VC9 Block-level functions

Defines
#define	GET_MQUANT()
	Get macroblock-level quantizer scale.
#define	GET_MVDATA(_dmv_x, _dmv_y)
	Get MV differentials.
Functions
static int	vc9_pred_dc (MpegEncContext *s, int n, uint16_t **dc_val_ptr, int *dir_ptr)
int	vc9_decode_block (VC9Context *v, DCTELEM block[64], int n, int coded, int mquant)

---

Detailed Description

See also:

7.1.4, p91 and 8.1.1.7, p(1)04

Todo:

TODO: Integrate to MpegEncContext facilities

---

Define Documentation

 

#define GET_MQUANT

(

)

Value:

if (v->dquantfrm)                                            \
  {                                                            \
    if (v->dqprofile == DQPROFILE_ALL_MBS)                     \
    {                                                          \
      if (v->dqbilevel)                                        \
      {                                                        \
        mquant = (get_bits(gb, 1)) ? v->pq : v->altpq;         \
      }                                                        \
      else                                                     \
      {                                                        \
        mqdiff = get_bits(gb, 3);                              \
        if (mqdiff != 7) mquant = v->pq + mqdiff;              \
        else mquant = get_bits(gb, 5);                         \
      }                                                        \
    }                                                          \
    else mquant = v->pq;                                       \
  }

Get macroblock-level quantizer scale.

Warning:

XXX: qdiff to the frame quant, not previous quant ? XXX: Don't know how to initialize mquant otherwise in last case

Definition at line 1689 of file vc9.c.

Referenced by vc9_decode_b_mb(), and vc9_decode_p_mb().

#define GET_MVDATA	(	_dmv_x,
		_dmv_y		)

Get MV differentials.

See also:

MVDATA decoding from 8.3.5.2, p(1)20

Parameters:

	_dmv_x	Horizontal differential for decoded MV
	_dmv_y	Vertical differential for decoded MV

Todo:

TODO: Use MpegEncContext arrays to store them

Definition at line 1716 of file vc9.c.

Referenced by vc9_decode_b_mb(), and vc9_decode_p_mb().

---

Function Documentation

int vc9_decode_block	(	VC9Context *	v,
		DCTELEM	block[64],
		int	n,
		int	coded,
		int	mquant
	)

Decode one block, inter or intra

Parameters:

	v	The VC9 context
	block	8x8 DCT block
	n	Block index in the current MB (<4=>luma)
	coded	If the block is coded
	mquant	Quantizer step for the current block

See also:

Inter TT: Table 21, p73 + p91-85

Intra TT: Table 20, p72 + p(1)05-(1)07

Todo:

TODO: Process the blocks

Todo:

TODO: Use M$ MPEG-4 cbp prediction

Definition at line 1835 of file vc9.c.

References MpegEncContext::ac_pred, av_log(), AV_LOG_ERROR, MpegEncContext::avctx, MpegEncContext::block_last_index, MpegEncContext::c_dc_scale, MpegEncContext::c_dc_scale_table, MpegEncContext::dc_table_index, DC_VLC_BITS, decode012(), MpegEncContext::gb, get_bits(), get_vlc2(), MpegEncContext::mb_intra, mpeg4_pred_ac(), VC9Context::pq, s, VC9Context::s, VLC::table, VC9Context::tt_index, VC9Context::ttblk4x4, VC9Context::ttmb, vc9_pred_dc(), VC9_SUBBLKPAT_VLC_BITS, VC9_TTBLK_VLC_BITS, MpegEncContext::y_dc_scale, and MpegEncContext::y_dc_scale_table.

Referenced by vc9_decode_b_mb(), vc9_decode_i_mb(), and vc9_decode_p_mb().

Here is the call graph for this function:

static int vc9_pred_dc	(	MpegEncContext *	s,
		int	n,
		uint16_t **	dc_val_ptr,
		int *	dir_ptr
	)			[inline, static]

Get predicted DC value prediction dir: left=0, top=1

Parameters:

	s	MpegEncContext
[in]	n	block index in the current MB
	dc_val_ptr	Pointer to DC predictor
	dir_ptr	Prediction direction for use in AC prediction

Todo:

TODO: Actually do it the VC9 way

Todo:

TODO: Handle properly edges

Definition at line 1759 of file vc9.c.

References a, b, BI_TYPE, MpegEncContext::block_index, MpegEncContext::block_wrap, c, MpegEncContext::c_dc_scale, MpegEncContext::dc_val, MpegEncContext::first_slice_line, I_TYPE, MpegEncContext::mb_x, MpegEncContext::pict_type, pred(), and MpegEncContext::y_dc_scale.

Referenced by vc9_decode_block().

Here is the call graph for this function: