Get the most recent info and news about Vidio on HackerNoon, where 10k+ technologists publish stories for 4M+ monthly readers.

Vidio began as a service to empower Emtek's media portfolio, but it has since evolved into a platform that supports SCM, the KapanLagi network, and hundreds of Indonesian content producers. Vidio is one of Indonesia's most popular news and entertainment websites, with live streaming channels, critically acclaimed original content, real-time audience involvement, and talent spotting competitions. Users can enjoy live streaming, watch TV and anytime, as well as other exclusive shows on Vidio.com anytime and anywhere. Continuous Improvement, Open and Collaborative Environment, Embrace Challenges, Passionate for Excellence, and Work-Life Harmony are five values that have always been our guiding principles and that we implement in our work processes at Vidio. Together with the purpose of becoming Indonesia's largest OTT platform, offering a diverse range of content such as live streaming channels, news, entertainment, live sports, and the most comprehensive soap operas. Of course, we want to provide our viewers with high-quality local and premium content that is tailored around any of Our Values. Let's join Vidio and grow together!

Vidio

vidio

Siaran Langsung SPOTV 2 (Men&#x27;s Singles: Jonatan Christie (INA) vs Cho Geon Yeop (KOR) | Thomas Cup Quarter Finals - TotalEnergies BWF Thomas &amp; Uber Cup) ● Nonton Live Streaming, Highlights &amp; Jadwal Pertandingan hari ini di <strong>Vidio</strong>.

Live Streaming Men's Singles: Jonatan Christie (INA) vs Cho Geon Yeop (KOR) | Thomas Cup Quarter Finals - TotalEnergies BWF Thomas & Uber Cup - 3 Mei 2024 | Vidio

