{"title":"Low-Frequency Non-Separable Transform Hardware System Design for the VVC Encoder","authors":"J. Goebel, L. Agostini, B. Zatt, M. Porto","doi":"10.1109/SBCCI55532.2022.9893228","DOIUrl":null,"url":null,"abstract":"This paper presents a dedicated hardware system design for the Low-Frequency Non-Separable Transform (LFNST) of the Versatile Video Coding (VVC/H.266) standard. The LFNST is a secondary transform used to transform the coefficients that came from the DCT-II transform. The developed design exploits two clock domains, where the LFNST core is working at 746.48 MHz and the primary transform can operate at a slower clock of only 186.62MHz to be able to process Ultra-High Definition (UHD) videos with $4098\\times 2160$ pixels (4K) at 60 frames per second. The whole LFNST hardware system design presents an area utilization of 57.3 Kgates and a power dissipation of 32.22 mW (processing the LFNST $4\\times 4$ through TU size of $4\\times 4$) when synthesized for an ASIC implementation with a 40nm technology standard cells library.","PeriodicalId":231587,"journal":{"name":"2022 35th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design (SBCCI)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 35th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design (SBCCI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/SBCCI55532.2022.9893228","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
This paper presents a dedicated hardware system design for the Low-Frequency Non-Separable Transform (LFNST) of the Versatile Video Coding (VVC/H.266) standard. The LFNST is a secondary transform used to transform the coefficients that came from the DCT-II transform. The developed design exploits two clock domains, where the LFNST core is working at 746.48 MHz and the primary transform can operate at a slower clock of only 186.62MHz to be able to process Ultra-High Definition (UHD) videos with $4098\times 2160$ pixels (4K) at 60 frames per second. The whole LFNST hardware system design presents an area utilization of 57.3 Kgates and a power dissipation of 32.22 mW (processing the LFNST $4\times 4$ through TU size of $4\times 4$) when synthesized for an ASIC implementation with a 40nm technology standard cells library.