Go to most recent revision | Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 4349 | Serge | 1 | /* |
| 2 | * Interface to libmp3lame for mp3 encoding |
||
| 3 | * Copyright (c) 2002 Lennert Buytenhek |
||
| 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 | /** |
||
| 23 | * @file |
||
| 24 | * Interface to libmp3lame for mp3 encoding. |
||
| 25 | */ |
||
| 26 | |||
| 27 | #include |
||
| 28 | |||
| 29 | #include "libavutil/channel_layout.h" |
||
| 30 | #include "libavutil/common.h" |
||
| 31 | #include "libavutil/float_dsp.h" |
||
| 32 | #include "libavutil/intreadwrite.h" |
||
| 33 | #include "libavutil/log.h" |
||
| 34 | #include "libavutil/opt.h" |
||
| 35 | #include "avcodec.h" |
||
| 36 | #include "audio_frame_queue.h" |
||
| 37 | #include "internal.h" |
||
| 38 | #include "mpegaudio.h" |
||
| 39 | #include "mpegaudiodecheader.h" |
||
| 40 | |||
| 41 | #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. |
||
| 42 | |||
| 43 | typedef struct LAMEContext { |
||
| 44 | AVClass *class; |
||
| 45 | AVCodecContext *avctx; |
||
| 46 | lame_global_flags *gfp; |
||
| 47 | uint8_t *buffer; |
||
| 48 | int buffer_index; |
||
| 49 | int buffer_size; |
||
| 50 | int reservoir; |
||
| 51 | int joint_stereo; |
||
| 52 | float *samples_flt[2]; |
||
| 53 | AudioFrameQueue afq; |
||
| 54 | AVFloatDSPContext fdsp; |
||
| 55 | } LAMEContext; |
||
| 56 | |||
| 57 | |||
| 58 | static int realloc_buffer(LAMEContext *s) |
||
| 59 | { |
||
| 60 | if (!s->buffer || s->buffer_size - s->buffer_index < BUFFER_SIZE) { |
||
| 61 | uint8_t *tmp; |
||
| 62 | int new_size = s->buffer_index + 2 * BUFFER_SIZE; |
||
| 63 | |||
| 64 | av_dlog(s->avctx, "resizing output buffer: %d -> %d\n", s->buffer_size, |
||
| 65 | new_size); |
||
| 66 | tmp = av_realloc(s->buffer, new_size); |
||
| 67 | if (!tmp) { |
||
| 68 | av_freep(&s->buffer); |
||
| 69 | s->buffer_size = s->buffer_index = 0; |
||
| 70 | return AVERROR(ENOMEM); |
||
| 71 | } |
||
| 72 | s->buffer = tmp; |
||
| 73 | s->buffer_size = new_size; |
||
| 74 | } |
||
| 75 | return 0; |
||
| 76 | } |
||
| 77 | |||
| 78 | static av_cold int mp3lame_encode_close(AVCodecContext *avctx) |
||
| 79 | { |
||
| 80 | LAMEContext *s = avctx->priv_data; |
||
| 81 | |||
| 82 | av_freep(&s->samples_flt[0]); |
||
| 83 | av_freep(&s->samples_flt[1]); |
||
| 84 | av_freep(&s->buffer); |
||
| 85 | |||
| 86 | ff_af_queue_close(&s->afq); |
||
| 87 | |||
| 88 | lame_close(s->gfp); |
||
| 89 | return 0; |
||
| 90 | } |
||
| 91 | |||
| 92 | static av_cold int mp3lame_encode_init(AVCodecContext *avctx) |
||
| 93 | { |
||
| 94 | LAMEContext *s = avctx->priv_data; |
||
| 95 | int ret; |
||
| 96 | |||
| 97 | s->avctx = avctx; |
||
| 98 | |||
| 99 | /* initialize LAME and get defaults */ |
||
