Isize Bitsave:High-Quality Video Streaming at Lower Bitrates Via

Total Page:16

File Type:pdf, Size:1020Kb

Isize Bitsave:High-Quality Video Streaming at Lower Bitrates Via White Paper iSIZE BitSave: High-Quality Video Streaming at Lower Bitrates via Amazon ECS Fast and easily integrated with any encoder including Streaming high-resolution video comes with an inevitable trade-off between AVC, HEVC, VP9 and AV1, without breaking standards or requiring any changes on end user device available bandwidth and quality-of-experience for the end user. Delivering iSIZE BitSave pre-processing is single-pass through the uncompromising video quality typically requires excessively high bitrates, content with 1 frame latency – it can be used with VoD, which can result in slow starts, video buffering and high CDN and live streaming, gaming, IoT storage costs. As the percentage of IP traffic attributed to high-resolution video iSIZE BitSave supports: multi-codec, multi-recipe, multi- increases, these problems are only exacerbated. bitrate and multi-resolution ABR ladders iSIZE deep neural network models improve significantly Current solutions that attempt to minimize bandwidth without compromising on with each new generation, providing increased gains with each new generation of pre-processing models quality are centered around the development of more intelligent video encoders; either by replacing parts, or the entirety of a standard video coding Next generation sustainable results and significant reduction in the energy consumption for video delivery pipeline. The latter, in particular, is a risky proposition for video encoding services, given the inevitable requirement for bespoke transport mechanisms and decoders or post-processors on client devices. On the other hand, any improvement to a standards-based coder is severely constrained by the need for compliance to the utilized standard. iSIZE’s innovative BitSave technology comprises an AI-based perceptual optimization that adaptively filters the video content before it reaches the video encoder (Fig. 1), such that the output after compressing with any iSIZE Technologies standard video codec is perceptually optimized (e.g., less motion artifacts or Russell Anam blurring), for the same bitrate. BitSave neural networks are able to isolate areas of perceptual importance, such as those with high motion or detailed texture, Ilya Fadeev and optimize their structure so that they are preserved better at lower bitrate by any subsequent encoder. As shown in our recent paper [1], when integrated as a Aaron Chadha preprocessor prior to a video encoder, BitSave is able to reduce bitrate Yiannis Andreopoulos requirements for a given quality level by up to 20% versus that encoder. When coupled with open-source AVC, HEVC and AV1 encoders, BitSave allows for up to 40% bitrate reduction at no quality compromise in comparison to leading third-party AVC, HEVC and AV1 encoding services on Amazon Web Services. Lower bitrates equate to smaller filesizes, lower storage and distribution costs, reduced energy consumption and more satisfied customers. iSIZE BitSave Storage in multiple Container containers like on Amazon ECS: Any External Any Streaming mp4, webm, Preprocessing only Video Encoder or CDN Service Apple QuickTime Source Content Preprocessed raw video AVC, HEVC or AV1 Reducing cost up to 20% iSIZE BitSave Storage in multiple Container on Amazon ECS: containers like Any Streaming Preprocessing and Encoding with mp4, webm, or CDN Service Open-Source Encoders Apple QuickTime Source Content AVC, HEVC or AV1 Reducing cost up to 40% Figure 1. iSIZE’s BitSave Container pipeline for preprocessing and encoding with third-party encoders (top) and preprocessing and encoding with open-source encoders (bottom) Importantly, given that BitSave pre-processes video content without changing the encoding, these gains are attainable without any of the disadvantages associated with