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.

317 lines
9.8KB

  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. #if FF_API_LAME_GLOBAL_OPTS
  72. s->reservoir = avctx->flags2 & CODEC_FLAG2_BIT_RESERVOIR;
  73. #endif
  74. lame_set_disable_reservoir(s->gfp, !s->reservoir);
  75. if (lame_init_params(s->gfp) < 0)
  76. goto err_close;
  77. avctx->frame_size = lame_get_framesize(s->gfp);
  78. if(!(avctx->coded_frame= avcodec_alloc_frame())) {
  79. lame_close(s->gfp);
  80. return AVERROR(ENOMEM);
  81. }
  82. avctx->coded_frame->key_frame = 1;
  83. if(AV_SAMPLE_FMT_S32 == avctx->sample_fmt && s->stereo) {
  84. int nelem = 2 * avctx->frame_size;
  85. if(! (s->s32_data.left = av_malloc(nelem * sizeof(int)))) {
  86. av_freep(&avctx->coded_frame);
  87. lame_close(s->gfp);
  88. return AVERROR(ENOMEM);
  89. }
  90. s->s32_data.right = s->s32_data.left + avctx->frame_size;
  91. }
  92. return 0;
  93. err_close:
  94. lame_close(s->gfp);
  95. err:
  96. return -1;
  97. }
  98. static const int sSampleRates[] = {
  99. 44100, 48000, 32000, 22050, 24000, 16000, 11025, 12000, 8000, 0
  100. };
  101. static const int sBitRates[2][3][15] = {
  102. {
  103. { 0, 32, 64, 96, 128, 160, 192, 224, 256, 288, 320, 352, 384, 416, 448 },
  104. { 0, 32, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320, 384 },
  105. { 0, 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320 }
  106. },
  107. {
  108. { 0, 32, 48, 56, 64, 80, 96, 112, 128, 144, 160, 176, 192, 224, 256 },
  109. { 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160 },
  110. { 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160 }
  111. },
  112. };
  113. static const int sSamplesPerFrame[2][3] = {
  114. { 384, 1152, 1152 },
  115. { 384, 1152, 576 }
  116. };
  117. static const int sBitsPerSlot[3] = { 32, 8, 8 };
  118. static int mp3len(void *data, int *samplesPerFrame, int *sampleRate)
  119. {
  120. uint32_t header = AV_RB32(data);
  121. int layerID = 3 - ((header >> 17) & 0x03);
  122. int bitRateID = ((header >> 12) & 0x0f);
  123. int sampleRateID = ((header >> 10) & 0x03);
  124. int bitsPerSlot = sBitsPerSlot[layerID];
  125. int isPadded = ((header >> 9) & 0x01);
  126. static int const mode_tab[4] = { 2, 3, 1, 0 };
  127. int mode = mode_tab[(header >> 19) & 0x03];
  128. int mpeg_id = mode > 0;
  129. int temp0, temp1, bitRate;
  130. if (((header >> 21) & 0x7ff) != 0x7ff || mode == 3 || layerID == 3 ||
  131. sampleRateID == 3) {
  132. return -1;
  133. }
  134. if (!samplesPerFrame)
  135. samplesPerFrame = &temp0;
  136. if (!sampleRate)
  137. sampleRate = &temp1;
  138. //*isMono = ((header >> 6) & 0x03) == 0x03;
  139. *sampleRate = sSampleRates[sampleRateID] >> mode;
  140. bitRate = sBitRates[mpeg_id][layerID][bitRateID] * 1000;
  141. *samplesPerFrame = sSamplesPerFrame[mpeg_id][layerID];
  142. //av_log(NULL, AV_LOG_DEBUG,
  143. // "sr:%d br:%d spf:%d l:%d m:%d\n",
  144. // *sampleRate, bitRate, *samplesPerFrame, layerID, mode);
  145. return *samplesPerFrame * bitRate / (bitsPerSlot * *sampleRate) + isPadded;
  146. }
  147. static int MP3lame_encode_frame(AVCodecContext *avctx, unsigned char *frame,
  148. int buf_size, void *data)
  149. {
  150. Mp3AudioContext *s = avctx->priv_data;
  151. int len;
  152. int lame_result;
  153. /* lame 3.91 dies on '1-channel interleaved' data */
  154. if (!data){
  155. lame_result= lame_encode_flush(
  156. s->gfp,
  157. s->buffer + s->buffer_index,
  158. BUFFER_SIZE - s->buffer_index
  159. );
  160. #if 2147483647 == INT_MAX