| 100 | if ((s->gfp = lame_init()) == NULL) |
||
| 101 | return AVERROR(ENOMEM); |
||
| 102 | |||
| 103 | |||
| 104 | lame_set_num_channels(s->gfp, avctx->channels); |
||
| 105 | lame_set_mode(s->gfp, avctx->channels > 1 ? s->joint_stereo ? JOINT_STEREO : STEREO : MONO); |
||
| 106 | |||
| 107 | /* sample rate */ |
||
| 108 | lame_set_in_samplerate (s->gfp, avctx->sample_rate); |
||
| 109 | lame_set_out_samplerate(s->gfp, avctx->sample_rate); |
||
| 110 | |||
| 111 | /* algorithmic quality */ |
||
| 112 | if (avctx->compression_level == FF_COMPRESSION_DEFAULT) |
||
| 113 | lame_set_quality(s->gfp, 5); |
||
| 114 | else |
||
| 115 | lame_set_quality(s->gfp, avctx->compression_level); |
||
| 116 | |||
| 117 | /* rate control */ |
||
| 118 | if (avctx->flags & CODEC_FLAG_QSCALE) { // VBR |
||
| 119 | lame_set_VBR(s->gfp, vbr_default); |
||
| 120 | lame_set_VBR_quality(s->gfp, avctx->global_quality / (float)FF_QP2LAMBDA); |
||
| 121 | } else { |
||
| 122 | if (avctx->bit_rate) // CBR |
||
| 123 | lame_set_brate(s->gfp, avctx->bit_rate / 1000); |
||
| 124 | } |
||
| 125 | |||
| 126 | /* do not get a Xing VBR header frame from LAME */ |
||
| 127 | lame_set_bWriteVbrTag(s->gfp,0); |
||
| 128 | |||
| 129 | /* bit reservoir usage */ |
||
| 130 | lame_set_disable_reservoir(s->gfp, !s->reservoir); |
||
| 131 | |||
| 132 | /* set specified parameters */ |
||
| 133 | if (lame_init_params(s->gfp) < 0) { |
||
| 134 | ret = -1; |
||
| 135 | goto error; |
||
| 136 | } |
||
| 137 | |||
| 138 | /* get encoder delay */ |
||
| 139 | avctx->delay = lame_get_encoder_delay(s->gfp) + 528 + 1; |
||
| 140 | ff_af_queue_init(avctx, &s->afq); |
||
| 141 | |||
| 142 | avctx->frame_size = lame_get_framesize(s->gfp); |
||
| 143 | |||
| 144 | /* allocate float sample buffers */ |
||
| 145 | if (avctx->sample_fmt == AV_SAMPLE_FMT_FLTP) { |
||
| 146 | int ch; |
||
| 147 | for (ch = 0; ch < avctx->channels; ch++) { |
||
| 148 | s->samples_flt[ch] = av_malloc(avctx->frame_size * |
||
| 149 | sizeof(*s->samples_flt[ch])); |
||
| 150 | if (!s->samples_flt[ch]) { |
||
| 151 | ret = AVERROR(ENOMEM); |
||
| 152 | goto error; |
||
| 153 | } |
||
| 154 | } |
||
| 155 | } |
||
| 156 | |||
| 157 | ret = realloc_buffer(s); |
||
| 158 | if (ret < 0) |
||
| 159 | goto error; |
||
| 160 | |||
| 161 | avpriv_float_dsp_init(&s->fdsp, avctx->flags & CODEC_FLAG_BITEXACT); |
||
| 162 | |||
| 163 | return 0; |
||
| 164 | error: |
||
| 165 | mp3lame_encode_close(avctx); |
||
| 166 | return ret; |
||
| 167 | } |
||
| 168 | |||
| 169 | #define ENCODE_BUFFER(func, buf_type, buf_name) do { \ |
||
| 170 | lame_result = func(s->gfp, \ |
||
| 171 | (const buf_type *)buf_name[0], \ |
||
| 172 | (const buf_type *)buf_name[1], frame->nb_samples, \ |
||
| 173 | s->buffer + s->buffer_index, \ |
