506 * pkt->duration);
  239.  
  240.     if (ctx->counter < 0) {
  241.         int size2 = ctx->counter + read_size;
  242.  
  243.         ret = avio_read(s->pb, ctx->dss_sp_buf + offset + buff_offset,
  244.                         size2 - offset);
  245.         if (ret < size2 - offset)
  246.             goto error_eof;
  247.  
  248.         dss_skip_audio_header(s, pkt);
  249.         offset = size2;
  250.     }
  251.  
  252.     ret = avio_read(s->pb, ctx->dss_sp_buf + offset + buff_offset,
  253.                     read_size - offset);
  254.     if (ret < read_size - offset)
  255.         goto error_eof;
  256.  
  257.     dss_sp_byte_swap(ctx, pkt->data, ctx->dss_sp_buf);
  258.  
  259.     if (ctx->dss_sp_swap_byte < 0) {
  260.         ret = AVERROR(EAGAIN);
  261.         goto error_eof;
  262.     }
  263.  
  264.     if (pkt->data[0] == 0xff)
  265.         return AVERROR_INVALIDDATA;
  266.  
  267.     return pkt->size;
  268.  
  269. error_eof:
  270.     av_free_packet(pkt);
  271.     return ret < 0 ? ret : AVERROR_EOF;
  272. }
  273.  
  274. static int dss_723_1_read_packet(AVFormatContext *s, AVPacket *pkt)
  275. {
  276.     DSSDemuxContext *ctx = s->priv_data;
  277.     AVStream *st = s->streams[0];
  278.     int size, byte, ret, offset;
  279.     int64_t pos = avio_tell(s->pb);
  280.  
  281.     if (ctx->counter == 0)
  282.         dss_skip_audio_header(s, pkt);
  283.  
  284.     /* We make one byte-step here. Don't forget to add offset. */
  285.     byte = avio_r8(s->pb);
  286.     if (byte == 0xff)
  287.         return AVERROR_INVALIDDATA;
  288.  
  289.     size = frame_size[byte & 3];
  290.  
  291.     ctx->packet_size = size;
  292.     ctx->counter -= size;
  293.  
  294.     ret = av_new_packet(pkt, size);
  295.     if (ret < 0)
  296.         return ret;
  297.     pkt->pos = pos;
  298.  
  299.     pkt->data[0]  = byte;
  300.     offset        = 1;
  301.     pkt->duration = 240;
  302.     s->bit_rate = 8LL * size * st->codec->sample_rate * 512 / (506 * pkt->duration);
  303.  
  304.     pkt->stream_index = 0;
  305.  
  306.     if (ctx->counter < 0) {
  307.         int size2 = ctx->counter + size;
  308.  
  309.         ret = avio_read(s->pb, pkt->data + offset,
  310.                         size2 - offset);
  311.         if (ret < size2 - offset) {
  312.             av_free_packet(pkt);
  313.             return ret < 0 ? ret : AVERROR_EOF;
  314.         }
  315.  
  316.         dss_skip_audio_header(s, pkt);
  317.         offset = size2;
  318.     }
  319.  
  320.     ret = avio_read(s->pb, pkt->data + offset, size - offset);
  321.     if (ret < size - offset) {
  322.         av_free_packet(pkt);
  323.         return ret < 0 ? ret : AVERROR_EOF;
  324.     }
  325.  
  326.     return pkt->size;
  327. }
  328.  
  329. static int dss_read_packet(AVFormatContext *s, AVPacket *pkt)
  330. {
  331.     DSSDemuxContext *ctx = s->priv_data;
  332.  
  333.     if (ctx->audio_codec == DSS_ACODEC_DSS_SP)
  334.         return dss_sp_read_packet(s, pkt);
  335.     else
  336.         return dss_723_1_read_packet(s, pkt);
  337. }
  338.  
  339. static int dss_read_close(AVFormatContext *s)
  340. {
  341.     DSSDemuxContext *ctx = s->priv_data;
  342.  
  343.     av_freep(&ctx->dss_sp_buf);
  344.  
  345.     return 0;
  346. }
  347.  
  348. static int dss_read_seek(AVFormatContext *s, int stream_index,
  349.                          int64_t timestamp, int flags)
  350. {
  351.     DSSDemuxContext *ctx = s->priv_data;
  352.     int64_t ret, seekto;
  353.     uint8_t header[DSS_AUDIO_BLOCK_HEADER_SIZE];
  354.     int offset;
  355.  
  356.     if (ctx->audio_codec == DSS_ACODEC_DSS_SP)
  357.         seekto = timestamp / 264 * 41 / 506 * 512;
  358.     else
  359.         seekto = timestamp / 240 * ctx->packet_size / 506 * 512;
  360.  
  361.     if (seekto < 0)
  362.         seekto = 0;
  363.  
  364.     seekto += DSS_HEADER_SIZE;
  365.  
  366.     ret = avio_seek(s->pb, seekto, SEEK_SET);
  367.     if (ret < 0)
  368.         return ret;
  369.  
  370.     avio_read(s->pb, header, DSS_AUDIO_BLOCK_HEADER_SIZE);
  371.     ctx->swap = !!(header[0] & 0x80);
  372.     offset = 2*header[1] + 2*ctx->swap;
  373.     if (offset < DSS_AUDIO_BLOCK_HEADER_SIZE)
  374.         return AVERROR_INVALIDDATA;
  375.     if (offset == DSS_AUDIO_BLOCK_HEADER_SIZE) {
  376.         ctx->counter = 0;
  377.         offset = avio_skip(s->pb, -DSS_AUDIO_BLOCK_HEADER_SIZE);
  378.     } else {
  379.         ctx->counter = DSS_BLOCK_SIZE - offset;
  380.         offset = avio_skip(s->pb, offset - DSS_AUDIO_BLOCK_HEADER_SIZE);
  381.     }
  382.     ctx->dss_sp_swap_byte = -1;
  383.     return 0;
  384. }
  385.  
  386.  
  387. AVInputFormat ff_dss_demuxer = {
  388.     .name           = "dss",
  389.     .long_name      = NULL_IF_CONFIG_SMALL("Digital Speech Standard (DSS)"),
  390.     .priv_data_size = sizeof(DSSDemuxContext),
  391.     .read_probe     = dss_probe,
  392.     .read_header    = dss_read_header,
  393.     .read_packet    = dss_read_packet,
  394.     .read_close     = dss_read_close,
  395.     .read_seek      = dss_read_seek,
  396.     .extensions     = "dss"
  397. };
  398.