9

NALU（AVC Decoder Configuration Record）でSPSを取得し、そこからビデオの幅/高さを解析しようとしています。

67 64 00 15 ac c8 60 20 09 6c 04 40 00 00 03 00 40 00 00 07 a3 c5 8b 67 80

これは SPS を解析する私のコードですが、間違った値を取得します。pic_width_in_mbs_minus1 は 5、pic_height_in_map_units_minus1 は 1 です。実際の動画は 512 X 288px です

typedef struct _SequenceParameterSet
{
private:
    const unsigned char * m_pStart;
    unsigned short m_nLength;
    int m_nCurrentBit;

    unsigned int ReadBit()
    {
        ATLASSERT(m_nCurrentBit <= m_nLength * 8);
        int nIndex = m_nCurrentBit / 8;
        int nOffset = m_nCurrentBit % 8 + 1;

        m_nCurrentBit ++;
        return (m_pStart[nIndex] >> (8-nOffset)) & 0x01;
    }

    unsigned int ReadBits(int n)
    {
        int r = 0;
        int i;
        for (i = 0; i < n; i++)
        {
            r |= ( ReadBit() << ( n - i - 1 ) );
        }
        return r;
    }

    unsigned int ReadExponentialGolombCode()
    {
        int r = 0;
        int i = 0;

        while( (ReadBit() == 0) && (i < 32) )
        {
            i++;
        }
        r = ReadBits(i);
        r += (1 << i) - 1;
        return r;
    }


    unsigned int ReadSE() 
    {
        int r = ReadExponentialGolombCode();
        if (r & 0x01)
        {
            r = (r+1)/2;
        }
        else
        {
            r = -(r/2);
        }
        return r;
    }
public:

    void Parse(const unsigned char * pStart, unsigned short nLen)
    {
        m_pStart = pStart;
        m_nLength = nLen;
        m_nCurrentBit = 0;
        int profile_idc = ReadBits(8);
        int constraint_set0_flag = ReadBit();
        int constraint_set1_flag = ReadBit();
        int constraint_set2_flag = ReadBit();
        int constraint_set3_flag = ReadBit();
        int constraint_set4_flag = ReadBit();
        int constraint_set5_flag = ReadBit();
        int reserved_zero_2bits  = ReadBits(2);
        int level_idc = ReadBits(8);
        int seq_parameter_set_id = ReadExponentialGolombCode();

        if( profile_idc == 100 || profile_idc == 110 ||
            profile_idc == 122 || profile_idc == 144 )
        {
            int chroma_format_idc = ReadExponentialGolombCode();
            if( chroma_format_idc == 3 )
            {
                int residual_colour_transform_flag = ReadBit();
            }
            int bit_depth_luma_minus8 = ReadExponentialGolombCode();
            int bit_depth_chroma_minus8 = ReadExponentialGolombCode();
            int qpprime_y_zero_transform_bypass_flag = ReadBit();
            int seq_scaling_matrix_present_flag = ReadBit();
            if( seq_scaling_matrix_present_flag )
            {
                for( int i = 0; i < 8; i++ )
                {
                    int seq_scaling_list_present_flag = ReadBit();
                    if( seq_scaling_list_present_flag )
                    {
                        /*
                        if( i < 6 )
                        {
                            read_scaling_list( b, sps->ScalingList4x4[ i ], 16,
                                sps->UseDefaultScalingMatrix4x4Flag[ i ]);
                        }
                        else
                        {
                            read_scaling_list( b, sps->ScalingList8x8[ i - 6 ], 64,
                                sps->UseDefaultScalingMatrix8x8Flag[ i - 6 ] );
                        }
                        */
                    }
                }
            }
        }
        int log2_max_frame_num_minus4 = ReadExponentialGolombCode();
        int pic_order_cnt_type = ReadExponentialGolombCode();
        if( pic_order_cnt_type == 0 )
        {
            int log2_max_pic_order_cnt_lsb_minus4 = ReadExponentialGolombCode();
        }
        else if( pic_order_cnt_type == 1 )
        {
            int delta_pic_order_always_zero_flag = ReadBit();
            int offset_for_non_ref_pic = ReadSE();
            int offset_for_top_to_bottom_field = ReadSE();
            int num_ref_frames_in_pic_order_cnt_cycle = ReadExponentialGolombCode();
            for( int i = 0; i < num_ref_frames_in_pic_order_cnt_cycle; i++ )
            {
                ReadSE();
                //sps->offset_for_ref_frame[ i ] = ReadSE();
            }
        }
        int num_ref_frames = ReadExponentialGolombCode();
        int gaps_in_frame_num_value_allowed_flag = ReadBit();
        int pic_width_in_mbs_minus1 = ReadExponentialGolombCode();
        int pic_height_in_map_units_minus1 = ReadExponentialGolombCode();
        int frame_mbs_only_flag = ReadBit();
        if( !frame_mbs_only_flag )
        {
            int mb_adaptive_frame_field_flag = ReadBit();
        }
        int direct_8x8_inference_flag = ReadBit();
        int frame_cropping_flag = ReadBit();
        if( frame_cropping_flag )
        {
            int frame_crop_left_offset = ReadExponentialGolombCode();
            int frame_crop_right_offset = ReadExponentialGolombCode();
            int frame_crop_top_offset = ReadExponentialGolombCode();
            int frame_crop_bottom_offset = ReadExponentialGolombCode();
        }
        int vui_parameters_present_flag = ReadBit();

        pStart++;
    }
}SequenceParameterSet, *LPSequenceParameterSet;

これは SPS を解析する私のコードですが、間違った値を取得します。pic_width_in_mbs_minus1 は 5、pic_height_in_map_units_minus1 は 1 です。実際の動画は 512 X 288px です

ありがとう

4

4 に答える 4

4

はSPS（抜粋）：

profile_idc 100 
constraint_set0_flag 0 
constraint_set1_flag 0 
constraint_set2_flag 0 
constraint_set3_flag 0 
level_idc 21 
seq_parameter_set_id 0 
chroma_format_idc 1 
num_ref_frames 5 
gaps_in_frame_num_value_allowed_flag 0 
pic_width_in_mbs_minus1 31 
pic_height_in_map_units_minus1 17 
frame_mbs_only_flag 1 
direct_8x8_inference_flag 1 
frame_cropping_flag 0 
vui_parameters_present_flag 1

これは 512x288 ビデオです。データのデコード方法の詳細については、こちらを参照してください: H264Video ストリームのサイズの取得

于 2012-08-18T14:38:48.833 に答える