Nonton Takdir Lonceng Cinta - 03 Mei 2024 di <strong>Vidio</strong>. Streaming

Takdir Lonceng Cinta - 03 Mei 2024 | Vidio

B2B Tools You Need To Have in Your Arsenal in 2024

Video is everywhere. From repurposed image sensors in bird feeders to the seamless integration of connected doorbells, the addition of video to everyday devices is fundamentally reshaping surveillance. As a result, video is no longer a primary feature but rather a secondary attribute that permeates the Internet of Things (IoT). The implications of this transformation are profound, urging the security industry to adapt to a landscape where surveillance cameras are distributed across devices with various functions. In turn, this shift ushers in a new era of consumer demands, driving product revamps and evolving opportunities within the market. Of course, with more cameras in the modern home and office, safely navigating this trend requires careful consideration. Let’s consider the path ahead for security professionals with five key insights into the IoT video revolution. The Evolution of Surveillance We’ve come a long way from CCTV. In the early days of surveillance, fixed cameras were singular analog tools. Advancements led to the creation of digital multiplexing, which brought about motion recorders and time-lapse capabilities. Fast forward and the market continued to innovate. Today, surveillance encompasses everything from Internet Protocol (IP) cameras to cloud-based storage, remote alerts, and smart interconnectivity. This story isn’t static and the advent of video-enabled IoT is the next chapter. Thanks to the upcycling of video sensors from older mobile phones, surveillance increasingly comes in all shapes and sizes. This opens up a world of possibilities for surveillance, automation, and connectivity that promises to redefine the years to come. Beyond Cameras Smart doorbells, pet cameras, baby monitors: surveillance devices are no longer just cameras. Inversely, regular devices now count surveillance features. And the market is more diverse and lucrative as a result. This first product – the smart doorbell – is an example of this in action. Part surveillance and part convenience, this product does far more than alert occupants to the arrival of visitors. It monitors the front yard, enables two-way communication, and adds a layer of home security. Ten years after entering the market, smart doorbells are worth more than $3B. By the end of the decade, the segment is expected to grow 17% per year to a valuation of $11B. The expansion of surveillance into devices represents an expansion of market opportunities for security companies. In 2026, for example, the global surveillance market is estimated to grow to $54B, more than double the size of 2019. Therefore, this is a lucrative opportunity for those willing to pivot. To do so successfully, however, demands a keen focus on research and development (R&D) and versatile skills. A Fusion of Skills Technically, there are important considerations in offering video-enabled devices. This is because incorporating video into devices goes far beyond simply adding a camera. Rather, it necessitates the development of an all-in-one system that not only captures image streams but also processes the data on-board. Companies in this space quickly realize that crafting such devices requires specialized components. For instance, real-time operating systems may struggle to efficiently process and transmit video data, often necessitating a shift to a high-end operating system like Linux. Yet, this transition can introduce energy considerations, especially for battery-powered devices. Further, video encoding and compression is no mean feat. Given that video significantly multiplies data volumes, developers must also incorporate chipset add-ons to encrypt the imagery for secure transmission. In essence, creating video-capable devices is a formidable task, rife with software, hardware, and development challenges. In addition to these hurdles, companies will also need to develop companion apps and platforms that facilitate user-device interaction while prioritizing data transmission, security, and privacy. This introduces complexity on top of complexity. These obstacles underscore the need for companies developing devices in this direction to possess versatile skills (and plenty of patience). The Vendor's Dilemma Therefore, security companies must consider if this evolution is right for them. Of course, not every operator will find it necessary or feasible to develop video-enabled devices. Leaders must carefully evaluate the relevance of this shift within the context of their existing product lines. This evaluation process involves consideration of specific use cases and the flexibility to adapt as needed. Incorporating video into devices shouldn’t be a hasty response to passing trends. Rather, it should align with a well-founded business and technological strategy. It’s also crucial to recognize that “device vision” doesn’t invariably translate to the inclusion of traditional cameras. Alternative technologies such as Lidar sensors, which determine distances using laser targeting and measuring the return times of reflected light, offer viable options. These sensors can be more cost-effective and less invasive for certain use cases. Ultimately, the sector must remain adaptable and discerning when addressing the market’s ever-evolving demands. Security in Flux Finally, we must address the cybersecurity elephant in the room. If cameras are in everything, and those devices connect to the internet, the reality is that hackers have a larger attack surface to exploit. And, a particularly sensitive attack surface at that. Unfortunately, hackers can and will target companies that don’t adequately protect themselves or their customers. This happened earlier in the year when this surveillance camera manufacturer recalled almost 5M cameras with easy-to-guess passwords. Security companies venturing into connected video devices face a dual responsibility. First, they must achieve robust user authentication. This is possible by employing solutions like public key infrastructure (PKI) for cryptographic key-based security. This method works to strengthen resistance against brute force attacks and eliminate the vulnerability of shared secrets in the event of a breach. Additionally, secure data transfer is vital. Peer-to-peer connections that eschew intermediary servers in favor of end-to-end encrypted transmission best achieve this. Not only does this connection type enhance cybersecurity, but it also offers cost savings by avoiding cloud service fees based on file size, an important detail when handling high-quality video. The IoT video revolution is transforming surveillance and security, offering new growth opportunities As video sensors become integral to an array of devices, executives must weigh the benefits carefully. This expansion requires a fusion of hardware and software skills, robust cybersecurity, and considerable R&D. Flexibility and adaptability are crucial to meet evolving market demands. It’s a significant shift, but one that must be approached cautiously to ensure both success and safety. Video is everywhere. From repurposed image sensors in bird feeders to the seamless integration of connected doorbells, the addition of video to everyday devices is fundamentally reshaping surveillance. As a result, video is no longer a primary feature but rather a secondary attribute that permeates the Internet of Things (IoT). The implications of this transformation are profound, urging the security industry to adapt to a landscape where surveillance cameras are distributed across devices with various functions. In turn, this shift ushers in a new era of consumer demands, driving product revamps and evolving opportunities within the market. Of course, with more cameras in the modern home and office, safely navigating this trend requires careful consideration. Let’s consider the path ahead for security professionals with five key insights into the IoT video revolution. The Evolution of Surveillance We’ve come a long way from CCTV. In the early days of surveillance, fixed cameras were singular analog tools. Advancements led to the creation of digital multiplexing, which brought about motion recorders and time-lapse capabilities. Fast forward and the market continued to innovate. Today, surveillance encompasses everything from Internet Protocol (IP) cameras to cloud-based storage, remote alerts, and smart interconnectivity. This story isn’t static and the advent of video-enabled IoT is the next chapter. Thanks to the upcycling of video sensors from older mobile phones, surveillance increasingly comes in all shapes and sizes. This opens up a world of possibilities for surveillance, automation, and connectivity that promises to redefine the years to come. Beyond Cameras Smart doorbells, pet cameras, baby monitors: surveillance devices are no longer just cameras. Inversely, regular devices now count surveillance features. And the market is more diverse and lucrative as a result. This first product – the smart doorbell – is an example of this in action. Part surveillance and part convenience, this product does far more than alert occupants to the arrival of visitors. It monitors the front yard, enables two-way communication, and adds a layer of home security. Ten years after entering the market, smart doorbells are worth more than $3B. By the end of the decade, the segment is expected to grow 17% per year to a valuation of $11B. worth worth The expansion of surveillance into devices represents an expansion of market opportunities for security companies. In 2026, for example, the global surveillance market is estimated to grow to $54B, more than double the size of 2019. Therefore, this is a lucrative opportunity for those willing to pivot. To do so successfully, however, demands a keen focus on research and development (R&D) and versatile skills. estimated estimated A Fusion of Skills Technically, there are important considerations in offering video-enabled devices. This is because incorporating video into devices goes far beyond simply adding a camera. Rather, it necessitates the development of an all-in-one system that not only captures image streams but also processes the data on-board. Companies in this space quickly realize that crafting such devices requires specialized components. For instance, real-time operating systems may struggle to efficiently process and transmit video data, often necessitating a shift to a high-end operating system like Linux. Yet, this transition can introduce energy considerations, especially for battery-powered devices. Further, video encoding and compression is no mean feat. Given that video significantly multiplies data volumes, developers must also incorporate chipset add-ons to encrypt the imagery for secure transmission. In essence, creating video-capable devices is a formidable task, rife with software, hardware, and development challenges. In addition to these hurdles, companies will also need to develop companion apps and platforms that facilitate user-device interaction while prioritizing data transmission, security, and privacy. This introduces complexity on top of complexity. These obstacles underscore the need for companies developing devices in this direction to possess versatile skills (and plenty of patience). The Vendor's Dilemma Therefore, security companies must consider if this evolution is right for them. Of course, not every operator will find it necessary or feasible to develop video-enabled devices. Leaders must carefully evaluate the relevance of this shift within the context of their existing product lines. This evaluation process involves consideration of specific use cases and the flexibility to adapt as needed. Incorporating video into devices shouldn’t be a hasty response to passing trends. Rather, it should align with a well-founded business and technological strategy. It’s also crucial to recognize that “device vision” doesn’t invariably translate to the inclusion of traditional cameras. Alternative technologies such as Lidar sensors, which determine distances using laser targeting and measuring the return times of reflected light, offer viable options. These sensors can be more cost-effective and less invasive for certain use cases. Ultimately, the sector must remain adaptable and discerning when addressing the market’s ever-evolving demands. Security in Flux Finally, we must address the cybersecurity elephant in the room. If cameras are in everything, and those devices connect to the internet, the reality is that hackers have a larger attack surface to exploit. And, a particularly sensitive attack surface at that. Unfortunately, hackers can and will target companies that don’t adequately protect themselves or their customers. This happened earlier in the year when this surveillance camera manufacturer recalled almost 5M cameras with easy-to-guess passwords. this surveillance camera manufacturer this surveillance camera manufacturer Security companies venturing into connected video devices face a dual responsibility. First, they must achieve robust user authentication. This is possible by employing solutions like public key infrastructure (PKI) for cryptographic key-based security. This method works to strengthen resistance against brute force attacks and eliminate the vulnerability of shared secrets in the event of a breach. Additionally, secure data transfer is vital. Peer-to-peer connections that eschew intermediary servers in favor of end-to-end encrypted transmission best achieve this. Not only does this connection type enhance cybersecurity, but it also offers cost savings by avoiding cloud service fees based on file size, an important detail when handling high-quality video. The IoT video revolution is transforming surveillance and security, offering new growth opportunities As video sensors become integral to an array of devices, executives must weigh the benefits carefully. This expansion requires a fusion of hardware and software skills, robust cybersecurity, and considerable R&D. Flexibility and adaptability are crucial to meet evolving market demands. It’s a significant shift, but one that must be approached cautiously to ensure both success and safety.

