< Back to glossary

CTU/CU/PU/LCU

Coding tree unit (CTU) is the basic processing unit of the High Efficiency Video Coding (HEVC) and Versatile Video Coding (VVC) video standard and conceptually corresponds in structure to macroblock units that were used in several previous video standards. CTU is also referred to as the largest coding unit (LCU)

CTU (Coding Tree Unit, coding tree unit), each coding tree unit can be recursively divided, using the quad-tree structure, divided into 32×32, 16×16, 8×8 sub-regions, the following figure is a 64× 64 coding tree unit partition example.


Coding tree unit

CU (Coding Unit), each CTU can be further evenly divided into 4 square CUs, and a CU can be recursively divided into 4 small CUs according to the quad-tree structure, with a maximum depth of 3, as shown in the following figure:


Coding Unit

The CU block is the basic unit for decision-making in inter-frame, intra-frame, and Skip/Merge modes.

The resolution (length and width parameters) of the video sequence will also be transmitted in SPS. The length and width must be an integer multiple of the minimum CU size, but it may not be an integer multiple of the CTU size. For the case where the length and width are not integer multiples of the CTU size, the CTU at the image boundary is considered to have been segmented to coincide with the image boundary. For the default segmentation at this boundary, there is no need to transmit the split_cu_flags flag.

PU (Prediction Unit, prediction unit), PU specifies all the prediction modes of the coding unit, including all prediction-related information, such as intra prediction direction, inter prediction division method, motion vector prediction, and inter prediction reference Image index number, etc. PU is the basic unit for prediction. In different prediction methods, the division of PU is shown in the figure below:


Prediction Unit, prediction unit