ffmpeg 从mp4上提取H264的nalu

http://blog.csdn.net/gavinr/article/details/7183499

1.获取数据
ffmpeg读取mp4中的H264数据，并不能直接得到NALU，文件中也没有储存0x00000001的分隔符。下面这张图为packet.data中的数据


从图中可以发现，packet中的数据起始处没有分隔符(0x00000001), 也不是0x65、0x67、0x68、0x41等字节，所以可以肯定这不是标准的nalu。

其实，前4个字0x000032ce表示的是nalu的长度，从第5个字节开始才是nalu的数据。所以直接将前4个字节替换为0x00000001即可得到标准的nalu数据。


2.获取pps及sps

pps及sps不能从packet获得，而是保存在AVCodecContext的extradata数据域中。如下：




如何从extradata中解析出sps及pps呢？ffmpeg中提供了一个流过滤器"h264_mp4toannexb"完成这项工作，关键代码如下

//h264_mp4toannexb_bsf.c
static int h264_mp4toannexb_filter(AVBitStreamFilterContext *bsfc,
                                   AVCodecContext *avctx,  const char *args,
                                   uint8_t  **poutbuf,  int *poutbuf_size,
                                   const uint8_t *buf,int      buf_size,
                                   int keyframe) {
    H264BSFContext *ctx = bsfc->priv_data;
    uint8_t unit_type;
    int32_t nal_size;
    uint32_t cumul_size = 0;
    const uint8_t *buf_end = buf + buf_size;


    /* nothing to filter */
    if (!avctx->extradata || avctx->extradata_size < 6) {
        *poutbuf = (uint8_t*) buf;
        *poutbuf_size = buf_size;
        return 0;
    }

    //
    //从extradata中分析出SPS、PPS
    //
    /* retrieve sps and pps NAL units from extradata */
    if (!ctx->extradata_parsed) {
        uint16_t unit_size;
        uint64_t total_size = 0;
        uint8_t *out = NULL, unit_nb, sps_done = 0, sps_seen = 0, pps_seen = 0;
        const uint8_t *extradata = avctx->extradata+4; //跳过前4个字节
        static const uint8_t nalu_header[4] = {0, 0, 0, 1};


        /* retrieve length coded size */
        ctx->length_size = (*extradata++ & 0x3) + 1;     //用于指示表示编码数据长度所需字节数
        if (ctx->length_size == 3)
            return AVERROR(EINVAL);


        /* retrieve sps and pps unit(s) */
        unit_nb = *extradata++ & 0x1f; /* number of sps unit(s) */
        if (!unit_nb) {
            goto pps;
        } else {
            sps_seen = 1;
        }


        while (unit_nb--) {
            void *tmp;


            unit_size = AV_RB16(extradata);
            total_size += unit_size+4;
            if (total_size > INT_MAX - FF_INPUT_BUFFER_PADDING_SIZE ||
                extradata+2+unit_size > avctx->extradata+avctx->extradata_size) {
                av_free(out);
                return AVERROR(EINVAL);
            }
            tmp = av_realloc(out, total_size + FF_INPUT_BUFFER_PADDING_SIZE);
            if (!tmp) {
                av_free(out);
                return AVERROR(ENOMEM);
            }
            out = tmp;
            memcpy(out+total_size-unit_size-4, nalu_header, 4);
            memcpy(out+total_size-unit_size,   extradata+2, unit_size);
            extradata += 2+unit_size;
pps:
            if (!unit_nb && !sps_done++) {
                unit_nb = *extradata++; /* number of pps unit(s) */
                if (unit_nb)
                    pps_seen = 1;
            }
        }


        if(out)
            memset(out + total_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);


        if (!sps_seen)
            av_log(avctx, AV_LOG_WARNING, "Warning: SPS NALU missing or invalid. The resulting stream may not play.\n");
        if (!pps_seen)
            av_log(avctx, AV_LOG_WARNING, "Warning: PPS NALU missing or invalid. The resulting stream may not play.\n");


        av_free(avctx->extradata);
        avctx->extradata      = out;
        avctx->extradata_size = total_size;
        ctx->first_idr        = 1;
        ctx->extradata_parsed = 1;
    }