What The IoT Video Revolution Means For Security

Enhancing Digital Security with AI Image and Video Detectors

I still remember my first video made with Windows Movie Maker back in the early 2000s (I'm feeling old)... editing those shaky family home videos, filmed by the only family member who had the only phone with an actual camera feature (a VGA one, but still). It was more of a slide show than an actual video. Video editing has changed a lot since its early days. If you don't edit videos regularly or have only used simple filters on Instagram and Snapchat, you might not realize how much the technology has improved. Let's jump right into what the future holds for video editing and every wannabe Spielbergs out there. AI-U Ready for This? The first extra large slice of the future 'video editing pie' has to be AI, our tech friend who is more than ready to make all our editing life just a tad simpler. Forget about spending hours sorting through an entire day's shoot because sophisticated algorithms have arrived on the scene that can: Pick out the best moments from the footage
Create rough edits (automatically!) based on your set parameters
Smoothen out those pixels with color correction and noise reduction Next time you're working in Adobe Premiere Pro or Final Cut Pro X, let the AI do its thing and let you flex your creative muscles where they matter. For example, adding cinematic slow pans to your drone shots rather than wasting time trying to fix a poorly lit frame. Virtual Reality: It's Here and It's Unreal It isn't a new trend by any stretch, but it is gaining momentum. Now, you can immerse viewers in 360° videos that make them feel as if they've been transported to another universe entirely, right from the comfort of their couch. Editing these immersive visuals requires specialized tools, techniques, and a broader imagination to truly harness the potential of full 360-degree spatial awareness. The Augmented Side of Things Having a Pokemon suddenly pop up in your living room was not only an exciting moment for all those 90s kids but also a step towards how AR has leveled up from its sci-fi roots. Video editing fluidly merging with reality is a major big deal! With AR, editors now have the chance to blend virtual elements into real-world footage seamlessly. Think of showcasing real-estate properties enhanced with virtual furniture or adding special effects that look almost palpable. Web Browser Video Editors Not all of us have a powerful computer for heavy-duty video editing software, and that's okay. For any amateur videographer or someone who's not yet ready to commit to professional editing software, web browser video editors offer a glimmer of hope. With a web browser video editor, there's no need for local file storage or endless rendering times; everything's processed in the cloud. Plus, forget about losing work due to system crashes (been there). All edits are saved instantly online. Social Media Ready Videos Now, this one isn't a prediction; it's happening already and will only continue to surge. Each social platform is unique in its video formats, dimensions, and lengths (TikTok's split-second videos are surely different from YouTube’s longer cuts). In line with that, the future of video editing has to keep up with these ever-changing demands. Video editors will increasingly cater to producing 'social media ready' clips– square for Instagram posts, vertical for IG and Snapchat stories, shorter versions for Twitter blasts... well, you get the drift. AI Voice-Overs and Subtitling I'm 99% sure that you've already seen a YouTube video with an AI voice that you mistook for a human narrator (Yep, it happens!). Although it happens to be a work in progress, AI voice-overs are starting to sound uncannily human. They're streamlined and time-efficient and provide you with numerous tonal options without having to hire someone to get behind the mic. AI is also sprinting ahead in perfectly syncing subtitles with speech (regardless of speed or delivery). Think about the amount of time we'll save not having to adjust the timing manually. Yes, accuracy still needs some fine-tuning, but once perfected? There's gonna be a drastic drop in many of our work hours. Text-Based Editing Can't be bothered to scrub through timelines? Text-based video editing is an emerging technology that just might simplify our lives in the near future. Simply jump directly to parts of a clip featuring specific words or phrases, just like finding instances of text in a Word document (Ctrl+F, anyone?) Yes, that's exactly what we're moving towards: editing via transcription. Want to take out a speaker from that mammoth two-hour seminar who used the phrase 'paradigm shift' one too many times? Find it and snip it just like that. Deep Fakes and Face Swaps This aspect tends to sit on the fine line between cool tech breakthroughs and ethically problematic uses. Advanced software can now replace faces completely from one person onto another's body in a video. This technology has been a hot topic, used in comedic fashion or controversial ways, and undeniably demonstrates the growing capabilities of video editing. While this area also needed to be regulated strictly for ethical use, we can't deny the wow factor of it in terms of technological progression. Interactive Video Editing Synchronize your exercise videos with your smartwatch or turn that cooking tutorial into an interactive experience by syncing instructions directly to your smart kitchen (yes, it exists). These are just a few of the exciting possibilities that interactive video editing offers. You can create videos that respond to user actions, perform actions like branching (choosing different paths in the video), shop directly from a product review video, or even change elements within the videos based on audience preferences. Fusing these 'Choose-Your-Own-Adventure' mechanics with movies and gaming will open up possibilities for whole new forms of media content. Can we just take a moment to picture how Netflix's interactive episode of Black Mirror: Bandersnatch was created? What About the Video Editors? We can't talk about the future of video editing without discussing those behind the lens, our video editors. No matter how much technology evolves, a human touch will always be essential. The intricate decisions that lead to storytelling and evoking feelings are often too complex (at least for now) for AI or software alone. However, this surge in tech doesn’t mean unemployment is looming on the horizon for all creative minds out there working as video editors...although there might be a slight shift in the required skill set. Surely being proficient with Adobe After Effects won't be enough, editors will need to develop new skills in AI programming, 360-degree video editing, AR and VR content creation, and managing cloud-based editing suites. Even with all these advancements, the need for human intuition, decision-making skills, and, ultimately, creativity will never lose their charm or importance. Remember: technology is an assistant, not a substitute for human creativity. Keep Calm and Click "Export" So, is it time to press the panic button with all these accelerating changes? Not at all. Trust me; you're not going to wake up tomorrow and suddenly have to edit a 360-degree video using AI tools in a virtual reality environment. (If that's happening, please tell me where you work—I want in.) These changes are ongoing; they're evolutionary, not revolutionary. Steps towards these technologies have been smoothly integrated into our workflows for years now – and just like how we made the leap from VCR to DVD or from landline to smartphone, we'll take all of this in stride as well. I still remember my first video made with Windows Movie Maker back in the early 2000s (I'm feeling old)... editing those shaky family home videos, filmed by the only family member who had the only phone with an actual camera feature (a VGA one, but still). It was more of a slide show than an actual video. Video editing has changed a lot since its early days. If you don't edit videos regularly or have only used simple filters on Instagram and Snapchat, you might not realize how much the technology has improved. Let's jump right into what the future holds for video editing and every wannabe Spielbergs out there. AI-U Ready for This? The first extra large slice of the future 'video editing pie' has to be AI, our tech friend who is more than ready to make all our editing life just a tad simpler. Forget about spending hours sorting through an entire day's shoot because sophisticated algorithms have arrived on the scene that can: Pick out the best moments from the footage Create rough edits (automatically!) based on your set parameters Smoothen out those pixels with color correction and noise reduction Pick out the best moments from the footage Create rough edits (automatically!) based on your set parameters Smoothen out those pixels with color correction and noise reduction Next time you're working in Adobe Premiere Pro or Final Cut Pro X, let the AI do its thing and let you flex your creative muscles where they matter. For example, adding cinematic slow pans to your drone shots rather than wasting time trying to fix a poorly lit frame. Virtual Reality: It's Here and It's Unreal It isn't a new trend by any stretch, but it is gaining momentum. Now, you can immerse viewers in 360° videos that make them feel as if they've been transported to another universe entirely, right from the comfort of their couch. 360° videos 360° videos Editing these immersive visuals requires specialized tools, techniques, and a broader imagination to truly harness the potential of full 360-degree spatial awareness. The Augmented Side of Things Having a Pokemon suddenly pop up in your living room was not only an exciting moment for all those 90s kids but also a step towards how AR has leveled up from its sci-fi roots. Video editing fluidly merging with reality is a major big deal! With AR, editors now have the chance to blend virtual elements into real-world footage seamlessly. Think of showcasing real-estate properties enhanced with virtual furniture or adding special effects that look almost palpable. showcasing real-estate properties showcasing real-estate properties Web Browser Video Editors Not all of us have a powerful computer for heavy-duty video editing software, and that's okay. For any amateur videographer or someone who's not yet ready to commit to professional editing software, web browser video editors offer a glimmer of hope. With a web browser video editor , there's no need for local file storage or endless rendering times; everything's processed in the cloud. Plus, forget about losing work due to system crashes (been there). All edits are saved instantly online. web browser video editor Social Media Ready Videos Now, this one isn't a prediction; it's happening already and will only continue to surge. Each social platform is unique in its video formats, dimensions, and lengths (TikTok's split-second videos are surely different from YouTube’s longer cuts). In line with that, the future of video editing has to keep up with these ever-changing demands. Video editors will increasingly cater to producing 'social media ready' clips– square for Instagram posts, vertical for IG and Snapchat stories, shorter versions for Twitter blasts... well, you get the drift. AI Voice-Overs and Subtitling I'm 99% sure that you've already seen a YouTube video with an AI voice that you mistook for a human narrator (Yep, it happens!). Although it happens to be a work in progress, AI voice-overs are starting to sound uncannily human. They're streamlined and time-efficient and provide you with numerous tonal options without having to hire someone to get behind the mic. AI is also sprinting ahead in perfectly syncing subtitles with speech (regardless of speed or delivery). Think about the amount of time we'll save not having to adjust the timing manually. Yes, accuracy still needs some fine-tuning, but once perfected? There's gonna be a drastic drop in many of our work hours. Text-Based Editing Can't be bothered to scrub through timelines? Text-based video editing is an emerging technology that just might simplify our lives in the near future. Simply jump directly to parts of a clip featuring specific words or phrases, just like finding instances of text in a Word document (Ctrl+F, anyone?) Yes, that's exactly what we're moving towards: editing via transcription. Want to take out a speaker from that mammoth two-hour seminar who used the phrase 'paradigm shift' one too many times? Find it and snip it just like that. Deep Fakes and Face Swaps This aspect tends to sit on the fine line between cool tech breakthroughs and ethically problematic uses . Advanced software can now replace faces completely from one person onto another's body in a video. This technology has been a hot topic, used in comedic fashion or controversial ways, and undeniably demonstrates the growing capabilities of video editing. ethically problematic uses ethically problematic uses While this area also needed to be regulated strictly for ethical use, we can't deny the wow factor of it in terms of technological progression. Interactive Video Editing Synchronize your exercise videos with your smartwatch or turn that cooking tutorial into an interactive experience by syncing instructions directly to your smart kitchen (yes, it exists). These are just a few of the exciting possibilities that interactive video editing offers. smart kitchen smart kitchen You can create videos that respond to user actions, perform actions like branching (choosing different paths in the video), shop directly from a product review video, or even change elements within the videos based on audience preferences. Fusing these 'Choose-Your-Own-Adventure' mechanics with movies and gaming will open up possibilities for whole new forms of media content. Can we just take a moment to picture how Netflix's interactive episode of Black Mirror: Bandersnatch was created? What About the Video Editors? We can't talk about the future of video editing without discussing those behind the lens, our video editors. No matter how much technology evolves, a human touch will always be essential. The intricate decisions that lead to storytelling and evoking feelings are often too complex (at least for now) for AI or software alone. However, this surge in tech doesn’t mean unemployment is looming on the horizon for all creative minds out there working as video editors...although there might be a slight shift in the required skill set. Surely being proficient with Adobe After Effects won't be enough, editors will need to develop new skills in AI programming, 360-degree video editing, AR and VR content creation, and managing cloud-based editing suites. Even with all these advancements, the need for human intuition, decision-making skills, and, ultimately, creativity will never lose their charm or importance. Remember: technology is an assistant, not a substitute for human creativity. Keep Calm and Click "Export" So, is it time to press the panic button with all these accelerating changes? Not at all. Trust me; you're not going to wake up tomorrow and suddenly have to edit a 360-degree video using AI tools in a virtual reality environment. (If that's happening, please tell me where you work—I want in.) These changes are ongoing; they're evolutionary, not revolutionary. Steps towards these technologies have been smoothly integrated into our workflows for years now – and just like how we made the leap from VCR to DVD or from landline to smartphone, we'll take all of this in stride as well.

