global.h

H.264编码实现

  int                 DFDisableIdc;
  int                 DFAlphaC0Offset;
  int                 DFBetaOffset;

  int                 skip_flag;

  char                intra_pred_modes[MB_BLOCK_PARTITIONS];
  char                intra_pred_modes8x8[MB_BLOCK_PARTITIONS];           //!< four 8x8 blocks in a macroblock

  int64               cbp_blk ;    //!< 1 bit set for every 4x4 block with coefs (not implemented for INTRA)
  int64               cbp_bits[3];
  int64               cbp_bits_8x8[3];

  short               bipred_me[4]; //!< bi prediction ME type for a 8x8 block (0 : not used, 1: type 1, 2: type 2)

  // rate control
  double              actj;               // macroblock activity measure for macroblock j
  int                 prev_qp;
  int                 prev_dqp;
  int                 prev_cbp;


  struct macroblock   *mb_available_up;   //!< pointer to neighboring MB (CABAC)
  struct macroblock   *mb_available_left; //!< pointer to neighboring MB (CABAC)

} Macroblock;



//! Bitstream
typedef struct
{
  int             byte_pos;           //!< current position in bitstream;
  int             bits_to_go;         //!< current bitcounter
  
  int             stored_byte_pos;    //!< storage for position in bitstream;
  int             stored_bits_to_go;  //!< storage for bitcounter
  int             byte_pos_skip;      //!< storage for position in bitstream;
  int             bits_to_go_skip;    //!< storage for bitcounter
  int             write_flag;         //!< Bitstream contains data and needs to be written

  byte            byte_buf;           //!< current buffer for last written byte
  byte            stored_byte_buf;    //!< storage for buffer of last written byte
  byte            byte_buf_skip;      //!< current buffer for last written byte
  byte            *streamBuffer;      //!< actual buffer for written bytes

#if TRACE
  Boolean             trace_enabled;
#endif

} Bitstream;

//! DataPartition
typedef struct datapartition
{
  Bitstream           *bitstream;
  NALU_t              *nal_unit;
  EncodingEnvironment ee_cabac;
  EncodingEnvironment ee_recode;
} DataPartition;

//! Slice
typedef struct
{
  int                 picture_id;
  int                 qp;
  int                 qs;
  int                 picture_type; //!< picture type
  int                 start_mb_nr;
  int                 max_part_nr;  //!< number of different partitions
  int                 num_mb;       //!< number of MBs in the slice

  int                 ref_pic_list_reordering_flag_l0;
  int                 *reordering_of_pic_nums_idc_l0;
  int                 *abs_diff_pic_num_minus1_l0;
  int                 *long_term_pic_idx_l0;
  int                 ref_pic_list_reordering_flag_l1;
  int                 *reordering_of_pic_nums_idc_l1;
  int                 *abs_diff_pic_num_minus1_l1;
  int                 *long_term_pic_idx_l1;
  int                 field_ctx[3][2]; //GB

  int                 symbol_mode;

  Boolean             (*slice_too_big)(int bits_slice); //!< for use of callback functions

  DataPartition       *partArr;     //!< array of partitions
  MotionInfoContexts  *mot_ctx;     //!< pointer to struct of context models for use in CABAC
  TextureInfoContexts *tex_ctx;     //!< pointer to struct of context models for use in CABAC

} Slice;



#define MAXSLICEPERPICTURE 100

enum {
  SSE              = 0,
  SSE_RGB          = 1,  
  PSNR             = 2,
  PSNR_RGB         = 3,
  SSIM             = 4,
  SSIM_RGB         = 5,
  MS_SSIM          = 6,
  MS_SSIM_RGB      = 7,
  TOTAL_DIST_TYPES = 8
} distortion_types;

//! DistortionParams
typedef struct
{
  float value[3];                   //!< current frame distortion
  float average[3];                 //!< average frame distortion
  float avslice[NUM_SLICE_TYPES][3]; //!< average frame type distortion
} DistMetric;

typedef struct
{
  int        frame_ctr;                        //!< number of coded frames
  DistMetric metric[TOTAL_DIST_TYPES];      //!< Distortion metrics
} DistortionParams;

typedef struct
{
  int   no_slices;
  int   bits_per_picture;
  Slice *slices[MAXSLICEPERPICTURE];