  161. }else if(AV_SAMPLE_FMT_S32 == avctx->sample_fmt){
  162. if (s->stereo) {
  163. int32_t *rp = data;
  164. int32_t *mp = rp + 2*avctx->frame_size;
  165. int *wpl = s->s32_data.left;
  166. int *wpr = s->s32_data.right;
  167. while (rp < mp) {
  168. *wpl++ = *rp++;
  169. *wpr++ = *rp++;
  170. }
  171. lame_result = lame_encode_buffer_int(
  172. s->gfp,
  173. s->s32_data.left,
  174. s->s32_data.right,
  175. avctx->frame_size,
  176. s->buffer + s->buffer_index,
  177. BUFFER_SIZE - s->buffer_index
  178. );
  179. } else {
  180. lame_result = lame_encode_buffer_int(
  181. s->gfp,
  182. data,
  183. data,
  184. avctx->frame_size,
  185. s->buffer + s->buffer_index,
  186. BUFFER_SIZE - s->buffer_index
  187. );
  188. }
  189. #endif
  190. }else{
  191. if (s->stereo) {
  192. lame_result = lame_encode_buffer_interleaved(
  193. s->gfp,
  194. data,
  195. avctx->frame_size,
  196. s->buffer + s->buffer_index,
  197. BUFFER_SIZE - s->buffer_index
  198. );
  199. } else {
  200. lame_result = lame_encode_buffer(
  201. s->gfp,
  202. data,
  203. data,
  204. avctx->frame_size,
  205. s->buffer + s->buffer_index,
  206. BUFFER_SIZE - s->buffer_index
  207. );
  208. }
  209. }
  210. if (lame_result < 0) {
  211. if (lame_result == -1) {
  212. /* output buffer too small */
  213. av_log(avctx, AV_LOG_ERROR,
  214. "lame: output buffer too small (buffer index: %d, free bytes: %d)\n",
  215. s->buffer_index, BUFFER_SIZE - s->buffer_index);
  216. }
  217. return -1;
  218. }
  219. s->buffer_index += lame_result;
  220. if (s->buffer_index < 4)
  221. return 0;
  222. len = mp3len(s->buffer, NULL, NULL);
  223. //av_log(avctx, AV_LOG_DEBUG, "in:%d packet-len:%d index:%d\n",
  224. // avctx->frame_size, len, s->buffer_index);
  225. if (len <= s->buffer_index) {
  226. memcpy(frame, s->buffer, len);
  227. s->buffer_index -= len;
  228. memmove(s->buffer, s->buffer + len, s->buffer_index);
  229. // FIXME fix the audio codec API, so we do not need the memcpy()
  230. /*for(i=0; i<len; i++) {
  231. av_log(avctx, AV_LOG_DEBUG, "%2X ", frame[i]);
  232. }*/
  233. return len;
  234. } else
  235. return 0;
  236. }
  237. static av_cold int MP3lame_encode_close(AVCodecContext *avctx)
  238. {
  239. Mp3AudioContext *s = avctx->priv_data;
  240. av_freep(&s->s32_data.left);
  241. av_freep(&avctx->coded_frame);
  242. lame_close(s->gfp);
  243. return 0;
  244. }
  245. #define OFFSET(x) offsetof(Mp3AudioContext, x)
  246. #define AE AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
  247. static const AVOption options[] = {
  248. { "reservoir", "Use bit reservoir.", OFFSET(reservoir), AV_OPT_TYPE_INT, { 1 }, 0, 1, AE },
  249. { NULL },
  250. };
  251. static const AVClass libmp3lame_class = {
  252. .class_name = "libmp3lame encoder",
  253. .item_name = av_default_item_name,
  254. .option = options,
  255. .version = LIBAVUTIL_VERSION_INT,
  256. };
  257. AVCodec ff_libmp3lame_encoder = {
  258. .name = "libmp3lame",
  259. .type = AVMEDIA_TYPE_AUDIO,
  260. .id = CODEC_ID_MP3,
  261. .priv_data_size = sizeof(Mp3AudioContext),
  262. .init = MP3lame_encode_init,
  263. .encode = MP3lame_encode_frame,
  264. .close = MP3lame_encode_close,
  265. .capabilities = CODEC_CAP_DELAY,
  266. .sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16,
  267. #if 2147483647 == INT_MAX
  268. AV_SAMPLE_FMT_S32,
  269. #endif
  270. AV_SAMPLE_FMT_NONE },
  271. .supported_samplerates = sSampleRates,
  272. .long_name = NULL_IF_CONFIG_SMALL("libmp3lame MP3 (MPEG audio layer 3)"),
  273. .priv_class = &libmp3lame_class,
  274. };