BEGIN:VCALENDAR VERSION:2.0 PRODID:-//swoogo.com//NONSGML kigkonsult.se iCalcreator 2.41.90// CALSCALE:GREGORIAN UID:65336234-6632-4037-b834-336434653163 BEGIN:VEVENT UID:ec137e2d73b6c923ce06cb98a90bbb4be261ec65@swoogo.com DTSTAMP:20260813T002452Z DESCRIPTION:Can we halve the cost of encoding a movie without sacrificing i ts perceptual quality?\nThe HEVC video compression standard is an essentia l technology for providing UltraHD and HDR content\, and a driver for sign ificant increase of coding efficiency for HD content. This increase comes at a cost of a significant increase in computational complexity. Combined with increase of number of encodes due to the number of rungs in adaptive streaming ladder and to content-adaptive encoding\, use of HEVC in multi-r ate streaming context requires substantial increase in computational capac ity. This\, in turn\, translates into increase in infrastructure or public compute costs. \nThe main goal of this paper is to show what can be done to optimize the throughput of an open source distributed encoding system r unning on commodity hardware either on premises or in the public cloud. Th is will be done using the popular open source x265 video encoder. It is ub iquitous in commercial cloud encoding implementations and is typically use d in conjunction with ffmpeg. It is a versatile encoder which allows user control over various aspects such as use HEVC coding tools\, rate-distorti on optimizations (RDO)\, perceptual quality\, instruction set extensions\, and concurrency at a very fine granularity. It also outputs a plethora of internal information through its logging system. This level of detail\, c ombined with the source code availability\, lets us do a detailed tradeoff analysis at both quality and code path level. \nWe start the analysis fro m examining the cost-benefit analysis of x265 presets which recently went through an overhaul. We will continue with effects of lesser-known options controlling RDO behavior\, such as SSIM-based RDO. We further examine the effects of disabling or limiting some of the coding tools offered by the HEVC standard\, such as in-loop filtering. We then evaluate CPU-level aspe cts of x265 performance\, such as core frequency and utilization\, and com pare performance on latest generations of server processors\, both in on-p remise and public cloud settings. \nLastly\, we analyze tradeoffs of massi vely parallel chunked encoding\, where chunk duration and number of concur rent encoding jobs per machines are an additional optimization variable.\n Our evaluations are done on large amounts of UltraHD and HD cinematic cont ent. Quality and quality constancy are evaluated using modern perceptual q uality metrics such as VMAF and SSIM+. \nThe value of this paper is in the detailed analysis of the codec performance. This information is currently not readily available to practitioners. DTSTART:20191021T183000Z DTEND:20191021T190000Z LAST-MODIFIED:20260813T002452Z LOCATION:San Francisco Room SEQUENCE:0 STATUS:CONFIRMED SUMMARY:Speed-Distortion Optimization: Tradeoffs in Open Source HEVC Encodi ng TRANSP:OPAQUE X-ALT-DESC;FMTTYPE=text/html:Can we halve the cost of encoding a movie with out sacrificing its perceptual quality?
\nThe HEVC video compression standard is an essential technology for providing UltraHD and HDR content\ , and a driver for significant increase of coding efficiency for HD conten t. This increase comes at a cost of a significant increase in computationa l complexity. Combined with increase of number of encodes due to the numbe r of rungs in adaptive streaming ladder and to content-adaptive encoding\, use of HEVC in multi-rate streaming context requires substantial increase in computational capacity. This\, in turn\, translates into increase in i nfrastructure or public compute costs.
\nThe main goal of this pape r is to show what can be done to optimize the throughput of an open source distributed encoding system running on commodity hardware either on premi ses or in the public cloud. This will be done using the popular open sourc e x265 video encoder. It is ubiquitous in commercial cloud encoding implem entations and is typically used in conjunction with ffmpeg. It is a versat ile encoder which allows user control over various aspects such as use HEV C coding tools\, rate-distortion optimizations (RDO)\, perceptual quality\ , instruction set extensions\, and concurrency at a very fine granularity. It also outputs a plethora of internal information through its logging sy stem. This level of detail\, combined with the source code availability\, lets us do a detailed tradeoff analysis at both quality and code path leve l.
\nWe start the analysis from examining the cost-benefit analysis of x265 presets which recently went through an overhaul. We will continue with effects of lesser-known options controlling RDO behavior\, such as SS IM-based RDO. We further examine the effects of disabling or limiting some of the coding tools offered by the HEVC standard\, such as in-loop filter ing. We then evaluate CPU-level aspects of x265 performance\, such as core frequency and utilization\, and compare performance on latest generations of server processors\, both in on-premise and public cloud settings.
\nLastly\, we analyze tradeoffs of massively parallel chunked encoding\, where chunk duration and number of concurrent encoding jobs per machines are an additional optimization variable.
\nOur evaluations are done o n large amounts of UltraHD and HD cinematic content. Quality and quality c onstancy are evaluated using modern perceptual quality metrics such as VMA F and SSIM+.
\nThe value of this paper is in the detailed analysis of the codec performance. This information is currently not readily availa ble to practitioners.
END:VEVENT END:VCALENDAR