Click, Cut, Create: The Future of Video Editing

Unlocking the Future of Bitcoin Liquidity: Portal to Bitcoin's Vision and Our Upcoming Testnet

বিটকয়েনের তারল্যের ভবিষ্যত আনলক করা: বিটকয়েনের দৃষ্টিভঙ্গি এবং আমাদের আসন্ন টেস্টনেটের পোর্টাল

Die Zukunft der Bitcoin-Liquidität erschließen: Portal zur Bitcoin-Vision und unserem kommenden Testnetz

Liberando el futuro de la liquidez de Bitcoin: portal a la visión de Bitcoin y nuestra próxima Testnet

Déverrouiller l'avenir de la liquidité Bitcoin : portail vers la vision de Bitcoin et notre prochain réseau de test

बिटकॉइन लिक्विडिटी के भविष्य को अनलॉक करना: बिटकॉइन के विज़न और हमारे आगामी टेस्टनेट का पोर्टल

ビットコイン流動性の未来を解き放つ: ビットコインのビジョンと今後のテストネットへのポータル

비트코인 유동성의 미래를 열다: 비트코인의 비전과 다가오는 테스트넷에 대한 포털

Desbloqueando o futuro da liquidez do Bitcoin: portal para a visão do Bitcoin e nossa próxima rede de testes

Открывая будущее ликвидности биткойнов: портал к видению биткойнов и нашей будущей тестовой сети

Bitcoin Likiditesinin Geleceğinin Kilidini Açmak: Bitcoin'in Vizyonuna ve Yaklaşan Test Ağımıza Açılan Portal

Mở khóa tương lai của tính thanh khoản Bitcoin: Cổng thông tin cho tầm nhìn của Bitcoin và mạng thử nghiệm sắp tới của chúng tôi

解锁比特币流动性的未来：比特币愿景和我们即将推出的测试网的门户

Building Advanced Video Search: Frame Search Versus Multi-Modal Embeddings

উন্নত ভিডিও অনুসন্ধান তৈরি করা: ফ্রেম অনুসন্ধান বনাম মাল্টি-মোডাল এম্বেডিং

Erstellen einer erweiterten Videosuche: Frame-Suche im Vergleich zu multimodalen Einbettungen

