libavformat/latmenc.c
0ca36b4d
 /*
  * LATM/LOAS muxer
  * Copyright (c) 2011 Kieran Kunhya <kieran@kunhya.com>
  *
d814a839
  * This file is part of FFmpeg.
0ca36b4d
  *
d814a839
  * FFmpeg is free software; you can redistribute it and/or
0ca36b4d
  * modify it under the terms of the GNU Lesser General Public
  * License as published by the Free Software Foundation; either
  * version 2.1 of the License, or (at your option) any later version.
  *
d814a839
  * FFmpeg is distributed in the hope that it will be useful,
0ca36b4d
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  * Lesser General Public License for more details.
  *
  * You should have received a copy of the GNU Lesser General Public
d814a839
  * License along with FFmpeg; if not, write to the Free Software
0ca36b4d
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
 #include "libavcodec/get_bits.h"
 #include "libavcodec/put_bits.h"
 #include "libavcodec/avcodec.h"
 #include "libavcodec/mpeg4audio.h"
 #include "libavutil/opt.h"
 #include "avformat.h"
9bf3d01d
 #include "internal.h"
c1fcd7e1
 #include "rawenc.h"
0ca36b4d
 
d1a58afb
 #define MAX_EXTRADATA_SIZE 1024
 
daf8cf35
 typedef struct LATMContext {
32dfd8ca
     AVClass *av_class;
0ca36b4d
     int off;
     int channel_conf;
     int object_type;
     int counter;
     int mod;
b44a5ec7
     uint8_t buffer[0x1fff + MAX_EXTRADATA_SIZE + 1024];
0ca36b4d
 } LATMContext;
 
 static const AVOption options[] = {
     {"smc-interval", "StreamMuxConfig interval.",
e6153f17
      offsetof(LATMContext, mod), AV_OPT_TYPE_INT, {.i64 = 0x0014}, 0x0001, 0xffff, AV_OPT_FLAG_ENCODING_PARAM},
0ca36b4d
     {NULL},
 };
 
 static const AVClass latm_muxer_class = {
     .class_name = "LATM/LOAS muxer",
     .item_name  = av_default_item_name,
     .option     = options,
     .version    = LIBAVUTIL_VERSION_INT,
 };
 
 static int latm_decode_extradata(LATMContext *ctx, uint8_t *buf, int size)
 {
     MPEG4AudioConfig m4ac;
 
d1a58afb
     if (size > MAX_EXTRADATA_SIZE) {
         av_log(ctx, AV_LOG_ERROR, "Extradata is larger than currently supported.\n");
         return AVERROR_INVALIDDATA;
     }
fd095539
     ctx->off = avpriv_mpeg4audio_get_config(&m4ac, buf, size * 8, 1);
0ca36b4d
     if (ctx->off < 0)
         return ctx->off;
 
a736eb4a
     if (ctx->object_type == AOT_ALS && (ctx->off & 7)) {
         // as long as avpriv_mpeg4audio_get_config works correctly this is impossible
         av_log(ctx, AV_LOG_ERROR, "BUG: ALS offset is not byte-aligned\n");
         return AVERROR_INVALIDDATA;
     }
0ca36b4d
     /* FIXME: are any formats not allowed in LATM? */
 
     if (m4ac.object_type > AOT_SBR && m4ac.object_type != AOT_ALS) {
         av_log(ctx, AV_LOG_ERROR, "Muxing MPEG-4 AOT %d in LATM is not supported\n", m4ac.object_type);
         return AVERROR_INVALIDDATA;
     }
     ctx->channel_conf = m4ac.chan_config;
     ctx->object_type  = m4ac.object_type;
 
     return 0;
 }
 
 static int latm_write_header(AVFormatContext *s)
 {
     LATMContext *ctx = s->priv_data;
9200514a
     AVCodecParameters *par = s->streams[0]->codecpar;
0ca36b4d
 