replacing components of encoders, streaming infrastructures, or client devices. iSIZE BitSave: High-quality Video Streaming at Lower Bitrates via Amazon ECS iSIZE’s proprietary deep learning-based models can run on all BitSave is compatible with any existing codec (including CPU and GPU hardware. The current generation of BitSave MPEG AVC/H.264, HEVC/H.265, EVC, VVC and models is able to run on 4 CPU cores or higher, achieving AOMedia VP9, AV1, AV2), and supports all resolutions up real-time preprocessing and AVC encoding on a to 8K. The preprocessing accepts all video formats as g4dn.8xlarge AWS instance with 12% average bitrate input and outputs same-resolution raw Y4M, which can be saving versus x264 ‘slow’ preset and 33% average bitrate optionally downscaled, before passing to any standard saving versus a leading AWS quality-based variable video encoder. Only a single pass processing is bitrate (QVBR) encoding service. The remainder of this required (with single frame latency), prior to any white paper details the expected bitrate savings and speed number of subsequent encodings with any video benchmarks on a range of CPU and GPU-enabled AWS encoder. instances. Figure 2. BitSave preprocessing performs a perceptual and rate-constrained decomposition of each frame to ensure that the output of any standard codec is perceptually optimized for a given bitrate Bitrate savings on typical full-HD test (1080p) videos are We also compare against a widely-used and highly- reported in Table 1 below for the regime of 2 to 10 mbps. respected third-party encoding service on AWS. The More negative values indicate higher bitrate saving versus corresponding bitrate savings versus that benchmark are the corresponding video encoder (anchor) with the same reported in Table 2. All BitSave encodings were done preset and the same bitrate targets. The average bitrate with the settings shown in Table 1. All third-party saving across multiple encoders and VMAF, VMAF_NEG encodings were done with the recommended settings of and SSIM was 11%, with BitSave offering up to 20% that external service, including the use of its proprietary saving on difficult content (e.g., ‘ducks take off’, QVBR rate control. The results show that, against well- ‘touchdown pass’, ‘park joy’). Importantly, all encoded established third-party encoders, BitSave and open- bitstreams are fully standard compliant and do not require source encoders lead to bitrate savings between 17% to any changes in content packaging, delivery or client side. 40%. Figure 3 (next page) illustrates some examples of This means they can be decoded by any device that typical segments of sequences at the same bitrate, supports AVC, HEVC or AV1 decoding without requiring showcasing the quality improvement obtained when any downscaling or upscaling on the client device. utilizing BitSave. Bitrate Saving (%) over Anchor Encoder Bitrate Saving (%) over Leading QVBR Encoding Service on AWS Sequence AVC x264 HEVC x265 AV1 svt-av1 Sequence AVC HEVC AV1 ‘slow’ preset ‘slow’ preset (cpu=5) two-pass two-pass single-pass red kayak -8.8 -5.8 -6.5 red kayak -28.4 -17.5 -24.4 ducks take off -15.2 -16.2 -20.3 ducks take off -35.9 -39.4 -30.8 riverbed -16.3 -3.3 -6.0 riverbed -26.1 -19.6 -22.9 west wind easy -10.4 -0.1 -6.0 west wind easy -41.4 -17.6 -22.3 rush field cuts -8.6 -8.8 -9.3 rush field cuts -32.0 -21.1 -21.8 crowd run -11.0 -9.2 -11.0 crowd run -29.6 -32.2 -25.6 park joy -12.8 -15.1 -18.4 park joy -34.0 -41.5 -32.4 touchdown pass -12.9 -18.8 -17.2 touchdown pass -36.0 -27.7 -20.9 Average -12.0 -9.7 -11.8 Average -32.9 -27.1 -25.1 Table 1: Example BD-rates (%) of BitSave+anchor encoder versus Table 2: Representative BD-rates (%) of BitSave+encoder (x264, anchor encoder on 8 XIPH sequences, with the encoder being x264, x265, svt-av1) versus a market-leading VBR-based external encoding x265 and svt-av1 with bitrate targets set to [2,4,6,8,10] mbps, service for on 