||
| 174 | s->buffer_size - s->buffer_index); \ |
||
| 175 | } while (0) |
||
| 176 | |||
| 177 | static int mp3lame_encode_frame(AVCodecContext *avctx, AVPacket *avpkt, |
||
| 178 | const AVFrame *frame, int *got_packet_ptr) |
||
| 179 | { |
||
| 180 | LAMEContext *s = avctx->priv_data; |
||
| 181 | MPADecodeHeader hdr; |
||
| 182 | int len, ret, ch; |
||
| 183 | int lame_result; |
||
| 184 | |||
| 185 | if (frame) { |
||
| 186 | switch (avctx->sample_fmt) { |
||
| 187 | case AV_SAMPLE_FMT_S16P: |
||
| 188 | ENCODE_BUFFER(lame_encode_buffer, int16_t, frame->data); |
||
| 189 | break; |
||
| 190 | case AV_SAMPLE_FMT_S32P: |
||
| 191 | ENCODE_BUFFER(lame_encode_buffer_int, int32_t, frame->data); |
||
| 192 | break; |
||
| 193 | case AV_SAMPLE_FMT_FLTP: |
||
| 194 | if (frame->linesize[0] < 4 * FFALIGN(frame->nb_samples, 8)) { |
||
| 195 | av_log(avctx, AV_LOG_ERROR, "inadequate AVFrame plane padding\n"); |
||
| 196 | return AVERROR(EINVAL); |
||
| 197 | } |
||
| 198 | for (ch = 0; ch < avctx->channels; ch++) { |
||
| 199 | s->fdsp.vector_fmul_scalar(s->samples_flt[ch], |
||
| 200 | (const float *)frame->data[ch], |
||
| 201 | 32768.0f, |
||
| 202 | FFALIGN(frame->nb_samples, 8)); |
||
| 203 | } |
||
| 204 | ENCODE_BUFFER(lame_encode_buffer_float, float, s->samples_flt); |
||
| 205 | break; |
||
| 206 | default: |
||
| 207 | return AVERROR_BUG; |
||
| 208 | } |
||
| 209 | } else { |
||
| 210 | lame_result = lame_encode_flush(s->gfp, s->buffer + s->buffer_index, |
||
| 211 | s->buffer_size - s->buffer_index); |
||
| 212 | } |
||
| 213 | if (lame_result < 0) { |
||
| 214 | if (lame_result == -1) { |
||
| 215 | av_log(avctx, AV_LOG_ERROR, |
||
| 216 | "lame: output buffer too small (buffer index: %d, free bytes: %d)\n", |
||
| 217 | s->buffer_index, s->buffer_size - s->buffer_index); |
||
| 218 | } |
||
| 219 | return -1; |
||
| 220 | } |
||
| 221 | s->buffer_index += lame_result; |
||
| 222 | ret = realloc_buffer(s); |
||
| 223 | if (ret < 0) { |
||
| 224 | av_log(avctx, AV_LOG_ERROR, "error reallocating output buffer\n"); |
||
| 225 | return ret; |
||
| 226 | } |
||
| 227 | |||
| 228 | /* add current frame to the queue */ |
||
| 229 | if (frame) { |
||
| 230 | if ((ret = ff_af_queue_add(&s->afq, frame)) < 0) |
||
| 231 | return ret; |
||
| 232 | } |
||
| 233 | |||
| 234 | /* Move 1 frame from the LAME buffer to the output packet, if available. |
||
| 235 | We have to parse the first frame header in the output buffer to |
||
| 236 | determine the frame size. */ |
||
| 237 | if (s->buffer_index < 4) |
||
| 238 | return 0; |
||
| 239 | if (avpriv_mpegaudio_decode_header(&hdr, AV_RB32(s->buffer))) { |
||
| 240 | av_log(avctx, AV_LOG_ERROR, "free format output not supported\n"); |
||
| 241 | return -1; |
||
