{"title":"HDR videocompression with VPX","authors":"P. Topiwala, W. Dai, M. Krishnan","doi":"10.1109/DMIAF.2016.7574904","DOIUrl":null,"url":null,"abstract":"This paper present approaches to code HDR video using VP10, an open source video codec which is part of the WebM project developed by Google. Three techniques are developed in this paper: 1) An HDR video coding processing chain developed recently in standards committees (the Joint Collaborative Team on Video Coding, or JCT-VC, of ISO/IEC/ITU), but adapted to the VP10 codec, herein called HDR_VP10. 2) A modified version of HDR_VP10 with different intermediate color transforms and advanced sampling filters herein called FastVDO_ECHDR_VP10 and 3) FastVDO_HDR_VP10 that uses a new video data adaptive tuning process, which differs from the ST.2084 transfer function used in the anchor. HDR_VP10 is used as reference point for subjective and objective comparisons. Representative objective results for these systems include: (a) FastVDO_ECHDR_VP10, Overall results for RGB-PSNR, DE100, MD100, PSNRL100 were 8.0%, -3.5%, 46.2%, 2.9%; (b) FastVDO_HDR_VP10, Overall results for RGB-PSNR, DE100, MD100, PSNRL100 were -40.2%, -17.0%, -5.5% and -2.8% respectively. It is asserted (and for which there is appears to be fairly broad agreement) that such objective metrics are currently not very predictive of visual quality in HDR coding studies. FastVDO_HDR_VP10 shows a significant visual quality improvement on HDR_VP10.","PeriodicalId":404025,"journal":{"name":"2016 Digital Media Industry & Academic Forum (DMIAF)","volume":"118 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 Digital Media Industry & Academic Forum (DMIAF)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/DMIAF.2016.7574904","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
This paper present approaches to code HDR video using VP10, an open source video codec which is part of the WebM project developed by Google. Three techniques are developed in this paper: 1) An HDR video coding processing chain developed recently in standards committees (the Joint Collaborative Team on Video Coding, or JCT-VC, of ISO/IEC/ITU), but adapted to the VP10 codec, herein called HDR_VP10. 2) A modified version of HDR_VP10 with different intermediate color transforms and advanced sampling filters herein called FastVDO_ECHDR_VP10 and 3) FastVDO_HDR_VP10 that uses a new video data adaptive tuning process, which differs from the ST.2084 transfer function used in the anchor. HDR_VP10 is used as reference point for subjective and objective comparisons. Representative objective results for these systems include: (a) FastVDO_ECHDR_VP10, Overall results for RGB-PSNR, DE100, MD100, PSNRL100 were 8.0%, -3.5%, 46.2%, 2.9%; (b) FastVDO_HDR_VP10, Overall results for RGB-PSNR, DE100, MD100, PSNRL100 were -40.2%, -17.0%, -5.5% and -2.8% respectively. It is asserted (and for which there is appears to be fairly broad agreement) that such objective metrics are currently not very predictive of visual quality in HDR coding studies. FastVDO_HDR_VP10 shows a significant visual quality improvement on HDR_VP10.