16 XIPH sequences, with bitrate targets and GOP GOP=200 for x264 and x265 and GOP=120 for svt-av1. All BD-rates sizes set as for Table 1. All BD-rates are average BD-rates for SSIM, are average BD-rates for the SSIM, VMAF_NEG and VMAF quality VMAF_NEG and VMAF. Higher negative values indicate higher metrics. Higher negative values indicate higher bitrate saving. bitrate saving. iSIZE BitSave: High-quality Video Streaming at Lower Bitrates via Amazon ECS Third-party AVC QVBR, 4.46mbps x264 slow, 4.15mbps BitSave + x264 slow, 4.15mbps VMAF=43.0, VMAF_NEG=41.8, SSIM=0.925 VMAF=50.2, VMAF_NEG=48.5, SSIM=0.927 VMAF=58.7, VMAF_NEG=54.9, SSIM=0.943 Third-party AVC QVBR, 2.12mbps x264 slow, 1.93mbps BitSave + x264 slow, 1.93mbps VMAF=50.4, VMAF_NEG=48.9, SSIM=0.940 VMAF=57.7, VMAF_NEG=55.8, SSIM=0.944 VMAF=62.4, VMAF_NEG=59.6, SSIM=0.947 Figure 3. Visual comparison of encodings corresponding to Table 1 and Table 2 (top: ‘crowd run’; bottom: ‘rush field cuts’). x264 and BitSave + x264 is at approx. 8% lower bitrate than the third-party AVC QVBR. Typical speed benchmarks obtained on various AWS and g4dn.8xlarge instances. Live processing and encoding instances are reported in Table 3. The g4 instances are is achieved for 1080p@50fps on a g4dn.8xlarge instance. GPU-enabled, while c5d is CPU-based. As shown in BitSave and HEVC and AV1 encoding remains at high Table 3, BitSave achieves over 50fps for 720p and over speed in comparison to third-party alternatives, while 24fps for 1080p processing on all instance types. When offering significant bitrate savings, as shown in Table 2. combined with x264 ‘slow’ preset, live processing and encoding is achieved for 720p@50fps on g4dn.4xlarge Frames-Per-Second (FPS) Workload c5d.12xlarge g4dn.4xlarge g4dn.8xlarge 720p BitSave only 60 267 267 720p BitSave + x264 slow 39 58 74 720p BitSave + x265 slow 12 9 11 720p BitSave + svt-av1 9 4 7 1080p BitSave only 26 115 115 1080p BitSave + x264 slow 19 36 56 1080p BitSave + x265 slow 8 6 9 1080p BitSave + svt-av1 5 2 4 Table 3: Runtime (FPS) for various AWS instances processing HD or full-HD input resolutions.
Recommended publications
  • An Effective Self-Adaptive Policy for Optimal Video Quality Over Heterogeneous Mobile Devices and Network Discovery Services
    Appl. Math. Inf. Sci. 13, No. 3, 489-505 (2019) 489 Applied Mathematics & Information Sciences An International Journal http://dx.doi.org/10.18576/amis/130322 An Effective Self-Adaptive Policy for Optimal Video Quality over Heterogeneous Mobile Devices and Network Discovery Services Saleh Ali Alomari∗1 , Mowafaq Salem Alzboon1, Belal Zaqaibeh1 and Mohammad Subhi Al-Batah1 1Faculty of Science and Information Technology, Jadara University, 21110 Irbid, Jordan Received: 12 Dec. 2018, Revised: 1 Feb. 2019, Accepted: 20 Feb. 2019 Published online: 1 May 2019 Abstract: The Video on Demand (VOD) system is considered a communicating multimedia system that can allow clients be interested whilst watching a video of their selection anywhere and anytime upon their convenient. The design of the VOD system is based on the process and location of its three basic contents, which are: the server, network configuration and clients. The clients are varied from numerous approaches, battery capacities, involving screen resolutions, capabilities and decoder features (frame rates, spatial dimensions and coding standards). The up-to-date systems deliver VOD services through to several devices by utilising the content of a single coded video without taking into account various features and platforms of a device, such as WMV9, 3GPP2 codec, H.264, FLV, MPEG-1 and XVID. This limitation only provides existing services to particular devices that are only able to play with a few certain videos. Multiple video codecs are stored by VOD systems store for a similar video into the storage server. The problems caused by the bandwidth overhead arise once the layers of a video are produced.
    [Show full text]
  • Download Media Player Codec Pack Version 4.1 Media Player Codec Pack
    download media player codec pack version 4.1 Media Player Codec Pack. Description: In Microsoft Windows 10 it is not possible to set all file associations using an installer. Microsoft chose to block changes of file associations with the introduction of their Zune players. Third party codecs are also blocked in some instances, preventing some files from playing in the Zune players. A simple workaround for this problem is to switch playback of video and music files to Windows Media Player manually. In start menu click on the "Settings". In the "Windows Settings" window click on "System". On the "System" pane click on "Default apps". On the "Choose default applications" pane click on "Films & TV" under "Video Player". On the "Choose an application" pop up menu click on "Windows Media Player" to set Windows Media Player as the default player for video files. Footnote: The same method can be used to apply file associations for music, by simply clicking on "Groove Music" under "Media Player" instead of changing Video Player in step 4. Media Player Codec Pack Plus. Codec's Explained: A codec is a piece of software on either a device or computer capable of encoding and/or decoding video and/or audio data from files, streams and broadcasts. The word Codec is a portmanteau of ' co mpressor- dec ompressor' Compression types that you will be able to play include: x264 | x265 | h.265 | HEVC | 10bit x265 | 10bit x264 | AVCHD | AVC DivX | XviD | MP4 | MPEG4 | MPEG2 and many more. File types you will be able to play include: .bdmv | .evo | .hevc | .mkv | .avi | .flv | .webm | .mp4 | .m4v | .m4a | .ts | .ogm .ac3 | .dts | .alac | .flac | .ape | .aac | .ogg | .ofr | .mpc | .3gp and many more.
    [Show full text]
  • Encoding H.264 Video for Streaming and Progressive Download
    What is Tuning? • Disable features that: • Improve subjective video quality but • Degrade objective scores • Example: adaptive quantization – changes bit allocation over frame depending upon complexity • Improves visual quality • Looks like “error” to metrics like PSNR/VMAF What is Tuning? • Switches in encoding string that enables tuning (and disables these features) ffmpeg –input.mp4 –c:v libx264 –tune psnr output.mp4 • With x264, this disables adaptive quantization and psychovisual optimizations Why So Important • Major point of contention: • “If you’re running a test with x264 or x265, and you wish to publish PSNR or SSIM scores, you MUST use –tune PSNR or –tune SSIM, or your results will be completely invalid.” • http://x265.org/compare-video-encoders/ • Absolutely critical when comparing codecs because some may or may not enable these adjustments • You don’t have to tune in your tests; but you should address the issue and explain why you either did or didn’t Does Impact Scores • 3 mbps football (high motion, lots of detail) • PSNR • No tuning – 32.00 dB • Tuning – 32.58 dB • .58 dB • VMAF • No tuning – 71.79 • Tuning – 75.01 • Difference – over 3 VMAF points • 6 is JND, so not a huge deal • But if inconsistent between test parameters, could incorrectly show one codec (or encoding configuration) as better than the other VQMT VMAF Graph Red – tuned Green – not tuned Multiple frames with 3-4-point differentials Downward spikes represent untuned frames that metric perceives as having lower quality Tuned Not tuned Observations • Tuning
    [Show full text]
  • Screen Capture Tools to Record Online Tutorials This Document Is Made to Explain How to Use Ffmpeg and Quicktime to Record Mini Tutorials on Your Own Computer
