WO2013102298A1

WO2013102298A1 - Signaling size information of coding unit and transform unit in hevc

Info

Publication number: WO2013102298A1
Application number: PCT/CN2012/070030
Authority: WO
Inventors: Xiang Li; Jicheng An; Xun Guo; Shaw-Min Lei
Original assignee: Mediatek Singapore Pte. Ltd.
Priority date: 2012-01-04
Filing date: 2012-01-04
Publication date: 2013-07-11

Abstract

To efficiently signal the size information of coding unit and transform unit in high efficiency video coding (HEVC), the related syntax and semantics is revised by considering both value range and dependency between parameters. Accordingly, the coding efficiency is improved and the parsing process is simplified. Moreover, the proposed method put explicit constraints over parameters such that potential setting errors can be effectively avoided.

Description

SIGNALING SIZE INFORMATION OF CODING UNIT AND

TRANSFORM UNIT IN HEVC

BACKGROUND OF THE INVENTION

Field of the Invention [0001] The invention relates generally to video coding. In particular, the present invention relates to methods and apparatuses for signaling size information of coding unit and transform unit in high efficiency video coding (HEVC).

Description of the Related Art

[0002] HEVC is an emerging video coding standard being developed by ITU-T and MPEG. In current HEVC (working draft 5 and test model 5), size information of coding unit (CU) and transform unit (TU) is signaled with variable length coding ue(v) in sequence parameter set (SPS) as highlighted in Figure 1.

[0003] In general, the signaling method in Figure 1 is not efficient since the related value range and dependency between parameters are not considered. BRIEF SUMMARY OF THE INVENTION

[0004] In this invention, the size information signaling of coding unit and transform unit in HEVC WD5 is improved. Generally, both value range and dependency between size parameters are considered so that a higher coding efficiency and a simpler parsing process are achieved. Moreover, potential setting errors of size parameters can be effectively avoided by the proposed method.

BRIEF DESCRIPTION OF DRAWINGS

[0005] The invention relates generally to video coding. In particular, the present invention can be more fully understood by reading the subsequent detailed description and examples with references made to the accompanying drawings, wherein:

[0006] Figure 1 illustrates a sequence parameter set RBSP syntax.

[0007] Figure 2 is an exemplary diagram illustrating a proposed SPS parsing process. DETAILED DESCRIPTION OF THE INVENTION

[0008] The following description is of the best-contemplated mode of carrying out the invention. This description is made for the purpose of illustrating the general principles of the invention and should not be taken in a limiting sense. The scope of the invention is best determined by reference to the appended claims.

[0009] As previously mentioned, the current signaling method of CU and TU size information (highlighted part in Figure 1) is not efficient.

[0010] Firstly, it is not efficient to code log2_diff_max_min_coding_block_size with variable length coding ue(v). In practice, CU is limited to certain sizes. Normally, log2_diff_max_min_coding_block_size shall between 0 and 3 (CU size is from 8x8 to 64x64). If further extension is preferred, a range between 0 and 7 is sufficient (CU size is from 8x8 to 1024x1024). Accordingly, fixed-length coding such as u(2) or u(3) is desired for log2_diff_max_min_coding_block_size.

[0011] Secondly, it is not efficient to independently code log2_min_transform _block_size_minus2. Considering that minimum TU size shall be smaller than minimum CU size which has already been signaled, it is more efficient to code the difference between the minimum TU size and the minimum CU size.

[0012] Thirdly, it is not necessary to use variable length coding for log2_diff_max_ min_transform_block_size. Since only 4 different sizes of transform, i.e., 4x4, 8x8, 16x16, and 32x32, are defined in HEVC, log2_diff_max_min_transform_block_size is between 0 and 3.

Consequently, 2-bit fixed length coding is sufficient for log2_diff_max_min_transform_block_size.

[0013] Similarly to log2_diff_max_min_transform_bk>ck_size, max_transform_ hierarchy_depth_inter and max_transform_hierarchy_depth_intra are both between 1 and 4. It is more efficient to code max_transform_hierarchy_depth_inter_minusl and max_transform_hierarchy_depth_intra_minusl with 2-bit fixed length coding.

[0014] Based on above analyses, the signaling method is proposed as in Figure 2 wherein the related semantics are as follows.

[0015] log2_min_cu_size_minus_min_tu_size_minusl specifies the minimum size of a transform block. The variable Log2MinTrafoSize is set equal to Log2MinCUSize - log2_min_ cu_size_minus_min_tu_size_minusl - 1.

[0016] max_transform_hierarchy_depth_intra_minusl + 1 specifies the maximum hierarchy depth for transform blocks of coding blocks coded in intra prediction mode.

[0017] max_transform_hierarchy_depth_inter_minusl + 1 specifies the maximum hierarchy depth for transform units of coding units coded in inter prediction mode.

Claims

1. A method of signaling size information of coding units (CU) and transform units (TU) in emerging video coding standard HEVC (high efficiency video coding), comprising:

using 2-bit or 3-bit fixed length coding to code syntax element log2_diff_max_min_coding_block_size,log2_diff_max_min_transform_block_size, max_ transform_hierarchy_depth_inter_minusl,and max_transform_hierarchy_depth_intra_ minusl; and

using variable length coding to code minimum TU size information based on minimum CU size information.

2. The method as claimed in claim 1, wherein the syntax element max_transform_hierarchy_depth_intra_minusl specifies a maximum hierarchy depth for transform blocks of coding blocks coded in intra prediction mode, and max_transform_hierarchy_depth_inter_minusl specifies a maximum hierarchy depth for transform units of coding units coded in inter prediction mode.

3. The method as claimed in claim 1, wherein a new syntax element log2_min_cu_size_minus_min_tu_size_minusl is employed to code minimum TU size information based on a minimum CU size.