  DistMetric distortion;
  byte  idr_flag;
} Picture;

Picture *p_frame_pic;
Picture **frame_pic;
Picture **field_pic;
Picture *frame_pic_si;

#ifdef _LEAKYBUCKET_
long *Bit_Buffer;
#endif

// global picture format dependend buffers, mem allocation in image.c
imgpel **pImgOrg[MAX_PLANE];
imgpel **pCurImg;            //!< Reference image. Luma for other profiles, can be any component for 4:4:4
int    **imgY_sub_tmp;       //!< Y picture temporary component (Quarter pel)

int **PicPos;
unsigned int log2_max_frame_num_minus4;
unsigned int log2_max_pic_order_cnt_lsb_minus4;
unsigned int max_frame_num;
unsigned int max_pic_order_cnt_lsb;


time_t  me_tot_time,me_time;
pic_parameter_set_rbsp_t *active_pps;
seq_parameter_set_rbsp_t *active_sps;


int dsr_new_search_range; //!<Dynamic Search Range.
//////////////////////////////////////////////////////////////////////////
// B pictures
// motion vector : forward, backward, direct
byte  MBPairIsField;     //!< For MB level field/frame coding tools


//Weighted prediction
int ***wp_weight;  // weight in [list][index][component] order
int ***wp_offset;  // offset in [list][index][component] order
int ****wbp_weight;  // weight in [list][fwd_index][bwd_idx][component] order
int luma_log_weight_denom;
int chroma_log_weight_denom;
int wp_luma_round;
int wp_chroma_round;

// global picture format dependend buffers, mem allocation in image.c (field picture)
imgpel   **img_org_top[MAX_PLANE];
imgpel   **img_org_bot[MAX_PLANE];
imgpel   **img_org_frm[MAX_PLANE];

imgpel   **imgY_com;               //!< Encoded luma images
imgpel  ***imgUV_com;              //!< Encoded croma images

char    ***direct_ref_idx;           //!< direct mode reference index buffer
char    **direct_pdir;              //!< direct mode reference index buffer

// Buffers for rd optimization with packet losses, Dim. Kontopodis
byte **pixel_map;   //!< Shows the latest reference frame that is reliable for each pixel
byte **refresh_map; //!< Stores the new values for pixel_map
int intras;         //!< Counts the intra updates in each frame.

int  frame_no, nextP_tr_fld, nextP_tr_frm;

int64 tot_time;

#define ET_SIZE 300      //!< size of error text buffer
char errortext[ET_SIZE]; //!< buffer for error message for exit with error()

// Residue Color Transform
char b8_ipredmode8x8[4][4], b8_intra_pred_modes8x8[16];

//! Info for the "decoders-in-the-encoder" used for rdoptimization with packet losses
typedef struct
{
  imgpel ***dec_mbY;        //!< Best decoded macroblock pixel values in RDO
  imgpel ***dec_mbY8x8;     //!< Best decoded 8x8 mode pixel values in RDO
  int    res_img[MAX_PLANE][16][16]; //Residual values for macroblock
} Decoders;
extern Decoders *decs;


#define FILE_NAME_SIZE 200

typedef struct
{
  double md;     //!< Mode decision Lambda
  double me[3];  //!< Motion Estimation Lambda
  int    mf[3];  //!< Integer formatted Motion Estimation Lambda
} LambdaParams;