    Screen capture tools to record online tutorials This document is made to explain how to use ffmpeg and QuickTime to record mini tutorials on your own computer. FFmpeg is a cross-platform tool available for Windows, Linux and Mac. Installation and use process depends on your operating system. This info is taken from (Bellard 2016). Quicktime Player is natively installed on most of Mac computers. This tutorial focuses on Linux and Mac. Table of content 1. Introduction.......................................................................................................................................1 2. Linux.................................................................................................................................................1 2.1. FFmpeg......................................................................................................................................1 2.1.1. installation for Linux..........................................................................................................1 2.1.1.1. Add necessary components........................................................................................1 2.1.2. Screen recording with FFmpeg..........................................................................................2 2.1.2.1. List devices to know which one to record..................................................................2 2.1.2.2. Record screen and audio from your computer...........................................................3 2.2. Kazam........................................................................................................................................4
    [Show full text]
  • Spin Digital HEVC/H.265 Software Encoder (Spin Enc) Enables Ultra-High-Quality Video with the Highest Compression Level
    Spin Digital HEVC/H.265 software encoder (Spin Enc) enables ultra-high-quality video with the highest compression level. Encoding, transmission, and storage of video in 4K, 8K, and beyond are now possible with commodity computing technologies. HEVC/H.265 encoder for production, contribution, and distribution of professional video for broadcast, VoD, and creative studios. Spin Digital HEVC/H.265 encoder is ready for the next generation of high-quality video systems, providing support for Ultra HD (UHD), High Dynamic Range (HDR), High Frame Rate (HFR), Wide Color Gamut (WCG), and Virtual Reality (360° video). Product Highlights HEVC/H.265 Encoder Package • Fast offline encoding software solution • Encoder: standalone HEVC/H.265 encoder • Ready for 8K and beyond • Transcoder: HEVC/H.265 decoder and encoder • Significantly better compression and quality integrated in FFmpeg than competing encoders • SDK: encoder plugin for FFmpeg (Libavcodec) • Enables WCG and HDR with up to 12-bit video • Compatible with ARIB STD-B32 standard • Versatile high-precision pre-processing filters • Preserves color resolution with 4:2:2, 4:4:4, and RGB formats • 22.2-ch AAC audio encoding and decoding SPIN DIGITAL HEVC/H.265 ENCODER Support for the HEVC standard: Main and Main 10 profiles Range Extensions (HEVCv2) profiles ARIB STD-B32 version 3.9 Resolutions: 4K, 8K, and beyond Color formats: 4:2:0, 4:2:2, 4:4:4, RGB Bit depths: 8-, 10-, 12-bit Color spaces: BT.601, BT.709, DCI-P3, BT.2020 HDR support: ST2084 transfer function, ST2086 HDR metadata, HLG Coding
    [Show full text]
  • EMA Mezzanine File Creation Specification and Best Practices Version 1.0.1 For
    16530 Ventura Blvd., Suite 400 Encino, CA 91436 818.385.1500 www.entmerch.org EMA Mezzanine File Creation Specification and Best Practices Version 1.0.1 for Digital Audio‐Visual Distribution January 7, 2014 EMA MEZZANINE FILE CREATION SPECIFICATION AND BEST PRACTICES The Mezzanine File Working Group of EMA’s Digital Supply Chain Committee developed the attached recommended Mezzanine File Specification and Best Practices. Why is the Specification and Best Practices document needed? At the request of their customers, content providers and post‐house have been creating mezzanine files unique to each of their retail partners. This causes unnecessary costs in the supply chain and constrains the flow of new content. There is a demand to make more content available for digital distribution more quickly. Sales are lost if content isn’t available to be merchandised. Today’s ecosystem is too manual. Standardization will facilitate automation, reducing costs and increasing speed. Quality control issues slow down today’s processes. Creating one standard mezzanine file instead of many files for the same content should reduce the quantity of errors. And, when an error does occur and is caught by a single customer, it can be corrected for all retailers/distributors. Mezzanine File Working Group Participants in the Mezzanine File Working Group were: Amazon – Ben Waggoner, Ryan Wernet Dish – Timothy Loveridge Google – Bill Kotzman, Doug Stallard Microsoft – Andy Rosen Netflix – Steven Kang , Nick Levin, Chris Fetner Redbox Instant – Joe Ambeault Rovi