    *poutbuf_size = 0;
    *poutbuf = NULL;
    do {
        if (buf + ctx->length_size > buf_end)
            goto fail;   //buf为NULL时，以下代码将不再执行


        //
        //用于保存数据长度的字节数，是在分析原extradata计算出来的
        //
        if (ctx->length_size == 1) {
            nal_size = buf[0];
        } else if (ctx->length_size == 2) {
            nal_size = AV_RB16(buf);
        } else
            nal_size = AV_RB32(buf);


        buf += ctx->length_size;
        unit_type = *buf & 0x1f;


        if (buf + nal_size > buf_end || nal_size < 0)
            goto fail;


        /* prepend only to the first type 5 NAL unit of an IDR picture */
        if (ctx->first_idr && unit_type == 5) {
            //
            //copy IDR 帧时，需要将sps及pps一同拷贝
            //
            if (alloc_and_copy(poutbuf, poutbuf_size,
                               avctx->extradata, avctx->extradata_size,
                               buf, nal_size) < 0)
                goto fail;
            ctx->first_idr = 0;
        } else {
            //
            //非IDR帧，没有sps及pps
            if (alloc_and_copy(poutbuf, poutbuf_size,
                               NULL, 0,
                               buf, nal_size) < 0)
                goto fail;
            if (!ctx->first_idr && unit_type == 1)
                ctx->first_idr = 1;
        }


        buf += nal_size;
        cumul_size += nal_size + ctx->length_size;
    } while (cumul_size < buf_size);


    return 1;


fail:
    av_freep(poutbuf);
    *poutbuf_size = 0;
    return AVERROR(EINVAL);
}

