Intra-Picture Prediction in HEVC - High Efficiency Video Coding (HEVC) - page 106

Tabl e 4. 5 Binarization of the syntax element intra_chroma_pred_mode
with respect to the chroma intra prediction mode
Chroma intra prediction mode Binarization of intra_chroma_pred_mode

INTRA_DERIVED 0

INTRA_PLANAR 100

INTRA_ANGULAR[26] 101

INTRA_ANGULAR[10] 110

INTRA_DC 111

4.4.3 Syntax Design for Intra Mode Coding

Let three sorted most probable modes in an ascending order for luma intra
mode be SMPM [0], SMPM [1] and SMPM [2], respectively. The syntax element
prev_intra_luma_pred_flag specifies whether the luma intra mode is
equal to one of the three most probable modes. In this case, the additional syntax
element mpm_idx indicates that SMPM [ mpm_idx ] is the selected luma intra
mode.
Otherwise ( prev_intra_luma_pred_flag is equal to 0), the syntax ele-
ment rem_intra_luma_pred_mode is used to specify the luma intra mode.
Since the total number of luma intra modes is equal to 35 and the number of most
probable modes is equal to 3, the range of rem_intra_luma_pred_mode is
from 0 to 31, inclusive, which can be expressed by a 5-bit codeword conveniently.
Given rem_intra_luma_pred_mode , the luma intra mode is derived as
follows:

1. Let L be equal to the value of rem_intra_luma_pred_mode .
2. For i D 0 :::2 (sequentially in an increasing order), L is incremented by one if L
is greater than or equal to SMPM [ i ].

For chroma intra mode coding, it has been observed that the derived mode
is used most frequently while other four chroma intra modes with their own
identifiers (planar, angular 26, angular 10 and DC modes) are typically used with
roughly similar probabilities. The syntax element intra_chroma_pred_mode
is binarized with respect to the above observations as given in Table 4.5 .

4.5 Encoding Algorithms

Due to the large number of intra modes in HEVC, the computations of the rate-
distortion costs for all intra modes are impractical for most of the applications. In
the HEVC reference software, the sum of absolute transformed differences (SATD)
between prediction and original samples is used to reduce the number of luma
intra mode candidates before applying rate-distortion optimization (RDO) [ 12 ].