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.

313 lines
9.6KB

  1. /*
  2. * Interface to libmp3lame for mp3 encoding
  3. * Copyright (c) 2002 Lennert Buytenhek <buytenh@gnu.org>
  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. * @file
  23. * Interface to libmp3lame for mp3 encoding.
  24. */
  25. #include "libavutil/intreadwrite.h"
  26. #include "libavutil/log.h"
  27. #include "libavutil/opt.h"
  28. #include "avcodec.h"
  29. #include "mpegaudio.h"
  30. #include <lame/lame.h>
  31. #define BUFFER_SIZE (7200 + 2 * MPA_FRAME_SIZE + MPA_FRAME_SIZE / 4+1000) // FIXME: Buffer size to small? Adding 1000 to make up for it.
  32. typedef struct Mp3AudioContext {
  33. AVClass *class;
  34. lame_global_flags *gfp;
  35. int stereo;
  36. uint8_t buffer[BUFFER_SIZE];
  37. int buffer_index;
  38. struct {
  39. int *left;
  40. int *right;
  41. } s32_data;
  42. int reservoir;
  43. } Mp3AudioContext;
  44. static av_cold int MP3lame_encode_init(AVCodecContext *avctx)
  45. {
  46. Mp3AudioContext *s = avctx->priv_data;
  47. if (avctx->channels > 2) {
  48. av_log(avctx, AV_LOG_ERROR,
  49. "Invalid number of channels %d, must be <= 2\n", avctx->channels);
  50. return AVERROR(EINVAL);
  51. }
  52. s->stereo = avctx->channels > 1 ? 1 : 0;
  53. if ((s->gfp = lame_init()) == NULL)
  54. goto err;
  55. lame_set_in_samplerate(s->gfp, avctx->sample_rate);
  56. lame_set_out_samplerate(s->gfp, avctx->sample_rate);
  57. lame_set_num_channels(s->gfp, avctx->channels);
  58. if (avctx->compression_level == FF_COMPRESSION_DEFAULT) {
  59. lame_set_quality(s->gfp, 5);
  60. } else {
  61. lame_set_quality(s->gfp, avctx->compression_level);
  62. }
  63. lame_set_mode(s->gfp, s->stereo ? JOINT_STEREO : MONO);
  64. lame_set_brate(s->gfp, avctx->bit_rate / 1000);
  65. if (avctx->flags & CODEC_FLAG_QSCALE) {
  66. lame_set_brate(s->gfp, 0);
  67. lame_set_VBR(s->gfp, vbr_default);
  68. lame_set_VBR_quality(s->gfp, avctx->global_quality / (float)FF_QP2LAMBDA);
  69. }
  70. lame_set_bWriteVbrTag(s->gfp,0);
  71. lame_set_disable_reservoir(s->gfp, !s->reservoir);
  72. if (lame_init_params(s->gfp) < 0)
  73. goto err_close;
  74. avctx->frame_size = lame_get_framesize(s->gfp);
  75. if(!(avctx->coded_frame= avcodec_alloc_frame())) {
  76. lame_close(s->gfp);
  77. return AVERROR(ENOMEM);
  78. }
  79. if(AV_SAMPLE_FMT_S32 == avctx->sample_fmt && s->stereo) {
  80. int nelem = 2 * avctx->frame_size;
  81. if(! (s->s32_data.left = av_malloc(nelem * sizeof(int)))) {
  82. av_freep(&avctx->coded_frame);
  83. lame_close(s->gfp);
  84. return AVERROR(ENOMEM);
  85. }
  86. s->s32_data.right = s->s32_data.left + avctx->frame_size;
  87. }
  88. return 0;
  89. err_close:
  90. lame_close(s->gfp);
  91. err:
  92. return -1;
  93. }
  94. static const int sSampleRates[] = {
  95. 44100, 48000, 32000, 22050, 24000, 16000, 11025, 12000, 8000, 0
  96. };
  97. static const int sBitRates[2][3][15] = {
  98. {