// VUI Parameters
typedef struct
{
  int aspect_ratio_info_present_flag;
  int aspect_ratio_idc;
  int sar_width;
  int sar_height;
  int overscan_info_present_flag;
  int overscan_appropriate_flag;
  int video_signal_type_present_flag;
  int video_format;
  int video_full_range_flag;
  int colour_description_present_flag;
  int colour_primaries;
  int transfer_characteristics; 
  int matrix_coefficients;
  int chroma_location_info_present_flag;
  int chroma_sample_loc_type_top_field;
  int chroma_sample_loc_type_bottom_field;
  int timing_info_present_flag;
  int num_units_in_tick;
  int time_scale;
  int fixed_frame_rate_flag;
  int nal_hrd_parameters_present_flag;
  int nal_cpb_cnt_minus1;
  int nal_bit_rate_scale;
  int nal_cpb_size_scale;
  int nal_bit_rate_value_minus1;
  int nal_cpb_size_value_minus1;
  int nal_vbr_cbr_flag;
  int nal_initial_cpb_removal_delay_length_minus1;
  int nal_cpb_removal_delay_length_minus1;
  int nal_dpb_output_delay_length_minus1;
  int nal_time_offset_length;
  int vcl_hrd_parameters_present_flag;
  int vcl_cpb_cnt_minus1;
  int vcl_bit_rate_scale;
  int vcl_cpb_size_scale;
  int vcl_bit_rate_value_minus1;
  int vcl_cpb_size_value_minus1;
  int vcl_vbr_cbr_flag;
  int vcl_initial_cpb_removal_delay_length_minus1;
  int vcl_cpb_removal_delay_length_minus1;
  int vcl_dpb_output_delay_length_minus1;
  int vcl_time_offset_length;
  int low_delay_hrd_flag;
  int pic_struct_present_flag;
  int bitstream_restriction_flag;
  int motion_vectors_over_pic_boundaries_flag;
  int max_bytes_per_pic_denom;
  int max_bits_per_mb_denom;
  int log2_max_mv_length_vertical;
  int log2_max_mv_length_horizontal;
  int num_reorder_frames;
  int max_dec_frame_buffering;
} VUIParameters;

//! block 8x8 temporary RD info
typedef struct 
{
   short  best8x8mode            [4];
   char   best8x8pdir   [MAXMODE][4];
   char   best8x8l0ref  [MAXMODE][4];
   char   best8x8l1ref  [MAXMODE][4];
   short  bipred8x8me   [MAXMODE][4];
} Block8x8Info;

                             
//! all input parameters
typedef struct
{
  int ProfileIDC;                    //!< value of syntax element profile_idc
  int LevelIDC;                      //!< value of syntax element level_idc
  int IntraProfile;                  //!< Enable Intra profiles

  int no_frames;                     //!< number of frames to be encoded
  int qp0;                           //!< QP of first frame
  int qpN;                           //!< QP of remaining frames
  int jumpd;                         //!< number of frames to skip in input sequence (e.g 2 takes frame 0,3,6,9...)
  int DisableSubpelME;               //!< Disable sub-pixel motion estimation
  int search_range;                  /*!< search range - integer pel search and 16x16 blocks.  The search window is
                                          generally around the predicted vector. Max vector is 2xmcrange.  For 8x8
                                          and 4x4 block sizes the search range is 1/2 of that for 16x16 blocks.       */
  int num_ref_frames;                //!< number of reference frames to be used
  int P_List0_refs;                  //!< number of reference picture in list 0 in P pictures
  int B_List0_refs;                  //!< number of reference picture in list 0 in B pictures
  int B_List1_refs;                  //!< number of reference picture in list 1 in B pictures
  int Log2MaxFNumMinus4;             //!< value of syntax element log2_max_frame_num
  int Log2MaxPOCLsbMinus4;           //!< value of syntax element log2_max_pic_order_cnt_lsb_minus4

  // Input/output sequence format related variables
  FrameFormat source;                   //!< source related information
  FrameFormat output;                   //!< output related information
  int rgb_input_flag;
  int src_resize;                    //!< Control if input sequence will be resized (currently only cropping is supported)
  int src_BitDepthRescale;           //!< Control if input sequence bitdepth should be adjusted
  int yuv_format;                    //!< YUV format (0=4:0:0, 1=4:2:0, 2=4:2:2, 3=4:4:4)
  int intra_upd;                     /*!< For error robustness. 0: no special action. 1: One GOB/frame is intra coded
                                          as regular 'update'. 2: One GOB every 2 frames is intra coded etc.
                                          In connection with this intra update, restrictions is put on motion vectors
                                          to prevent errors to propagate from the past                                */
  int blc_size[8][2];                //!< array for different block sizes
  int part_size[8][2];               //!< array for different partition sizes

