libmp3lame.c 7.97 KB
Newer Older
1 2 3 4
/*
 * Interface to libmp3lame for mp3 encoding
 * Copyright (c) 2002 Lennert Buytenhek <buytenh@gnu.org>
 *
5
 * This file is part of Libav.
6
 *
7
 * Libav is free software; you can redistribute it and/or
Fabrice Bellard's avatar
Fabrice Bellard committed
8 9
 * modify it under the terms of the GNU Lesser General Public
 * 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
 * Libav is distributed in the hope that it will be useful,
13
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Fabrice Bellard's avatar
Fabrice Bellard committed
14 15
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
16
 *
Fabrice Bellard's avatar
Fabrice Bellard committed
17
 * You should have received a copy of the GNU Lesser General Public
18
 * License along with Libav; if not, write to the Free Software
19
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20
 */
21

Michael Niedermayer's avatar
Michael Niedermayer committed
22
/**
23
 * @file
Michael Niedermayer's avatar
Michael Niedermayer committed
24 25
 * Interface to libmp3lame for mp3 encoding.
 */
26

27
#include "libavutil/intreadwrite.h"
28 29
#include "libavutil/log.h"
#include "libavutil/opt.h"
30 31 32 33
#include "avcodec.h"
#include "mpegaudio.h"
#include <lame/lame.h>

34
#define BUFFER_SIZE (7200 + 2 * MPA_FRAME_SIZE + MPA_FRAME_SIZE / 4)
35
typedef struct Mp3AudioContext {
36
    AVClass *class;
37 38 39
    lame_global_flags *gfp;
    uint8_t buffer[BUFFER_SIZE];
    int buffer_index;
40
    int reservoir;
41 42
} Mp3AudioContext;

43
static av_cold int MP3lame_encode_init(AVCodecContext *avctx)
44
{
45 46 47 48 49 50 51 52 53 54
    Mp3AudioContext *s = avctx->priv_data;

    if (avctx->channels > 2)
        return -1;

    if ((s->gfp = lame_init()) == NULL)
        goto err;
    lame_set_in_samplerate(s->gfp, avctx->sample_rate);
    lame_set_out_samplerate(s->gfp, avctx->sample_rate);
    lame_set_num_channels(s->gfp, avctx->channels);
55
    if (avctx->compression_level == FF_COMPRESSION_DEFAULT) {
56 57 58 59
        lame_set_quality(s->gfp, 5);
    } else {
        lame_set_quality(s->gfp, avctx->compression_level);
    }
60
    lame_set_mode(s->gfp, avctx->channels > 1 ? JOINT_STEREO : MONO);
61 62
    lame_set_brate(s->gfp, avctx->bit_rate / 1000);
    if (avctx->flags & CODEC_FLAG_QSCALE) {
63 64
        lame_set_brate(s->gfp, 0);
        lame_set_VBR(s->gfp, vbr_default);
65
        lame_set_VBR_quality(s->gfp, avctx->global_quality / (float)FF_QP2LAMBDA);
66
    }
67
    lame_set_bWriteVbrTag(s->gfp,0);
68
    lame_set_disable_reservoir(s->gfp, !s->reservoir);
69 70
    if (lame_init_params(s->gfp) < 0)
        goto err_close;
71

72 73 74
    avctx->frame_size             = lame_get_framesize(s->gfp);
    avctx->coded_frame            = avcodec_alloc_frame();
    avctx->coded_frame->key_frame = 1;
75

76
    return 0;
77 78

err_close:
79
    lame_close(s->gfp);
80
err:
81
    return -1;
82 83
}

84 85
static const int sSampleRates[] = {
    44100, 48000,  32000, 22050, 24000, 16000, 11025, 12000, 8000, 0
86 87 88
};

static const int sBitRates[2][3][15] = {
89 90 91 92
    {
        { 0, 32, 64, 96, 128, 160, 192, 224, 256, 288, 320, 352, 384, 416, 448 },
        { 0, 32, 48, 56, 64,  80,  96,  112, 128, 160, 192, 224, 256, 320, 384 },
        { 0, 32, 40, 48, 56,  64,  80,  96,  112, 128, 160, 192, 224, 256, 320 }
93
    },
94 95 96 97
    {
        { 0, 32, 48, 56, 64, 80, 96, 112, 128, 144, 160, 176, 192, 224, 256 },
        { 0,  8, 16, 24, 32, 40, 48,  56,  64,  80,  96, 112, 128, 144, 160 },
        { 0,  8, 16, 24, 32, 40, 48,  56,  64,  80,  96, 112, 128, 144, 160 }
98 99 100
    },
};

101 102 103
static const int sSamplesPerFrame[2][3] = {
    { 384, 1152, 1152 },
    { 384, 1152,  576 }
104 105
};

106
static const int sBitsPerSlot[3] = { 32, 8, 8 };
107 108 109

static int mp3len(void *data, int *samplesPerFrame, int *sampleRate)
{
110 111 112
    uint32_t header  = AV_RB32(data);
    int layerID      = 3 - ((header >> 17) & 0x03);
    int bitRateID    = ((header >> 12) & 0x0f);
113
    int sampleRateID = ((header >> 10) & 0x03);
114 115 116 117 118
    int bitsPerSlot  = sBitsPerSlot[layerID];
    int isPadded     = ((header >> 9) & 0x01);
    static int const mode_tab[4] = { 2, 3, 1, 0 };
    int mode    = mode_tab[(header >> 19) & 0x03];
    int mpeg_id = mode > 0;
119 120
    int temp0, temp1, bitRate;

