xwma.c 9.33 KB
Newer Older
Max Horn's avatar
Max Horn committed
1 2 3 4
/*
 * xWMA demuxer
 * Copyright (c) 2011 Max Horn
 *
5
 * This file is part of FFmpeg.
Max Horn's avatar
Max Horn committed
6
 *
7
 * FFmpeg is free software; you can redistribute it and/or
Max Horn's avatar
Max Horn committed
8 9 10 11
 * 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.
 *
12
 * FFmpeg is distributed in the hope that it will be useful,
Max Horn's avatar
Max Horn committed
13 14 15 16 17
 * 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
18
 * License along with FFmpeg; if not, write to the Free Software
Max Horn's avatar
Max Horn committed
19 20 21
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

22 23
#include <inttypes.h>

Max Horn's avatar
Max Horn committed
24
#include "avformat.h"
25
#include "internal.h"
Max Horn's avatar
Max Horn committed
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
#include "riff.h"

/*
 * Demuxer for xWMA, a Microsoft audio container used by XAudio 2.
 */

typedef struct {
    int64_t data_end;
} XWMAContext;

static int xwma_probe(AVProbeData *p)
{
    if (!memcmp(p->buf, "RIFF", 4) && !memcmp(p->buf + 8, "XWMA", 4))
        return AVPROBE_SCORE_MAX;
    return 0;
}

43
static int xwma_read_header(AVFormatContext *s)
Max Horn's avatar
Max Horn committed
44
{
45
    int64_t size;
46
    int ret;
Max Horn's avatar
Max Horn committed
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
    uint32_t dpds_table_size = 0;
    uint32_t *dpds_table = 0;
    unsigned int tag;
    AVIOContext *pb = s->pb;
    AVStream *st;
    XWMAContext *xwma = s->priv_data;
    int i;

    /* The following code is mostly copied from wav.c, with some
     * minor alterations.
     */

    /* check RIFF header */
    tag = avio_rl32(pb);
    if (tag != MKTAG('R', 'I', 'F', 'F'))
        return -1;
    avio_rl32(pb); /* file size */
    tag = avio_rl32(pb);
    if (tag != MKTAG('X', 'W', 'M', 'A'))
        return -1;

    /* parse fmt header */
    tag = avio_rl32(pb);
    if (tag != MKTAG('f', 'm', 't', ' '))
        return -1;
    size = avio_rl32(pb);
73
    st = avformat_new_stream(s, NULL);
Max Horn's avatar
Max Horn committed
74 75 76
    if (!st)
        return AVERROR(ENOMEM);

77 78 79
    ret = ff_get_wav_header(pb, st->codec, size);
    if (ret < 0)
        return ret;
Max Horn's avatar
Max Horn committed
80 81 82 83 84 85 86
    st->need_parsing = AVSTREAM_PARSE_NONE;

    /* All xWMA files I have seen contained WMAv2 data. If there are files
     * using WMA Pro or some other codec, then we need to figure out the right
     * extradata for that. Thus, ask the user for feedback, but try to go on
     * anyway.
     */
87
    if (st->codec->codec_id != AV_CODEC_ID_WMAV2) {
88
        avpriv_request_sample(s, "Unexpected codec (tag 0x04%x; id %d)",
Max Horn's avatar
Max Horn committed
89 90 91 92 93 94
                              st->codec->codec_tag, st->codec->codec_id);
    } else {
        /* In all xWMA files I have seen, there is no extradata. But the WMA
         * codecs require extradata, so we provide our own fake extradata.
         *
         * First, check that there really was no extradata in the header. If
Max Horn's avatar
Max Horn committed
95
         * there was, then try to use it, after asking the user to provide a
Max Horn's avatar
Max Horn committed
96 97 98 99 100 101 102
         * sample of this unusual file.
         */
        if (st->codec->extradata_size != 0) {
            /* Surprise, surprise: We *did* get some extradata. No idea
             * if it will work, but just go on and try it, after asking
             * the user for a sample.
             */
103
            avpriv_request_sample(s, "Unexpected extradata (%d bytes)",
Max Horn's avatar
Max Horn committed
104 105 106 107 108 109 110 111 112 113 114 115
                                  st->codec->extradata_size);
        } else {
            st->codec->extradata_size = 6;
            st->codec->extradata      = av_mallocz(6 + FF_INPUT_BUFFER_PADDING_SIZE);
            if (!st->codec->extradata)
                return AVERROR(ENOMEM);

            /* setup extradata with our experimentally obtained value */
            st->codec->extradata[4] = 31;
        }
    }

116 117 118 119 120 121 122 123 124 125 126
    if (!st->codec->channels) {
        av_log(s, AV_LOG_WARNING, "Invalid channel count: %d\n",
               st->codec->channels);
        return AVERROR_INVALIDDATA;
    }
    if (!st->codec->bits_per_coded_sample) {
        av_log(s, AV_LOG_WARNING, "Invalid bits_per_coded_sample: %d\n",
               st->codec->bits_per_coded_sample);
        return AVERROR_INVALIDDATA;
    }

Max Horn's avatar
Max Horn committed
127
    /* set the sample rate */
128
    avpriv_set_pts_info(st, 64, 1, st->codec->sample_rate);
Max Horn's avatar
Max Horn committed
129 130 131 132 133 134 135 136 137 138 139 140 141 142 143