//h264_mp4toannexb_bsf.c static int h264_mp4toannexb_filter(AVBitStreamFilterContext *bsfc, AVCodecContext *avctx, const char *args, uint8_t **poutbuf, int *poutbuf_size, const uint8_t *buf, int buf_size, int keyframe) { H264BSFContext *ctx = bsfc->priv_data; uint8_t unit_type; int32_t nal_size; uint32_t cumul_size = 0; const uint8_t *buf_end = buf + buf_size; /* nothing to filter */ if (!avctx->extradata || avctx->extradata_size < 6) { *poutbuf = (uint8_t*) buf; *poutbuf_size = buf_size; return 0; } // //从extradata中分析出SPS、PPS // /* retrieve sps and pps NAL units from extradata */ if (!ctx->extradata_parsed) { uint16_t unit_size; uint64_t total_size = 0; uint8_t *out = NULL, unit_nb, sps_done = 0, sps_seen = 0, pps_seen = 0; const uint8_t *extradata = avctx->extradata+4; //跳过前4个字节 static const uint8_t nalu_header[4] = {0, 0, 0, 1}; /* retrieve length coded size */ ctx->length_size = (*extradata++ & 0x3) + 1; //用于指示表示编码数据长度所需字节数 if (ctx->length_size == 3) return AVERROR(EINVAL); /* retrieve sps and pps unit(s) */ unit_nb = *extradata++ & 0x1f; /* number of sps unit(s) */ if (!unit_nb) { goto pps; } else { sps_seen = 1; } while (unit_nb--) { void *tmp; unit_size = AV_RB16(extradata); total_size += unit_size+4; if (total_size > INT_MAX - FF_INPUT_BUFFER_PADDING_SIZE || extradata+2+unit_size > avctx->extradata+avctx->extradata_size) { av_free(out); return AVERROR(EINVAL); } tmp = av_realloc(out, total_size + FF_INPUT_BUFFER_PADDING_SIZE); if (!tmp) { av_free(out); return AVERROR(ENOMEM); } out = tmp; memcpy(out+total_size-unit_size-4, nalu_header, 4); memcpy(out+total_size-unit_size, extradata+2, unit_size); extradata += 2+unit_size; pps: if (!unit_nb && !sps_done++) { unit_nb = *extradata++; /* number of pps unit(s) */ if (unit_nb) pps_seen = 1; } } if(out) memset(out + total_size, 0, FF_INPUT_BUFFER_PADDING_SIZE); if (!sps_seen) av_log(avctx, AV_LOG_WARNING, "Warning: SPS NALU missing or invalid. The resulting stream may not play.\n"); if (!pps_seen) av_log(avctx, AV_LOG_WARNING, "Warning: PPS NALU missing or invalid. The resulting stream may not play.\n"); av_free(avctx->extradata); avctx->extradata = out; avctx->extradata_size = total_size; ctx->first_idr = 1; ctx->extradata_parsed = 1; } *poutbuf_size = 0; *poutbuf = NULL; do { if (buf + ctx->length_size > buf_end) goto fail; //buf为NULL时，以下代码将不再执行 // //用于保存数据长度的字节数，是在分析原extradata计算出来的 // if (ctx->length_size == 1) { nal_size = buf[0]; } else if (ctx->length_size == 2) { nal_size = AV_RB16(buf); } else nal_size = AV_RB32(buf); buf += ctx->length_size; unit_type = *buf & 0x1f; if (buf + nal_size > buf_end || nal_size < 0) goto fail; /* prepend only to the first type 5 NAL unit of an IDR picture */ if (ctx->first_idr && unit_type == 5) { // //copy IDR 帧时，需要将sps及pps一同拷贝 // if (alloc_and_copy(poutbuf, poutbuf_size, avctx->extradata, avctx->extradata_size, buf, nal_size) < 0) goto fail; ctx->first_idr = 0; } else { // //非IDR帧，没有sps及pps if (alloc_and_copy(poutbuf, poutbuf_size, NULL, 0, buf, nal_size) < 0) goto fail; if (!ctx->first_idr && unit_type == 1) ctx->first_idr = 1; } buf += nal_size; cumul_size += nal_size + ctx->length_size; } while (cumul_size < buf_size); return 1; fail: av_freep(poutbuf); *poutbuf_size = 0; return AVERROR(EINVAL); }
一般情况下，extradata中包含一个sps、一个pps 的nalu, 从上面的代码中容易看出extradata的数据格式。分析后的sps及pps依然储存在extradata域中，并添加了起始符。从代码中还可以看出，上面的函数会将sps、pps及packet中的数据，都copy到poutbuf指示的内存中，如果不需要copy到指定内存，直接给buf参数传入空值即可。




3.使用ffmpeg的流过滤器获取sps及pps
流过滤器"h264_mp4toannexb", 在av_register_all()函数中会被注册。用法示例如下：

int ParseH264ExtraDataInMp4(int stream_id)
{
    uint8_t *dummy = NULL;
    int dummy_size;
    AVBitStreamFilterContext* bsfc =  av_bitstream_filter_init("h264_mp4toannexb");


    if(bsfc == NULL)
    {
        return -1;
    }


    av_bitstream_filter_filter(
            bsfc, format_ctx_->streams[stream_id]->codec, NULL, &dummy, &dummy_size, NULL, 0, 0);

int ParseH264ExtraDataInMp4(int stream_id) { uint8_t *dummy = NULL; int dummy_size; AVBitStreamFilterContext* bsfc = av_bitstream_filter_init("h264_mp4toannexb"); if(bsfc == NULL) { return -1; } av_bitstream_filter_filter( bsfc, format_ctx_->streams[stream_id]->codec, NULL, &dummy, &dummy_size, NULL, 0, 0);

    av_bitstream_filter_close(bsfc);


    return 0;
}