121 122
    if (((header >> 21) & 0x7ff) != 0x7ff || mode == 3 || layerID == 3 ||
        sampleRateID == 3) {
123 124
        return -1;
    }
125

126 127 128 129
    if (!samplesPerFrame)
        samplesPerFrame = &temp0;
    if (!sampleRate)
        sampleRate      = &temp1;
130

131
    //*isMono = ((header >>  6) & 0x03) == 0x03;
132

133 134
    *sampleRate      = sSampleRates[sampleRateID] >> mode;
    bitRate          = sBitRates[mpeg_id][layerID][bitRateID] * 1000;
135
    *samplesPerFrame = sSamplesPerFrame[mpeg_id][layerID];
136 137 138
    //av_log(NULL, AV_LOG_DEBUG,
    //       "sr:%d br:%d spf:%d l:%d m:%d\n",
    //       *sampleRate, bitRate, *samplesPerFrame, layerID, mode);
139

140 141 142
    return *samplesPerFrame * bitRate / (bitsPerSlot * *sampleRate) + isPadded;
}

143 144
static int MP3lame_encode_frame(AVCodecContext *avctx, unsigned char *frame,
                                int buf_size, void *data)
145
{
146 147 148
    Mp3AudioContext *s = avctx->priv_data;
    int len;
    int lame_result;
149

150
    /* lame 3.91 dies on '1-channel interleaved' data */
151

152
    if (data) {
153
        if (avctx->channels > 1) {
154 155 156 157
            lame_result = lame_encode_buffer_interleaved(s->gfp, data,
                                                         avctx->frame_size,
                                                         s->buffer + s->buffer_index,
                                                         BUFFER_SIZE - s->buffer_index);
158
        } else {
159 160 161 162
            lame_result = lame_encode_buffer(s->gfp, data, data,
                                             avctx->frame_size, s->buffer +
                                             s->buffer_index, BUFFER_SIZE -
                                             s->buffer_index);
163
        }
164 165 166
    } else {
        lame_result = lame_encode_flush(s->gfp, s->buffer + s->buffer_index,
                                        BUFFER_SIZE - s->buffer_index);
167 168
    }

169 170
    if (lame_result < 0) {
        if (lame_result == -1) {
Michael Niedermayer's avatar
Michael Niedermayer committed
171
            /* output buffer too small */
172 173 174
            av_log(avctx, AV_LOG_ERROR,
                   "lame: output buffer too small (buffer index: %d, free bytes: %d)\n",
                   s->buffer_index, BUFFER_SIZE - s->buffer_index);
Michael Niedermayer's avatar
Michael Niedermayer committed
175
        }
176
        return -1;
177 178 179 180
    }

    s->buffer_index += lame_result;

181
    if (s->buffer_index < 4)
182
        return 0;
183

184 185 186 187
    len = mp3len(s->buffer, NULL, NULL);
    //av_log(avctx, AV_LOG_DEBUG, "in:%d packet-len:%d index:%d\n",
    //       avctx->frame_size, len, s->buffer_index);
    if (len <= s->buffer_index) {
Michael Niedermayer's avatar
Michael Niedermayer committed
188 189
        memcpy(frame, s->buffer, len);
        s->buffer_index -= len;
190

191 192 193 194 195
        memmove(s->buffer, s->buffer + len, s->buffer_index);
        // FIXME fix the audio codec API, so we do not need the memcpy()
        /*for(i=0; i<len; i++) {
            av_log(avctx, AV_LOG_DEBUG, "%2X ", frame[i]);
        }*/
Michael Niedermayer's avatar
Michael Niedermayer committed
196
        return len;
197
    } else
Michael Niedermayer's avatar
Michael Niedermayer committed
198
        return 0;
199 200
}

201
static av_cold int MP3lame_encode_close(AVCodecContext *avctx)
202
{
203
    Mp3AudioContext *s = avctx->priv_data;
204

205
    av_freep(&avctx->coded_frame);
206

207 208
    lame_close(s->gfp);
    return 0;
209 210
}

211 212 213
#define OFFSET(x) offsetof(Mp3AudioContext, x)
#define AE AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
static const AVOption options[] = {
214
    { "reservoir", "Use bit reservoir.", OFFSET(reservoir), AV_OPT_TYPE_INT, { 1 }, 0, 1, AE },
215 216 217 218 219 220 221 222 223
    { NULL },
};

static const AVClass libmp3lame_class = {
    .class_name = "libmp3lame encoder",
    .item_name  = av_default_item_name,
    .option     = options,
    .version    = LIBAVUTIL_VERSION_INT,
};
224

225
AVCodec ff_libmp3lame_encoder = {
226 227 228 229 230 231 232 233 234 235 236 237 238
    .name                  = "libmp3lame",
    .type                  = AVMEDIA_TYPE_AUDIO,
    .id                    = CODEC_ID_MP3,
    .priv_data_size        = sizeof(Mp3AudioContext),
    .init                  = MP3lame_encode_init,
    .encode                = MP3lame_encode_frame,
    .close                 = MP3lame_encode_close,
    .capabilities          = CODEC_CAP_DELAY,
    .sample_fmts           = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16,
                                                             AV_SAMPLE_FMT_NONE },
    .supported_samplerates = sSampleRates,
    .long_name             = NULL_IF_CONFIG_SMALL("libmp3lame MP3 (MPEG audio layer 3)"),
    .priv_class            = &libmp3lame_class,
239
};