  99. { 0, 32, 64, 96, 128, 160, 192, 224, 256, 288, 320, 352, 384, 416, 448 },
  100. { 0, 32, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320, 384 },
  101. { 0, 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320 }
  102. },
  103. {
  104. { 0, 32, 48, 56, 64, 80, 96, 112, 128, 144, 160, 176, 192, 224, 256 },
  105. { 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160 },
  106. { 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160 }
  107. },
  108. };
  109. static const int sSamplesPerFrame[2][3] = {
  110. { 384, 1152, 1152 },
  111. { 384, 1152, 576 }
  112. };
  113. static const int sBitsPerSlot[3] = { 32, 8, 8 };
  114. static int mp3len(void *data, int *samplesPerFrame, int *sampleRate)
  115. {
  116. uint32_t header = AV_RB32(data);
  117. int layerID = 3 - ((header >> 17) & 0x03);
  118. int bitRateID = ((header >> 12) & 0x0f);
  119. int sampleRateID = ((header >> 10) & 0x03);
  120. int bitsPerSlot = sBitsPerSlot[layerID];
  121. int isPadded = ((header >> 9) & 0x01);
  122. static int const mode_tab[4] = { 2, 3, 1, 0 };
  123. int mode = mode_tab[(header >> 19) & 0x03];
  124. int mpeg_id = mode > 0;
  125. int temp0, temp1, bitRate;
  126. if (((header >> 21) & 0x7ff) != 0x7ff || mode == 3 || layerID == 3 ||
  127. sampleRateID == 3) {
  128. return -1;
  129. }
  130. if (!samplesPerFrame)
  131. samplesPerFrame = &temp0;
  132. if (!sampleRate)
  133. sampleRate = &temp1;
  134. //*isMono = ((header >> 6) & 0x03) == 0x03;
  135. *sampleRate = sSampleRates[sampleRateID] >> mode;
  136. bitRate = sBitRates[mpeg_id][layerID][bitRateID] * 1000;
  137. *samplesPerFrame = sSamplesPerFrame[mpeg_id][layerID];
  138. //av_log(NULL, AV_LOG_DEBUG,
  139. // "sr:%d br:%d spf:%d l:%d m:%d\n",
  140. // *sampleRate, bitRate, *samplesPerFrame, layerID, mode);
  141. return *samplesPerFrame * bitRate / (bitsPerSlot * *sampleRate) + isPadded;
  142. }
  143. static int MP3lame_encode_frame(AVCodecContext *avctx, unsigned char *frame,
  144. int buf_size, void *data)
  145. {
  146. Mp3AudioContext *s = avctx->priv_data;
  147. int len;
  148. int lame_result;
  149. /* lame 3.91 dies on '1-channel interleaved' data */
  150. if (!data){
  151. lame_result= lame_encode_flush(
  152. s->gfp,
  153. s->buffer + s->buffer_index,
  154. BUFFER_SIZE - s->buffer_index
  155. );
  156. #if 2147483647 == INT_MAX
  157. }else if(AV_SAMPLE_FMT_S32 == avctx->sample_fmt){
  158. if (s->stereo) {
  159. int32_t *rp = data;
  160. int32_t *mp = rp + 2*avctx->frame_size;
  161. int *wpl = s->s32_data.left;
  162. int *wpr = s->s32_data.right;
  163. while (rp < mp) {
  164. *wpl++ = *rp++;
  165. *wpr++ = *rp++;
  166. }
  167. lame_result = lame_encode_buffer_int(
  168. s->gfp,
  169. s->s32_data.left,
  170. s->s32_data.right,
  171. avctx->frame_size,
  172. s->buffer + s->buffer_index,
  173. BUFFER_SIZE - s->buffer_index
  174. );
  175. } else {
  176. lame_result = lame_encode_buffer_int(
  177. s->gfp,
  178. data,
  179. data,