Creación de búsqueda de vídeo avanzada: búsqueda de fotogramas frente a incrustaciones multimodales

Création d'une recherche vidéo avancée : recherche de trames et intégrations multimodales

उन्नत वीडियो खोज का निर्माण: फ़्रेम खोज बनाम मल्टी-मोडल एम्बेडिंग

高度なビデオ検索の構築: フレーム検索とマルチモーダル埋め込み

고급 비디오 검색 구축: 프레임 검색과 다중 모드 임베딩 비교

Construindo pesquisa avançada de vídeo: pesquisa de quadros versus incorporações multimodais

Создание расширенного поиска видео: поиск кадров в сравнении с мультимодальными встраиваниями

Gelişmiş Video Arama Oluşturma: Çerçeve Aramaya Karşı Çok Modlu Yerleştirmeler

Xây dựng Tìm kiếm video nâng cao: Tìm kiếm khung so với nhúng đa phương thức

构建高级视频搜索：帧搜索与多模态嵌入

In this article, I want to showcase AI tools for creating summaries from YouTube videos. These AI tools can quickly summarize a video's content so you don't have to watch the entire thing. I'll demonstrate how to use these AIs to rapidly extract the main points from videos. https://youtu.be/LjtI64uEU8w?si=qx1cHvTv--KDVbQZ&embedable=true How to summarize a YouTube video with AI in 2024 AI Tools for Video Summarization Video summaries can save time, help you grasp key points quickly, and allow you to decide if watching the full video is worthwhile. I will use these four popular AI tools: ChatGPT
Claude
Google Gemini
Perplexity How to Use Each Tool How to summarize the video with ChatGPT ChatGPT now has internet access, making it easier to summarize videos directly. Open the website https://chatgpt.com/ and register there.


Copy the video link from YouTube.


Ask ChatGPT to "Summarize the video" along with the link.


If needed, ask for additional information to get a more comprehensive summary. It usually works immediately out of the box, but sometimes, if the video is new, it could show an error. In this case, please follow the steps to get the summary from the transcript. How to summarize the video with Claude Create an account on the https://claude.ai/ website to get access to AI. Claude doesn't have direct internet access, so you cannot just pass the link to the video; you need to pass a transcript. How to Get a Transcript from YouTube Video Open the video on YouTube.


Click on the "Show Transcript" button below the video description.


Copy the entire transcript.



Paste the transcript into Claude and ask it to summarize it. How to summarize the video with Google Gemini If you haven't registered yet, go to https://gemini.google.com/ and register there. Gemini performs very well with video summarization without needing extra steps. It also works well with new videos. Copy the video link
Ask Gemini to summarize the video and provide the link.
Gemini will generate a detailed summary without extra steps.

* How to summarize the video with Perplexity AI Visit https://www.perplexity.ai and register there. Perplexity offers the most detailed summaries but doesn't work with newly added videos. Copy the video link


Ask Perplexity to summarize the video and provide the link


Perplexity will generate a comprehensive summary

* Conclusion Based on the comparison in the video, I’ll rank these tools in the following order: Perplexity
Google Gemini
ChatGPT
Claude Each tool has its strengths, and the best choice may depend on your specific needs and the video content you're summarizing. Try these tools and see which works best for you! Don’t forget to share your favorite in the comments! Cheers! In this article, I want to showcase AI tools for creating summaries from YouTube videos. These AI tools can quickly summarize a video's content so you don't have to watch the entire thing. I'll demonstrate how to use these AIs to rapidly extract the main points from videos. https://youtu.be/LjtI64uEU8w?si=qx1cHvTv--KDVbQZ&embedable=true https://youtu.be/LjtI64uEU8w?si=qx1cHvTv--KDVbQZ&embedable=true How to summarize a YouTube video with AI in 2024 How to summarize a YouTube video with AI in 2024 AI Tools for Video Summarization Video summaries can save time, help you grasp key points quickly, and allow you to decide if watching the full video is worthwhile. I will use these four popular AI tools: I will use these four popular AI tools: ChatGPT Claude Google Gemini Perplexity ChatGPT Claude Google Gemini Perplexity How to Use Each Tool How to summarize the video with ChatGPT ChatGPT now has internet access, making it easier to summarize videos directly. Open the website https://chatgpt.com/ and register there. Copy the video link from YouTube. Ask ChatGPT to "Summarize the video" along with the link. If needed, ask for additional information to get a more comprehensive summary. Open the website https://chatgpt.com/ and register there. Open the website https://chatgpt.com/ and register there. https://chatgpt.com/ https://chatgpt.com/ Copy the video link from YouTube. Copy the video link from YouTube. Ask ChatGPT to "Summarize the video" along with the link. Ask ChatGPT to "Summarize the video" along with the link. If needed, ask for additional information to get a more comprehensive summary. If needed, ask for additional information to get a more comprehensive summary. It usually works immediately out of the box, but sometimes, if the video is new, it could show an error. In this case, please follow the steps to get the summary from the transcript. How to summarize the video with Claude Create an account on the https://claude.ai/ website to get access to AI. https://claude.ai/ https://claude.ai/ Claude doesn't have direct internet access, so you cannot just pass the link to the video; you need to pass a transcript. How to Get a Transcript from YouTube Video How to Get a Transcript from YouTube Video Open the video on YouTube. Click on the "Show Transcript" button below the video description. Copy the entire transcript. Paste the transcript into Claude and ask it to summarize it. Open the video on YouTube. Open the video on YouTube. Click on the "Show Transcript" button below the video description. Click on the "Show Transcript" button below the video description. Copy the entire transcript. Copy the entire transcript. Paste the transcript into Claude and ask it to summarize it. Paste the transcript into Claude and ask it to summarize it. How to summarize the video with Google Gemini If you haven't registered yet, go to https://gemini.google.com/ and register there. Gemini performs very well with video summarization without needing extra steps. It also works well with new videos. Copy the video link Ask Gemini to summarize the video and provide the link. Gemini will generate a detailed summary without extra steps.

* Copy the video link Ask Gemini to summarize the video and provide the link. Gemini will generate a detailed summary without extra steps. * * * How to summarize the video with Perplexity AI Visit https://www.perplexity.ai and register there. Perplexity offers the most detailed summaries but doesn't work with newly added videos. https://www.perplexity.ai Copy the video link Ask Perplexity to summarize the video and provide the link Perplexity will generate a comprehensive summary

* Copy the video link Copy the video link Ask Perplexity to summarize the video and provide the link Ask Perplexity to summarize the video and provide the link Perplexity will generate a comprehensive summary * Perplexity will generate a comprehensive summary * * Conclusion Based on the comparison in the video, I’ll rank these tools in the following order: Perplexity Google Gemini ChatGPT Claude Perplexity Google Gemini ChatGPT Claude Each tool has its strengths, and the best choice may depend on your specific needs and the video content you're summarizing. Try these tools and see which works best for you! Don’t forget to share your favorite in the comments! Cheers!