6f69f7a8
     if (par->codec_id == AV_CODEC_ID_AAC_LATM)
c1fcd7e1
         return 0;
 
9200514a
     if (par->extradata_size > 0 &&
         latm_decode_extradata(ctx, par->extradata, par->extradata_size) < 0)
0ca36b4d
         return AVERROR_INVALIDDATA;
 
     return 0;
 }
 
fa8a6385
 static void latm_write_frame_header(AVFormatContext *s, PutBitContext *bs)
0ca36b4d
 {
     LATMContext *ctx = s->priv_data;
9200514a
     AVCodecParameters *par = s->streams[0]->codecpar;
0ca36b4d
     int header_size;
 
     /* AudioMuxElement */
     put_bits(bs, 1, !!ctx->counter);
 
     if (!ctx->counter) {
         /* StreamMuxConfig */
         put_bits(bs, 1, 0); /* audioMuxVersion */
         put_bits(bs, 1, 1); /* allStreamsSameTimeFraming */
         put_bits(bs, 6, 0); /* numSubFrames */
         put_bits(bs, 4, 0); /* numProgram */
         put_bits(bs, 3, 0); /* numLayer */
 
         /* AudioSpecificConfig */
         if (ctx->object_type == AOT_ALS) {
6f69f7a8
             header_size = par->extradata_size-(ctx->off >> 3);
             avpriv_copy_bits(bs, &par->extradata[ctx->off >> 3], header_size);
0ca36b4d
         } else {
eb24def4
             // + 3 assumes not scalable and dependsOnCoreCoder == 0,
             // see decode_ga_specific_config in libavcodec/aacdec.c
9200514a
             avpriv_copy_bits(bs, par->extradata, ctx->off + 3);
0ca36b4d
 
             if (!ctx->channel_conf) {
0c1e9c2d
                 GetBitContext gb;
6f69f7a8
                 int ret = init_get_bits8(&gb, par->extradata, par->extradata_size);
879603fa
                 av_assert0(ret >= 0); // extradata size has been checked already, so this should not fail
0c1e9c2d
                 skip_bits_long(&gb, ctx->off + 3);
831018b0
                 ff_copy_pce_data(bs, &gb);
0ca36b4d
             }
         }
 
         put_bits(bs, 3, 0); /* frameLengthType */
d4a544cb
         put_bits(bs, 8, 0xff); /* latmBufferFullness */
0ca36b4d
 
         put_bits(bs, 1, 0); /* otherDataPresent */
         put_bits(bs, 1, 0); /* crcCheckPresent */
     }
 
     ctx->counter++;
     ctx->counter %= ctx->mod;
 }
 
 static int latm_write_packet(AVFormatContext *s, AVPacket *pkt)
 {
b44a5ec7
     LATMContext *ctx = s->priv_data;
8b3ec51d
     AVCodecParameters *par = s->streams[0]->codecpar;
0ca36b4d
     AVIOContext *pb = s->pb;
     PutBitContext bs;
     int i, len;
     uint8_t loas_header[] = "\x56\xe0\x00";
 
8b3ec51d
     if (par->codec_id == AV_CODEC_ID_AAC_LATM)
c1fcd7e1
         return ff_raw_write_packet(s, pkt);
 
8b3ec51d
     if (!par->extradata) {
b454c64e
         if(pkt->size > 2 && pkt->data[0] == 0x56 && (pkt->data[1] >> 4) == 0xe &&
             (AV_RB16(pkt->data + 1) & 0x1FFF) + 3 == pkt->size)
             return ff_raw_write_packet(s, pkt);
8b3ec51d
         else {
             uint8_t *side_data;
             int side_data_size = 0, ret;
 
             side_data = av_packet_get_side_data(pkt, AV_PKT_DATA_NEW_EXTRADATA,
                                                 &side_data_size);
             if (side_data_size) {
                 if (latm_decode_extradata(ctx, side_data, side_data_size) < 0)
                     return AVERROR_INVALIDDATA;
                 ret = ff_alloc_extradata(par, side_data_size);
                 if (ret < 0)
                     return ret;
                 memcpy(par->extradata, side_data, side_data_size);
             }
         }
b454c64e
     }
 