| 242 | } |
||
| 243 | len = hdr.frame_size; |
||
| 244 | av_dlog(avctx, "in:%d packet-len:%d index:%d\n", avctx->frame_size, len, |
||
| 245 | s->buffer_index); |
||
| 246 | if (len <= s->buffer_index) { |
||
| 247 | if ((ret = ff_alloc_packet2(avctx, avpkt, len)) < 0) |
||
| 248 | return ret; |
||
| 249 | memcpy(avpkt->data, s->buffer, len); |
||
| 250 | s->buffer_index -= len; |
||
| 251 | memmove(s->buffer, s->buffer + len, s->buffer_index); |
||
| 252 | |||
| 253 | /* Get the next frame pts/duration */ |
||
| 254 | ff_af_queue_remove(&s->afq, avctx->frame_size, &avpkt->pts, |
||
| 255 | &avpkt->duration); |
||
| 256 | |||
| 257 | avpkt->size = len; |
||
| 258 | *got_packet_ptr = 1; |
||
| 259 | } |
||
| 260 | return 0; |
||
| 261 | } |
||
| 262 | |||
| 263 | #define OFFSET(x) offsetof(LAMEContext, x) |
||
| 264 | #define AE AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM |
||
| 265 | static const AVOption options[] = { |
||
| 266 | { "reservoir", "use bit reservoir", OFFSET(reservoir), AV_OPT_TYPE_INT, { .i64 = 1 }, 0, 1, AE }, |
||
| 267 | { "joint_stereo", "use joint stereo", OFFSET(joint_stereo), AV_OPT_TYPE_INT, { .i64 = 1 }, 0, 1, AE }, |
||
| 268 | { NULL }, |
||
| 269 | }; |
||
| 270 | |||
| 271 | static const AVClass libmp3lame_class = { |
||
| 272 | .class_name = "libmp3lame encoder", |
||
| 273 | .item_name = av_default_item_name, |
||
| 274 | .option = options, |
||
| 275 | .version = LIBAVUTIL_VERSION_INT, |
||
| 276 | }; |
||
| 277 | |||
| 278 | static const AVCodecDefault libmp3lame_defaults[] = { |
||
| 279 | { "b", "0" }, |
||
| 280 | { NULL }, |
||
| 281 | }; |
||
| 282 | |||
| 283 | static const int libmp3lame_sample_rates[] = { |
||
| 284 | 44100, 48000, 32000, 22050, 24000, 16000, 11025, 12000, 8000, 0 |
||
| 285 | }; |
||
| 286 | |||
| 287 | AVCodec ff_libmp3lame_encoder = { |
||
| 288 | .name = "libmp3lame", |
||
| 289 | .long_name = NULL_IF_CONFIG_SMALL("libmp3lame MP3 (MPEG audio layer 3)"), |
||
| 290 | .type = AVMEDIA_TYPE_AUDIO, |
||
| 291 | .id = AV_CODEC_ID_MP3, |
||
| 292 | .priv_data_size = sizeof(LAMEContext), |
||
| 293 | .init = mp3lame_encode_init, |
||
| 294 | .encode2 = mp3lame_encode_frame, |
||
| 295 | .close = mp3lame_encode_close, |
||
| 296 | .capabilities = CODEC_CAP_DELAY | CODEC_CAP_SMALL_LAST_FRAME, |
||
| 297 | .sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S32P, |
||
| 298 | AV_SAMPLE_FMT_FLTP, |
||
| 299 | AV_SAMPLE_FMT_S16P, |
||
| 300 | AV_SAMPLE_FMT_NONE }, |
||
| 301 | .supported_samplerates = libmp3lame_sample_rates, |
||
| 302 | .channel_layouts = (const uint64_t[]) { AV_CH_LAYOUT_MONO, |
||
| 303 | AV_CH_LAYOUT_STEREO, |
||
| 304 | |||
| 305 | .priv_class = &libmp3lame_class, |
||
| 306 | .defaults = libmp3lame_defaults, |
||
| 307 | }; |