You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

154 lines
5.1KB

  1. /*
  2. * Copyright (c) CMU 1993 Computer Science, Speech Group
  3. * Chengxiang Lu and Alex Hauptmann
  4. * Copyright (c) 2005 Steve Underwood <steveu at coppice.org>
  5. * Copyright (c) 2009 Kenan Gillet
  6. * Copyright (c) 2010 Martin Storsjo
  7. *
  8. * This file is part of Libav.
  9. *
  10. * Libav is free software; you can redistribute it and/or
  11. * modify it under the terms of the GNU Lesser General Public
  12. * License as published by the Free Software Foundation; either
  13. * version 2.1 of the License, or (at your option) any later version.
  14. *
  15. * Libav is distributed in the hope that it will be useful,
  16. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  18. * Lesser General Public License for more details.
  19. *
  20. * You should have received a copy of the GNU Lesser General Public
  21. * License along with Libav; if not, write to the Free Software
  22. * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  23. */
  24. /**
  25. * @file
  26. * G.722 ADPCM audio decoder
  27. *
  28. * This G.722 decoder is a bit-exact implementation of the ITU G.722
  29. * specification for all three specified bitrates - 64000bps, 56000bps
  30. * and 48000bps. It passes the ITU tests.
  31. *
  32. * @note For the 56000bps and 48000bps bitrates, the lowest 1 or 2 bits
  33. * respectively of each byte are ignored.
  34. */
  35. #include "libavutil/channel_layout.h"
  36. #include "libavutil/opt.h"
  37. #include "avcodec.h"
  38. #include "bitstream.h"
  39. #include "g722.h"
  40. #include "internal.h"
  41. #define OFFSET(x) offsetof(G722Context, x)
  42. #define AD AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_DECODING_PARAM
  43. static const AVOption options[] = {
  44. { "bits_per_codeword", "Bits per G722 codeword", OFFSET(bits_per_codeword), AV_OPT_TYPE_INT, { .i64 = 8 }, 6, 8, AD },
  45. { NULL }
  46. };
  47. static const AVClass g722_decoder_class = {
  48. .class_name = "g722 decoder",
  49. .item_name = av_default_item_name,
  50. .option = options,
  51. .version = LIBAVUTIL_VERSION_INT,
  52. };
  53. static av_cold int g722_decode_init(AVCodecContext * avctx)
  54. {
  55. G722Context *c = avctx->priv_data;
  56. avctx->channels = 1;
  57. avctx->channel_layout = AV_CH_LAYOUT_MONO;
  58. avctx->sample_fmt = AV_SAMPLE_FMT_S16;
  59. c->band[0].scale_factor = 8;
  60. c->band[1].scale_factor = 2;
  61. c->prev_samples_pos = 22;
  62. ff_g722dsp_init(&c->dsp);
  63. return 0;
  64. }
  65. static const int16_t low_inv_quant5[32] = {
  66. -35, -35, -2919, -2195, -1765, -1458, -1219, -1023,
  67. -858, -714, -587, -473, -370, -276, -190, -110,
  68. 2919, 2195, 1765, 1458, 1219, 1023, 858, 714,
  69. 587, 473, 370, 276, 190, 110, 35, -35
  70. };
  71. static const int16_t * const low_inv_quants[3] = { ff_g722_low_inv_quant6,
  72. low_inv_quant5,
  73. ff_g722_low_inv_quant4 };
  74. static int g722_decode_frame(AVCodecContext *avctx, void *data,
  75. int *got_frame_ptr, AVPacket *avpkt)
  76. {
  77. G722Context *c = avctx->priv_data;
  78. AVFrame *frame = data;
  79. int16_t *out_buf;
  80. int j, ret;
  81. const int skip = 8 - c->bits_per_codeword;
  82. const int16_t *quantizer_table = low_inv_quants[skip];
  83. BitstreamContext bc;
  84. /* get output buffer */
  85. frame->nb_samples = avpkt->size * 2;
  86. if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
  87. av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
  88. return ret;
  89. }
  90. out_buf = (int16_t *)frame->data[0];
  91. bitstream_init8(&bc, avpkt->data, avpkt->size);
  92. for (j = 0; j < avpkt->size; j++) {
  93. int ilow, ihigh, rlow, rhigh, dhigh;
  94. int xout[2];
  95. ihigh = bitstream_read(&bc, 2);
  96. ilow = bitstream_read(&bc, 6 - skip);
  97. bitstream_skip(&bc, skip);
  98. rlow = av_clip_intp2((c->band[0].scale_factor * quantizer_table[ilow] >> 10)
  99. + c->band[0].s_predictor, 14);
  100. ff_g722_update_low_predictor(&c->band[0], ilow >> (2 - skip));
  101. dhigh = c->band[1].scale_factor * ff_g722_high_inv_quant[ihigh] >> 10;
  102. rhigh = av_clip_intp2(dhigh + c->band[1].s_predictor, 14);
  103. ff_g722_update_high_predictor(&c->band[1], dhigh, ihigh);
  104. c->prev_samples[c->prev_samples_pos++] = rlow + rhigh;
  105. c->prev_samples[c->prev_samples_pos++] = rlow - rhigh;
  106. c->dsp.apply_qmf(c->prev_samples + c->prev_samples_pos - 24, xout);
  107. *out_buf++ = av_clip_int16(xout[0] >> 11);
  108. *out_buf++ = av_clip_int16(xout[1] >> 11);
  109. if (c->prev_samples_pos >= PREV_SAMPLES_BUF_SIZE) {
  110. memmove(c->prev_samples, c->prev_samples + c->prev_samples_pos - 22,
  111. 22 * sizeof(c->prev_samples[0]));
  112. c->prev_samples_pos = 22;
  113. }
  114. }
  115. *got_frame_ptr = 1;
  116. return avpkt->size;
  117. }
  118. AVCodec ff_adpcm_g722_decoder = {
  119. .name = "g722",
  120. .long_name = NULL_IF_CONFIG_SMALL("G.722 ADPCM"),
  121. .type = AVMEDIA_TYPE_AUDIO,
  122. .id = AV_CODEC_ID_ADPCM_G722,
  123. .priv_data_size = sizeof(G722Context),
  124. .init = g722_decode_init,
  125. .decode = g722_decode_frame,
  126. .capabilities = AV_CODEC_CAP_DR1,
  127. .priv_class = &g722_decoder_class,
  128. };