  180. avctx->frame_size,
  181. s->buffer + s->buffer_index,
  182. BUFFER_SIZE - s->buffer_index
  183. );
  184. }
  185. #endif
  186. }else{
  187. if (s->stereo) {
  188. lame_result = lame_encode_buffer_interleaved(
  189. s->gfp,
  190. data,
  191. avctx->frame_size,
  192. s->buffer + s->buffer_index,
  193. BUFFER_SIZE - s->buffer_index
  194. );
  195. } else {
  196. lame_result = lame_encode_buffer(
  197. s->gfp,
  198. data,
  199. data,
  200. avctx->frame_size,
  201. s->buffer + s->buffer_index,
  202. BUFFER_SIZE - s->buffer_index
  203. );
  204. }
  205. }
  206. if (lame_result < 0) {
  207. if (lame_result == -1) {
  208. /* output buffer too small */
  209. av_log(avctx, AV_LOG_ERROR,
  210. "lame: output buffer too small (buffer index: %d, free bytes: %d)\n",
  211. s->buffer_index, BUFFER_SIZE - s->buffer_index);
  212. }
  213. return -1;
  214. }
  215. s->buffer_index += lame_result;
  216. if (s->buffer_index < 4)
  217. return 0;
  218. len = mp3len(s->buffer, NULL, NULL);
  219. //av_log(avctx, AV_LOG_DEBUG, "in:%d packet-len:%d index:%d\n",
  220. // avctx->frame_size, len, s->buffer_index);
  221. if (len <= s->buffer_index) {
  222. memcpy(frame, s->buffer, len);
  223. s->buffer_index -= len;
  224. memmove(s->buffer, s->buffer + len, s->buffer_index);
  225. // FIXME fix the audio codec API, so we do not need the memcpy()
  226. /*for(i=0; i<len; i++) {
  227. av_log(avctx, AV_LOG_DEBUG, "%2X ", frame[i]);
  228. }*/
  229. return len;
  230. } else
  231. return 0;
  232. }
  233. static av_cold int MP3lame_encode_close(AVCodecContext *avctx)
  234. {
  235. Mp3AudioContext *s = avctx->priv_data;
  236. av_freep(&s->s32_data.left);
  237. av_freep(&avctx->coded_frame);
  238. lame_close(s->gfp);
  239. return 0;
  240. }
  241. #define OFFSET(x) offsetof(Mp3AudioContext, x)
  242. #define AE AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
  243. static const AVOption options[] = {
  244. { "reservoir", "Use bit reservoir.", OFFSET(reservoir), AV_OPT_TYPE_INT, { 1 }, 0, 1, AE },
  245. { NULL },
  246. };
  247. static const AVClass libmp3lame_class = {
  248. .class_name = "libmp3lame encoder",
  249. .item_name = av_default_item_name,
  250. .option = options,
  251. .version = LIBAVUTIL_VERSION_INT,
  252. };
  253. AVCodec ff_libmp3lame_encoder = {
  254. .name = "libmp3lame",
  255. .type = AVMEDIA_TYPE_AUDIO,
  256. .id = CODEC_ID_MP3,
  257. .priv_data_size = sizeof(Mp3AudioContext),
  258. .init = MP3lame_encode_init,
  259. .encode = MP3lame_encode_frame,
  260. .close = MP3lame_encode_close,
  261. .capabilities = CODEC_CAP_DELAY,
  262. .sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16,
  263. #if 2147483647 == INT_MAX
  264. AV_SAMPLE_FMT_S32,
  265. #endif
  266. AV_SAMPLE_FMT_NONE },
  267. .supported_samplerates = sSampleRates,
  268. .long_name = NULL_IF_CONFIG_SMALL("libmp3lame MP3 (MPEG audio layer 3)"),
  269. .priv_class = &libmp3lame_class,
  270. };