    [Show full text]
  • Home for Christmas S01E01 INTERNAL 720P WEB X264strife Eztv Thepiratebay
    1 / 2 Home For Christmas S01E01 INTERNAL 720p WEB X264-STRiFE [eztv] - ThePirateBay http://nextisp.com/index.php/Home/Drama/detail/p/a-little-love-never-hurts/uid/0.html - A ... [url=http://thebeststar.us/torrent/1663028501/UFC+210+Prelims+720p+WEB-DL+ ... in Fire S04E09 The Charay iNTERNAL 720p HDTV x264-DHD[ettv][/url] ... [url=http://almuzaffar.org/christmas-is-here-again-movie]Christmas Is Here .... Us.S01.COMPLETE.720p.WEB.x264-GalaxyTV2019-05-31 VIP 1.22 GiB 285 sotnikam · Video > HD MoviesThey.Shall.Not.Grow.Old.2018.1080p.BluRay.H264.. Download Chloe Neill - Chicagoland Vampires and Dark Elite torrent or any other torrent from the Other E-books. Direct download via magnet link.Missing: Home ‎Christmas ‎S01E01 ‎INTERNAL ‎WEB ‎x264- ‎[eztv] -. ... i ian dervish · file 3245112 trailer.park.boys.out.of.the.park.s02e01.720p.web.x264 strife ... file 3086561 unearthed.2016.s02e05.internal.720p.hevc.x265 megusta ... christmas cookie challenge s03e05 colors of christmas 480p x264 msd eztv ... 720p bluray yts yify · file 817715 miami.monkey.s01e01.480p.hdtv.x264 msd .... Nov 15, 2014 — Ninja x men 1 720p dual audio Любовь РїРѕРґ ... Home And Away S23E93 WS PDTV XviD-AUTV kapitein rob en het geheim van ... й”法提琴手 Нанолюбовь (10) web dl inside amy schumer ... Low Winter Sun S01E01 2013 HDTV x264 EZN ettv Mad Men (Season 06 Episode .... Home for Christmas S01E01 INTERNAL WEB x264-STRiFE EZTV torrent download - download for free Home for Christmas S01E01 INTERNAL WEB .... ... file 4554555 brave.new.world.s01e07.internal.1080p.hevc.x265 megusta eztv ..
    [Show full text]
  • On Engagement with ICT Standards and Their Implementations in Open Source Software Projects: Experiences and Insights from the Multimedia Field
    International Journal of Standardization Research Volume 19 • Issue 1 On Engagement With ICT Standards and Their Implementations in Open Source Software Projects: Experiences and Insights From the Multimedia Field Jonas Gamalielsson, University of Skövde, Sweden Björn Lundell, University of Skövde, Sweden ABSTRACT The overarching goal in this paper is to investigate organisational engagement with an ICT standard and open source software (OSS) projects that implement the standard, with a specific focus on the multimedia field, which is relevant in light of the wide deployment of standards and different legal challenges in this field. The first part reports on experiences and insights from engagement with standards in the multimedia field and from implementation of such standards in OSS projects. The second part focuses on the case of the ITU-T H.264 standard and the two OSS projects OpenH264 and x264 that implement the standard, and reports on a characterisation of organisations that engage with and control the H.264 standard, and organisations that engage with and control OSS projects implementing the H.264 standard. Further, projects for standardisation and implementation of H.264 are contrasted with respect to mix of contributing organisations, and findings are related to organisational strategies of contributing organisations and previous research. KEywordS AVC, H.264, Involvement, ISO, ITU-T, OpenH264, Participation, x264 1 INTROdUCTION There are a number of different challenges related to provision of standards in the software sector, that can impact on the extent to which it is possible to faithfully implement the specification of a standard in software systems (Blind and Böhm, 2019; Gamalielsson and Lundell, 2013; Lundell et al., 2019; UK, 2015).