8e357e8e
     if (pkt->size > 0x1fff)
         goto too_large;
0ca36b4d
 
b44a5ec7
     init_put_bits(&bs, ctx->buffer, pkt->size+1024+MAX_EXTRADATA_SIZE);
0ca36b4d
 
     latm_write_frame_header(s, &bs);
 
     /* PayloadLengthInfo() */
     for (i = 0; i <= pkt->size-255; i+=255)
         put_bits(&bs, 8, 255);
 
     put_bits(&bs, 8, pkt->size-i);
 
     /* The LATM payload is written unaligned */
 
9540476b
     /* PayloadMux() */
912a838e
     if (pkt->size && (pkt->data[0] & 0xe1) == 0x81) {
         // Convert byte-aligned DSE to non-aligned.
         // Due to the input format encoding we know that
         // it is naturally byte-aligned in the input stream,
         // so there are no padding bits to account for.
         // To avoid having to add padding bits and rearrange
         // the whole stream we just remove the byte-align flag.
         // This allows us to remux our FATE AAC samples into latm
         // files that are still playable with minimal effort.
         put_bits(&bs, 8, pkt->data[0] & 0xfe);
9540476b
         avpriv_copy_bits(&bs, pkt->data + 1, 8*pkt->size - 8);
     } else
         avpriv_copy_bits(&bs, pkt->data, 8*pkt->size);
0ca36b4d
 
9f51c682
     avpriv_align_put_bits(&bs);
0ca36b4d
     flush_put_bits(&bs);
 
     len = put_bits_count(&bs) >> 3;
 
8e357e8e
     if (len > 0x1fff)
         goto too_large;
 
0ca36b4d
     loas_header[1] |= (len >> 8) & 0x1f;
     loas_header[2] |= len & 0xff;
 
     avio_write(pb, loas_header, 3);
b44a5ec7
     avio_write(pb, ctx->buffer, len);
0ca36b4d
 
     return 0;
8e357e8e
 
 too_large:
     av_log(s, AV_LOG_ERROR, "LATM packet size larger than maximum size 0x1fff\n");
     return AVERROR_INVALIDDATA;
0ca36b4d
 }
 
9bf3d01d
 static int latm_check_bitstream(struct AVFormatContext *s, const AVPacket *pkt)
 {
     int ret = 1;
     AVStream *st = s->streams[pkt->stream_index];
 
6f69f7a8
     if (st->codecpar->codec_id == AV_CODEC_ID_AAC) {
9bf3d01d
         if (pkt->size > 2 && (AV_RB16(pkt->data) & 0xfff0) == 0xfff0)
             ret = ff_stream_add_bitstream_filter(st, "aac_adtstoasc", NULL);
     }
 
     return ret;
 }
 
0ca36b4d
 AVOutputFormat ff_latm_muxer = {
     .name           = "latm",
     .long_name      = NULL_IF_CONFIG_SMALL("LOAS/LATM"),
     .mime_type      = "audio/MP4A-LATM",
f6e99bf5
     .extensions     = "latm,loas",
0ca36b4d
     .priv_data_size = sizeof(LATMContext),
36ef5369
     .audio_codec    = AV_CODEC_ID_AAC,
     .video_codec    = AV_CODEC_ID_NONE,
0ca36b4d
     .write_header   = latm_write_header,
     .write_packet   = latm_write_packet,
     .priv_class     = &latm_muxer_class,
9bf3d01d
     .check_bitstream= latm_check_bitstream,
f792d3cb
     .flags          = AVFMT_NOTIMESTAMPS,
0ca36b4d
 };