How to Quickly Summarize YouTube Videos Using Gemini, ChatGPT, Claude, and Perplexity in 2024

Authors:
(1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu);
(2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work;
(3) Yixu Chen, Amazon Prime Video;
(4) Yongjun Wu, Amazon Prime Video;
(5) Hai Wei,  Amazon Prime Video;
(6)Sriram Sethuraman, Amazon Prime Video;
(7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction
Related Work
Details of Subjective Study
Subjective Analysis
Objective Assessment
Conclusion, Acknowledgment and References IV. SUBJECTIVE ANALYSIS A. Internal Correlation The internal correlations for the three groups were calculated as follows. Let the score given by subject i for video j on TV t be given by uijt. The Z score is computed as The subjects were randomly divided into two equal groups and the average Z score computed for each video across all subjects in that group. The correlation between the scores provided by these groups for all the videos was computed over 100 trails with different random groupings. The median inter-subject correlation for viewers of TV1 was found to be 0.95, for TV2 was 0.94, and for TV3 was 0.93. These data indicate a high degree of internal consistency and reliability. B. Calculation of MOS The Mean Opinion Scores (MOS) were obtained using the Maximum Likelihood Estimation method proposed in ITU 910 [30]. The MOS is modelled as a random variable where Ψjn is the true quality of video j viewed on TVn, ∆i is the bias of subject i, νi represents the inconsistency of subject i, and X ∼ N(0, 1) are i.i.d. Gaussian random variables. Given the scores uijn, the true score for each video on each television is estimated by treating Ψjn, ∆i , νi as free parameters that are solved so that the model in (2) is the best fit to the observed MOS. Specifically, Ψjt is by maximizing the log-likelihood of the observations using the Newton-Raphson solver. The Differential Mean Opinion Scores (DMOS) were calculated between the distorted and reference videos by finding the differences in MOS as follows: where j0 is the index of the reference video corresponding to video j viewed on television TVn. The MOS and DMOS are thus computed for each video and separately for each television. C. Analysis of Scores The MOS of 20 videos of 4 contents (1 content from each source group and 5 videos per content) are plotted against bitrate in Figs. 2, 3, 4, and 5. The Forge video is from group 1, EPL is from group 2, ColorDJ is from group 3, and Porsche is from group 4. Screenshots from the SDR versions of the source contents are shown in the first column. The points on each line correspond to 540p, 720p, 1080p, 1440p, 2160p, and 2160p in increasing order, following Table I. The Forge video shows a sword being forged in a smithy. The light from the hot iron contrasts strongly with the darkness and shadows around it and in the background. The HDR version of the Forge video is rated higher than the SDR version on TV1, since details are clearer in the HDR version. However, on TV2 and TV3, due to their reduced ability to display contrasts, higher compression of the HDR video offsets the relative quality gain from the increased contrast, and the SDR version was thus rated as better. The HDR version appears darker and dimmer than the SDR version, due to the way in which HDR is displayed differently from SDR (as discussed in the Introduction) as well as the way in which the video was graded. The lower average brightness may also contribute to the lower perceived quality of the HDR version of the video content. The lower average brightness of the HDR version, the reduced contrast of TV2 and TV3, and the high spatial complexity of the scene, may also explain why, at higher bitrates, the SDR version was still rated higher. Similar observations can be made about the EPL video taken from group 2, where the HDR version was rated better on TV1, while the SDR version was rated better at lower bitrates on TV2 and TV3. At higher bitrates, the HDR version was rated higher, as compressive artifacts had less visual impact, despite the high temporal complexity of the soccer scene. The SDR version of the ColorDJ video suffers from oversaturation and overexposure due to the wide range of colors and brightnesses present. The HDR version, on the other hand, does not exhibit over-saturation or overexposure because of its greater bit-depth and wider color gamut. This may explain why the HDR version was generally rated better on all the televisions. The Porsche video has a bright red paint on the car which is accurately represented in HDR, but looks saturated in SDR. The video is also not spatially or temporally complex, which may be because of the superior ability of HDR to represent contrasts and bright colors, outweighing the effects of compression. However, there is a sharp drop in the MOS vs bitrate curve for the 720p HDR version of the video, encoded at 1515 kbps. The encoder made the decision to encode the 1080p version at 1300 kbps, yet the 1080p version is still rated higher than the 720p version. This is likely because rescaling artifacts are more prominent than the compressive artifacts on this content due to its low spatial complexity. The average difference between the MOS of the HDR and SDR versions of all the videos are plotted against the maxrate of each television in Fig. 6. As may be seen, at lower bitrates SDR was rated higher than HDR, while at higher bitrates the difference becomes positive. The drops in the curve of 1080p content at maxrate 3000 kbps and of 2160p at maxrate 6000 kbps are indicative of how for an optimal bitrate ladder, these resolutions should be encoded at higher bitrates. It is important to note that these compression levels were included in the study to represent a wide range of quality, and not to represent an optimal bitrate ladder. For TV1, HDR was better than SDR at even low bitrates (3000 kbps) and the difference in quality increased to 5 MOS units at the highest bitrate. For TV2 and TV3, SDR quality was better than HDR quality at low bitrates but the differences decreased and became slightly positive at higher bitrates. This also indicates how the capabilities of the display devices strongly influence the quality of HDR relative to that of SDR content. D. Combining Databases The scores given to the 27 anchor videos from the LIVE HDR, LIVE AQ HDR, and the APV HDR Sports VQA datasets were used to map scores from those databases to the current database. A logistic function was fitted to map the scores assigned to the anchor videos in those databases to the scores of the same videos in the LIVE HDRvsSDR database for each television. The logistic function is where x are the scores of the videos in prior databases, f(x) is the mapping to scores in the LIVE HDRvsSDR database for a particular television, and a, b, c and s were separately solved for on each database using the anchor videos for that database. Since each television and each (prior) database was used to generate a different fitting function, a total of nine functions were derived from the data. The functions that map scores from the LIVE HDR database and the LIVE AQ HDR database to the scores obtained from the three televisions used in the LIVE HDRvsSDR database are plotted in Figs. 12 and 13 in supplementary material, respectively. The mappings and results for the APV HDR Sports database cannot be shown, for proprietary reasons. Deriving these functions from the scores of the anchor videos enables the merging of the three LIVE databases into a single, large-scale database of 1066 videos that were collected in a controlled laboratory environment. This paper is available on arxiv under CC 4.0 license. Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Authors: Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction Related Work Details of Subjective Study Subjective Analysis Objective Assessment Conclusion, Acknowledgment and References Abstract and Introduction Abstract and Introduction Related Work Related Work Details of Subjective Study Details of Subjective Study Subjective Analysis Subjective Analysis Objective Assessment Objective Assessment Conclusion, Acknowledgment and References Conclusion, Acknowledgment and References IV. SUBJECTIVE ANALYSIS A. Internal Correlation A. Internal Correlation The internal correlations for the three groups were calculated as follows. Let the score given by subject i for video j on TV t be given by uijt. The Z score is computed as The subjects were randomly divided into two equal groups and the average Z score computed for each video across all subjects in that group. The correlation between the scores provided by these groups for all the videos was computed over 100 trails with different random groupings. The median inter-subject correlation for viewers of TV1 was found to be 0.95, for TV2 was 0.94, and for TV3 was 0.93. These data indicate a high degree of internal consistency and reliability. B. Calculation of MOS B. Calculation of MOS The Mean Opinion Scores (MOS) were obtained using the Maximum Likelihood Estimation method proposed in ITU 910 [30]. The MOS is modelled as a random variable where Ψjn is the true quality of video j viewed on TVn, ∆i is the bias of subject i, νi represents the inconsistency of subject i, and X ∼ N(0, 1) are i.i.d. Gaussian random variables. Given the scores uijn, the true score for each video on each television is estimated by treating Ψjn, ∆i , νi as free parameters that are solved so that the model in (2) is the best fit to the observed MOS. Specifically, Ψjt is by maximizing the log-likelihood of the observations using the Newton-Raphson solver. The Differential Mean Opinion Scores (DMOS) were calculated between the distorted and reference videos by finding the differences in MOS as follows: where j0 is the index of the reference video corresponding to video j viewed on television TVn. The MOS and DMOS are thus computed for each video and separately for each television. C. Analysis of Scores C. Analysis of Scores The MOS of 20 videos of 4 contents (1 content from each source group and 5 videos per content) are plotted against bitrate in Figs. 2, 3, 4, and 5. The Forge video is from group 1, EPL is from group 2, ColorDJ is from group 3, and Porsche is from group 4. Screenshots from the SDR versions of the source contents are shown in the first column. The points on each line correspond to 540p, 720p, 1080p, 1440p, 2160p, and 2160p in increasing order, following Table I. The Forge video shows a sword being forged in a smithy. The light from the hot iron contrasts strongly with the darkness and shadows around it and in the background. The HDR version of the Forge video is rated higher than the SDR version on TV1, since details are clearer in the HDR version. However, on TV2 and TV3, due to their reduced ability to display contrasts, higher compression of the HDR video offsets the relative quality gain from the increased contrast, and the SDR version was thus rated as better. The HDR version appears darker and dimmer than the SDR version, due to the way in which HDR is displayed differently from SDR (as discussed in the Introduction) as well as the way in which the video was graded. The lower average brightness may also contribute to the lower perceived quality of the HDR version of the video content. The lower average brightness of the HDR version, the reduced contrast of TV2 and TV3, and the high spatial complexity of the scene, may also explain why, at higher bitrates, the SDR version was still rated higher. Similar observations can be made about the EPL video taken from group 2, where the HDR version was rated better on TV1, while the SDR version was rated better at lower bitrates on TV2 and TV3. At higher bitrates, the HDR version was rated higher, as compressive artifacts had less visual impact, despite the high temporal complexity of the soccer scene. The SDR version of the ColorDJ video suffers from oversaturation and overexposure due to the wide range of colors and brightnesses present. The HDR version, on the other hand, does not exhibit over-saturation or overexposure because of its greater bit-depth and wider color gamut. This may explain why the HDR version was generally rated better on all the televisions. The Porsche video has a bright red paint on the car which is accurately represented in HDR, but looks saturated in SDR. The video is also not spatially or temporally complex, which may be because of the superior ability of HDR to represent contrasts and bright colors, outweighing the effects of compression. However, there is a sharp drop in the MOS vs bitrate curve for the 720p HDR version of the video, encoded at 1515 kbps. The encoder made the decision to encode the 1080p version at 1300 kbps, yet the 1080p version is still rated higher than the 720p version. This is likely because rescaling artifacts are more prominent than the compressive artifacts on this content due to its low spatial complexity. The average difference between the MOS of the HDR and SDR versions of all the videos are plotted against the maxrate of each television in Fig. 6. As may be seen, at lower bitrates SDR was rated higher than HDR, while at higher bitrates the difference becomes positive. The drops in the curve of 1080p content at maxrate 3000 kbps and of 2160p at maxrate 6000 kbps are indicative of how for an optimal bitrate ladder, these resolutions should be encoded at higher bitrates. It is important to note that these compression levels were included in the study to represent a wide range of quality, and not to represent an optimal bitrate ladder. For TV1, HDR was better than SDR at even low bitrates (3000 kbps) and the difference in quality increased to 5 MOS units at the highest bitrate. For TV2 and TV3, SDR quality was better than HDR quality at low bitrates but the differences decreased and became slightly positive at higher bitrates. This also indicates how the capabilities of the display devices strongly influence the quality of HDR relative to that of SDR content. D. Combining Databases D. Combining Databases The scores given to the 27 anchor videos from the LIVE HDR, LIVE AQ HDR, and the APV HDR Sports VQA datasets were used to map scores from those databases to the current database. A logistic function was fitted to map the scores assigned to the anchor videos in those databases to the scores of the same videos in the LIVE HDRvsSDR database for each television. The logistic function is where x are the scores of the videos in prior databases, f(x) is the mapping to scores in the LIVE HDRvsSDR database for a particular television, and a, b, c and s were separately solved for on each database using the anchor videos for that database. Since each television and each (prior) database was used to generate a different fitting function, a total of nine functions were derived from the data. The functions that map scores from the LIVE HDR database and the LIVE AQ HDR database to the scores obtained from the three televisions used in the LIVE HDRvsSDR database are plotted in Figs. 12 and 13 in supplementary material, respectively. The mappings and results for the APV HDR Sports database cannot be shown, for proprietary reasons. Deriving these functions from the scores of the anchor videos enables the merging of the three LIVE databases into a single, large-scale database of 1066 videos that were collected in a controlled laboratory environment. This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

