Back to Search Document details
13th Meeting: Marrakech, January 2019 2019-01-09 18:07
CE10.1.4: Simplification of combined inter and intra prediction
Abstract
Since this is an early draft, topics outside of the specific aspects that have been established by recorded meeting agreements are not included in the specification. Such aspects are to be determined by further development of the VVC project in JVET. The high-level syntax for the standard is yet to be developed. The aspects of high-level syntax in this early draft are provided only to show how certain features are likely to be controlled by some high-level syntax that may have a sequence level, a picture level, and a slice level (a picture spatial region level that includes a subset of the CTUs of the picture).
JVET-M0177 CE10.1.4: Simplification of combined inter and intra prediction [M.-S. Chiang, C.-W. Hsu, Y.-W. Huang, S.-M. Lei (MediaTek)]

This contribution proposes to simplify combined inter and intra prediction (CIIP). The concept is to disable CIIP for coding units (CUs) with luma coding block (CB) sizes larger than 1024 luma samples for reducing intra prediction buffer size when CIIP is applied. It is reported that, compared to VTM3.0, the proposed simplification method results in 0.01% and 0.04% luma BD-rates with 1% and 0% encoding time decreases for RA and LB, respectively.

Decisions
This contribution proposes to simplify combined inter and intra prediction (CIIP). The concept is to disable CIIP for coding units (CUs) with luma coding block (CB) sizes larger than 1024 luma samples for reducing intra prediction buffer size when CIIP is applied. It is reported that, compared to VTM3.0, the proposed simplification method results in 0.01% and 0.04% luma BD-rates with 1% and 0% encoding time decreases for RA and LB, respectively.
Citation