    /* parse the remaining RIFF chunks */
    for (;;) {
        if (pb->eof_reached)
            return -1;
        /* read next chunk tag */
        tag = avio_rl32(pb);
        size = avio_rl32(pb);
        if (tag == MKTAG('d', 'a', 't', 'a')) {
            /* We assume that the data chunk comes last. */
            break;
        } else if (tag == MKTAG('d','p','d','s')) {
            /* Quoting the MSDN xWMA docs on the dpds chunk: "Contains the
             * decoded packet cumulative data size array, each element is the
             * number of bytes accumulated after the corresponding xWMA packet
Max Horn's avatar
Max Horn committed
144
             * is decoded in order."
Max Horn's avatar
Max Horn committed
145 146 147 148 149 150 151 152 153 154 155 156 157 158
             *
             * Each packet has size equal to st->codec->block_align, which in
             * all cases I saw so far was always 2230. Thus, we can use the
             * dpds data to compute a seeking index.
             */

            /* Error out if there is more than one dpds chunk. */
            if (dpds_table) {
                av_log(s, AV_LOG_ERROR, "two dpds chunks present\n");
                return -1;
            }

            /* Compute the number of entries in the dpds chunk. */
            if (size & 3) {  /* Size should be divisible by four */
159 160
                av_log(s, AV_LOG_WARNING,
                       "dpds chunk size %"PRId64" not divisible by 4\n", size);
Max Horn's avatar
Max Horn committed
161 162 163
            }
            dpds_table_size = size / 4;
            if (dpds_table_size == 0 || dpds_table_size >= INT_MAX / 4) {
164 165
                av_log(s, AV_LOG_ERROR,
                       "dpds chunk size %"PRId64" invalid\n", size);
Max Horn's avatar
Max Horn committed
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
                return -1;
            }

            /* Allocate some temporary storage to keep the dpds data around.
             * for processing later on.
             */
            dpds_table = av_malloc(dpds_table_size * sizeof(uint32_t));
            if (!dpds_table) {
                return AVERROR(ENOMEM);
            }

            for (i = 0; i < dpds_table_size; ++i) {
                dpds_table[i] = avio_rl32(pb);
                size -= 4;
            }
        }
        avio_skip(pb, size);
    }

    /* Determine overall data length */
    if (size < 0)
        return -1;
    if (!size) {
        xwma->data_end = INT64_MAX;
    } else
        xwma->data_end = avio_tell(pb) + size;


    if (dpds_table && dpds_table_size) {
        int64_t cur_pos;
        const uint32_t bytes_per_sample
                = (st->codec->channels * st->codec->bits_per_coded_sample) >> 3;

        /* Estimate the duration from the total number of output bytes. */
        const uint64_t total_decoded_bytes = dpds_table[dpds_table_size - 1];
201 202 203 204 205 206

        if(!bytes_per_sample) {
            av_log(s, AV_LOG_ERROR, "bytes_per_sample is 0\n");
            return AVERROR_INVALIDDATA;
        }

Max Horn's avatar
Max Horn committed
207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
        st->duration = total_decoded_bytes / bytes_per_sample;

        /* Use the dpds data to build a seek table.  We can only do this after
         * we know the offset to the data chunk, as we need that to determine
         * the actual offset to each input block.
         * Note: If we allowed ourselves to assume that the data chunk always
         * follows immediately after the dpds block, we could of course guess
         * the data block's start offset already while reading the dpds chunk.
         * I decided against that, just in case other chunks ever are
         * discovered.
         */
        cur_pos = avio_tell(pb);
        for (i = 0; i < dpds_table_size; ++i) {
            /* From the number of output bytes that would accumulate in the
             * output buffer after decoding the first (i+1) packets, we compute
             * an offset / timestamp pair.
             */
            av_add_index_entry(st,
                               cur_pos + (i+1) * st->codec->block_align, /* pos */
                               dpds_table[i] / bytes_per_sample,         /* timestamp */
                               st->codec->block_align,                   /* size */
                               0,                                        /* duration */
                               AVINDEX_KEYFRAME);
        }
    } else if (st->codec->bit_rate) {
        /* No dpds chunk was present (or only an empty one), so estimate
         * the total duration using the average bits per sample and the
         * total data length.
         */
        st->duration = (size<<3) * st->codec->sample_rate / st->codec->bit_rate;
    }

    av_free(dpds_table);

    return 0;
}

static int xwma_read_packet(AVFormatContext *s, AVPacket *pkt)
{
    int ret, size;
    int64_t left;
    AVStream *st;
    XWMAContext *xwma = s->priv_data;

    st = s->streams[0];

    left = xwma->data_end - avio_tell(s->pb);
    if (left <= 0) {
        return AVERROR_EOF;
    }

    /* read a single block; the default block size is 2230. */
    size = (st->codec->block_align > 1) ? st->codec->block_align : 2230;
    size = FFMIN(size, left);

    ret  = av_get_packet(s->pb, pkt, size);
    if (ret < 0)
        return ret;

    pkt->stream_index = 0;
    return ret;
}

AVInputFormat ff_xwma_demuxer = {
271 272 273 274 275 276
    .name           = "xwma",
    .long_name      = NULL_IF_CONFIG_SMALL("Microsoft xWMA"),
    .priv_data_size = sizeof(XWMAContext),
    .read_probe     = xwma_probe,
    .read_header    = xwma_read_header,
    .read_packet    = xwma_read_packet,
Max Horn's avatar
Max Horn committed
277
};