    [Show full text]
  • Format Support
    Episode 6 Format Support FILE FORMAT CODEC Episode Episode Episode Pro EngineCOMMENTS Adaptive bitrate streaming Microsoft Smooth Streaming H.264 (AAC audio) O Windows OS only. Available with Episode Engine License. Apple HLS H.264 (AAC audio) O Available with Episode Engine License. Windows Media WMV, ASF VC-1 O O O WM9 I/O I/O I/O WMV7 and 8 through F4M component on Mac WMA I/O I/O I/O WMA Pro I/O I/O I/O Flash FLV Flash 8 (VP6s/VP6e) I/O I/O I/O SWF Flash 8 (VP6s/VP6e) I/O I/O I/O MOV/MP4/F4V Flash 9 (H.264) I/O I/O I/O F4V as extension to MP4 WebM WebM VP8 O O O Vorbis O O O 3GPP 3GPP AAC I/O I/O I/O H.263 I/O I/O I/O H.264 I/O I/O I/O MainConcept and x264 MPEG-4 I/O I/O I/O 3GPP2 3GPP2 AAC I/O I/O I/O H.263 I/O I/O I/O H.264 I/O I/O I/O MainConcept and x264 MPEG-4 I/O I/O I/O MPEG Elementary Streams MPEG-1 Elementary Stream MPEG-1 (video) I/O I/O I/O MPEG-2 Elementary Stream MPEG-2 I/O I/O I/O MPEG Program Streams PS AAC O O O MainConcept and x264 H.264 I/O I/O I/O MPEG-1/2 (audio) I/O I/O I/O MPEG-2 I/O I/O I/O MPEG-4 I/O I/O I/O MPEG Transport Streams TS AAC I O O AES I I/O I/O H.264 I I/O I/O MainConcept and x264 AVCHD I I I HDV I I/O I/O MPEG - 1/2 (audio) I I/O I/O MPEG - 2 I I/O I/O MPEG - 4 I I/O I/O PCM I I I Matrox MAX H.264 I/O I/O I/O QT codec (*output possible via QT), Requires Matrox MAX hardware - Mac OS X only MPEG System Streams M1A MPEG-1 (audio) I/O I/O I/O M1V MPEG-1 (audio) I/O I/O I/O Episode 6 Format Support Format Support FILE FORMAT CODEC Episode Episode Episode Pro EngineCOMMENTS MPEG-4 MP4 AAC I/O I/O I/O
    [Show full text]
  • Ffmpeg Codecs Documentation Table of Contents
    FFmpeg Codecs Documentation Table of Contents 1 Description 2 Codec Options 3 Decoders 4 Video Decoders 4.1 hevc 4.2 rawvideo 4.2.1 Options 5 Audio Decoders 5.1 ac3 5.1.1 AC-3 Decoder Options 5.2 flac 5.2.1 FLAC Decoder options 5.3 ffwavesynth 5.4 libcelt 5.5 libgsm 5.6 libilbc 5.6.1 Options 5.7 libopencore-amrnb 5.8 libopencore-amrwb 5.9 libopus 6 Subtitles Decoders 6.1 dvbsub 6.1.1 Options 6.2 dvdsub 6.2.1 Options 6.3 libzvbi-teletext 6.3.1 Options 7 Encoders 8 Audio Encoders 8.1 aac 8.1.1 Options 8.2 ac3 and ac3_fixed 8.2.1 AC-3 Metadata 8.2.1.1 Metadata Control Options 8.2.1.2 Downmix Levels 8.2.1.3 Audio Production Information 8.2.1.4 Other Metadata Options 8.2.2 Extended Bitstream Information 8.2.2.1 Extended Bitstream Information - Part 1 8.2.2.2 Extended Bitstream Information - Part 2 8.2.3 Other AC-3 Encoding Options 8.2.4 Floating-Point-Only AC-3 Encoding Options 8.3 flac 8.3.1 Options 8.4 opus 8.4.1 Options 8.5 libfdk_aac 8.5.1 Options 8.5.2 Examples 8.6 libmp3lame 8.6.1 Options 8.7 libopencore-amrnb 8.7.1 Options 8.8 libopus 8.8.1 Option Mapping 8.9 libshine 8.9.1 Options 8.10 libtwolame 8.10.1 Options 8.11 libvo-amrwbenc 8.11.1 Options 8.12 libvorbis 8.12.1 Options 8.13 libwavpack 8.13.1 Options 8.14 mjpeg 8.14.1 Options 8.15 wavpack 8.15.1 Options 8.15.1.1 Shared options 8.15.1.2 Private options 9 Video Encoders 9.1 Hap 9.1.1 Options 9.2 jpeg2000 9.2.1 Options 9.3 libkvazaar 9.3.1 Options 9.4 libopenh264 9.4.1 Options 9.5 libtheora 9.5.1 Options 9.5.2 Examples 9.6 libvpx 9.6.1 Options 9.7 libwebp 9.7.1 Pixel Format 9.7.2 Options 9.8 libx264, libx264rgb 9.8.1 Supported Pixel Formats 9.8.2 Options 9.9 libx265 9.9.1 Options 9.10 libxvid 9.10.1 Options 9.11 mpeg2 9.11.1 Options 9.12 png 9.12.1 Private options 9.13 ProRes 9.13.1 Private Options for prores-ks 9.13.2 Speed considerations 9.14 QSV encoders 9.15 snow 9.15.1 Options 9.16 vc2 9.16.1 Options 10 Subtitles Encoders 10.1 dvdsub 10.1.1 Options 11 See Also 12 Authors 1 Description# TOC This document describes the codecs (decoders and encoders) provided by the libavcodec library.