  int slice_mode;                    //!< Indicate what algorithm to use for setting slices
  int slice_argument;                //!< Argument to the specified slice algorithm
  int UseConstrainedIntraPred;       //!< 0: Inter MB pixels are allowed for intra prediction 1: Not allowed
  int  infile_header;                //!< If input file has a header set this to the length of the header
  char infile     [FILE_NAME_SIZE];  //!< YUV 4:2:0 input format
  char outfile    [FILE_NAME_SIZE];  //!< H.264 compressed output bitstream
  char ReconFile  [FILE_NAME_SIZE];  //!< Reconstructed Pictures
  char TraceFile  [FILE_NAME_SIZE];  //!< Trace Outputs
  char StatsFile  [FILE_NAME_SIZE];  //!< Stats File
  char QmatrixFile[FILE_NAME_SIZE];  //!< Q matrix cfg file
  int  EnableOpenGOP;                 //!< support for open gops.
  int  EnableIDRGOP;                  //!< support for open gops.

  int idr_period;                    //!< IDR picture period
  int intra_period;                  //!< intra picture period
  int intra_delay;                   //!< IDR picture delay
  int adaptive_idr_period;
  int adaptive_intra_period;         //!< reinitialize start of intra period

  int start_frame;                   //!< Encode sequence starting from Frame start_frame

  int GenerateMultiplePPS;
  int GenerateSEIMessage;
  char SEIMessageText[500];

  int ResendSPS;
  int ResendPPS;

  // B pictures
  int successive_Bframe;             //!< number of B frames that will be used
  int PReplaceBSlice;
  int qpB;                           //!< QP for non-reference B slice coded pictures
  int qpBRSOffset;                   //!< QP for reference B slice coded pictures
  int direct_spatial_mv_pred_flag;   //!< Direct Mode type to be used (0: Temporal, 1: Spatial)
  int directInferenceFlag;           //!< Direct Mode Inference Flag

  int BiPredMotionEstimation;
  int BiPredSearch[4];               //!< Use Bi prediction for modes 16x16, 16x8, 8x16, and 8x8
  
  int BiPredMERefinements;           //!< Number of bipred refinements
  int BiPredMESearchRange;
  int BiPredMESubPel;

  // SP Pictures
  int sp_periodicity;                //!< The periodicity of SP-pictures
  int qpsp;                          //!< SP Picture QP for prediction error
  int qpsp_pred;                     //!< SP Picture QP for predicted block

  int si_frame_indicator;            //!< Flag indicating whether SI frames should be encoded rather than SP frames (0: not used, 1: used)
  int sp2_frame_indicator;           //!< Flag indicating whether switching SP frames should be encoded rather than SP frames (0: not used, 1: used)
  int sp_output_indicator;           //!< Flag indicating whether coefficients are output to allow future encoding of switchin SP frames (0: not used, 1: used)
  char sp_output_filename[FILE_NAME_SIZE];    //!<Filename where SP coefficients are output
  char sp2_input_filename1[FILE_NAME_SIZE];   //!<Filename of coefficients of the first bitstream when encoding SP frames to switch bitstreams
  char sp2_input_filename2[FILE_NAME_SIZE];   //!<Filenames of coefficients of the second bitstream when encoding SP frames to switch bitstreams

  int WeightedPrediction;            //!< Weighted prediction for P frames (0: not used, 1: explicit)
  int WeightedBiprediction;          //!< Weighted prediction for B frames (0: not used, 1: explicit, 2: implicit)
  int WPMethod;                      //!< WP method (0: DC, 1: LMS)
  int WPMCPrecision;
  int WPMCPrecFullRef;
  int WPMCPrecBSlice;
  int EnhancedBWeightSupport;
  int ChromaWeightSupport;           //!< Weighted prediction support for chroma (0: disabled, 1: enabled)
  int UseWeightedReferenceME;        //!< Use Weighted Reference for ME.
  int RDPictureDecision;             //!< Perform RD optimal decision between various coded versions of same picture