    [Show full text]
  • AVC, HEVC, VP9, AVS2 Or AV1? — a Comparative Study of State-Of-The-Art Video Encoders on 4K Videos
    AVC, HEVC, VP9, AVS2 or AV1? | A Comparative Study of State-of-the-art Video Encoders on 4K Videos Zhuoran Li, Zhengfang Duanmu, Wentao Liu, and Zhou Wang University of Waterloo, Waterloo, ON N2L 3G1, Canada fz777li,zduanmu,w238liu,[email protected] Abstract. 4K, ultra high-definition (UHD), and higher resolution video contents have become increasingly popular recently. The largely increased data rate casts great challenges to video compression and communication technologies. Emerging video coding methods are claimed to achieve su- perior performance for high-resolution video content, but thorough and independent validations are lacking. In this study, we carry out an in- dependent and so far the most comprehensive subjective testing and performance evaluation on videos of diverse resolutions, bit rates and content variations, and compressed by popular and emerging video cod- ing methods including H.264/AVC, H.265/HEVC, VP9, AVS2 and AV1. Our statistical analysis derived from a total of more than 36,000 raw sub- jective ratings on 1,200 test videos suggests that significant improvement in terms of rate-quality performance against the AVC encoder has been achieved by state-of-the-art encoders, and such improvement is increas- ingly manifest with the increase of resolution. Furthermore, we evaluate state-of-the-art objective video quality assessment models, and our re- sults show that the SSIMplus measure performs the best in predicting 4K subjective video quality. The database will be made available online to the public to facilitate future video encoding and video quality research. Keywords: Video compression, quality-of-experience, subjective qual- ity assessment, objective quality assessment, 4K video, ultra-high-definition (UHD), video coding standard 1 Introduction 4K, ultra high-definition (UHD), and higher resolution video contents have en- joyed a remarkable growth in recent years.
    [Show full text]
  • MPEG-4 AVC/H.264 Video Codecs Comparison Short Version of Report
    MPEG-4 AVC/H.264 Video Codecs Comparison Short version of report Project head: Dr. Dmitriy Vatolin Measurements, analysis: Dr. Dmitriy Kulikov, Alexander Parshin Codecs: DivX H.264 Elecard H.264 Intel® MediaSDK AVC/H.264 MainConcept H.264 Microsoft Expression Encoder Theora x264 XviD (MPEG-4 ASP codec) April 2010 CS MSU Graphics&Media Lab Video Group http://www.compression.ru/video/codec_comparison/index_en.html [email protected] VIDEO MPEG-4 AVC/H.264 CODECS COMPARISON MOSCOW, APR 2010 CS MSU GRAPHICS & MEDIA LAB VIDEO GROUP SHORT VERSION Contents Contents .................................................................................................................... 2 1 Acknowledgments ............................................................................................. 4 2 Overview ........................................................................................................... 5 2.1 Sequences .....................................................................................................................5 2.2 Codecs ...........................................................................................................................5 2.3 Objectives and Testing Rules ........................................................................................6 2.4 H.264 Codec Testing Objectives ...................................................................................6 2.5 Testing Rules .................................................................................................................6
    [Show full text]