HDR or SDR? A Study of Scaled and Compressed Videos: Subjective Analysis

HDR or SDR? A Study of Scaled and Compressed Videos: Details of Subjective Study

Authors:
(1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu);
(2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work;
(3) Yixu Chen, Amazon Prime Video;
(4) Yongjun Wu, Amazon Prime Video;
(5) Hai Wei,  Amazon Prime Video;
(6)Sriram Sethuraman, Amazon Prime Video;
(7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction
Related Work
Details of Subjective Study
Subjective Analysis
Objective Assessment
Conclusion, Acknowledgment and References II. RELATED WORK To the best of our knowledge, there do not exist any studies that compare the subjective qualities of videos as the dynamic range, resolution, and compression levels are all varied. Existing databases such as LIVE Livestream [2], LIVE ETRI [3], LIVE YTHFR [4], AVT UHD [5], and APV LBMFR [6] study the subjective quality of professionallygenerated SDR videos under conditions of downsampling, compression, and source distortions. Other datasets including Konvid-1k [7], YouTube UGC [8], and LSVQ [9] study the quality of SDR user-generated content. UGC databases are typically much larger than those that study the quality of professional-grade content because they can be conducted online via crowdsourcing owing to looser requirements on the display devices, resolution, and bitrate. LIVE HDR [10], LIVE AQ HDR [11], and APV HDR Sports [12] are recent databases that study the quality of professionally-created HDR videos that have been downsampled and compressed at various resolutions and bitrates. Each of the above-mentioned databases study the quality of either SDR or HDR videos, but not both., that have been subject to distortions. Here we present the first subjective study that compares the quality of HDR and SDR videos of the same content, that have been processed by downscaling and compression. We conducted the study on a variety of display devices using different technologies and having differing capabilities. While subjective human scores from studies like the one we conducted are considered the gold standards of video quality, conducting such studies is expensive is not scalable. However, objective video quality metrics are designed and trained to automatically predict video quality and can be quite economic and scalable. These fall into two categories: FullReference (FR) and No-Reference (NR) models. FR VQA models require as take as input both pristine and distorted videos to measure the quality of the distorted videos. NR metrics only have access to distorted videos when predicting quality, hence designing them is a more challenging problem. NR VQA models are relevant for video source inspection as well as when measuring quality with no available source video. PSNR measures the peak signal to noise ratio between a reference frame and a distorted version of the same frame. SSIM [13] incorporates luminance, contrast, and structure features to predict the quality of distorted images. VMAF [14] models the statistics of the wavelet coefficients of video frames, as well as the detail losses from distortions. SpEED [15] measures the difference in entropy of bandpass coefficients of reference and distorted videos. STRRED [16] models the statistics of space-time video wavelet coefficients altered by distortions. STGREED [17] measures differences in temporal and spatial entropy arising from distortions to model the quality of videos having varying frame rates and bitrates. BRISQUE [18], VBLIINDS [19], VIDEVAL [20], RAPIQUE [21], ChipQA [22], HDR ChipQA [23] and NIQE [24] are NR video quality metrics that rely on neurostatistical models of visual perception. Pristine videos are known to follow certain regular statistics when processed using visual neural models. Distortions predictably alter the statistics of perceptually processed videos, allowing for the design of accurate VQA models. RAPIQUE combines features developed under these models with (semantic) video features provided by a pre-trained deep network. TLVQM [25] explicitly models common distortions such as compression, blur, and flicker, using a variety of spatial and temporal filters and heuristics. This paper is available on arxiv under CC 4.0 license. Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Authors: Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction Related Work Details of Subjective Study Subjective Analysis Objective Assessment Conclusion, Acknowledgment and References Abstract and Introduction Abstract and Introduction Related Work Related Work Details of Subjective Study Details of Subjective Study Subjective Analysis Subjective Analysis Objective Assessment Objective Assessment Conclusion, Acknowledgment and References Conclusion, Acknowledgment and References II. RELATED WORK To the best of our knowledge, there do not exist any studies that compare the subjective qualities of videos as the dynamic range, resolution, and compression levels are all varied. Existing databases such as LIVE Livestream [2], LIVE ETRI [3], LIVE YTHFR [4], AVT UHD [5], and APV LBMFR [6] study the subjective quality of professionallygenerated SDR videos under conditions of downsampling, compression, and source distortions. Other datasets including Konvid-1k [7], YouTube UGC [8], and LSVQ [9] study the quality of SDR user-generated content. UGC databases are typically much larger than those that study the quality of professional-grade content because they can be conducted online via crowdsourcing owing to looser requirements on the display devices, resolution, and bitrate. LIVE HDR [10], LIVE AQ HDR [11], and APV HDR Sports [12] are recent databases that study the quality of professionally-created HDR videos that have been downsampled and compressed at various resolutions and bitrates. Each of the above-mentioned databases study the quality of either SDR or HDR videos, but not both., that have been subject to distortions. Here we present the first subjective study that compares the quality of HDR and SDR videos of the same content, that have been processed by downscaling and compression. We conducted the study on a variety of display devices using different technologies and having differing capabilities. While subjective human scores from studies like the one we conducted are considered the gold standards of video quality, conducting such studies is expensive is not scalable. However, objective video quality metrics are designed and trained to automatically predict video quality and can be quite economic and scalable. These fall into two categories: FullReference (FR) and No-Reference (NR) models. FR VQA models require as take as input both pristine and distorted videos to measure the quality of the distorted videos. NR metrics only have access to distorted videos when predicting quality, hence designing them is a more challenging problem. NR VQA models are relevant for video source inspection as well as when measuring quality with no available source video. PSNR measures the peak signal to noise ratio between a reference frame and a distorted version of the same frame. SSIM [13] incorporates luminance, contrast, and structure features to predict the quality of distorted images. VMAF [14] models the statistics of the wavelet coefficients of video frames, as well as the detail losses from distortions. SpEED [15] measures the difference in entropy of bandpass coefficients of reference and distorted videos. STRRED [16] models the statistics of space-time video wavelet coefficients altered by distortions. STGREED [17] measures differences in temporal and spatial entropy arising from distortions to model the quality of videos having varying frame rates and bitrates. BRISQUE [18], VBLIINDS [19], VIDEVAL [20], RAPIQUE [21], ChipQA [22], HDR ChipQA [23] and NIQE [24] are NR video quality metrics that rely on neurostatistical models of visual perception. Pristine videos are known to follow certain regular statistics when processed using visual neural models. Distortions predictably alter the statistics of perceptually processed videos, allowing for the design of accurate VQA models. RAPIQUE combines features developed under these models with (semantic) video features provided by a pre-trained deep network. TLVQM [25] explicitly models common distortions such as compression, blur, and flicker, using a variety of spatial and temporal filters and heuristics. This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

HDR or SDR? A Study of Scaled and Compressed Videos: Related Work

Authors:
(1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu);
(2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work;
(3) Yixu Chen, Amazon Prime Video;
(4) Yongjun Wu, Amazon Prime Video;
(5) Hai Wei,  Amazon Prime Video;
(6)Sriram Sethuraman, Amazon Prime Video;
(7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction
Related Work
Details of Subjective Study
Subjective Analysis
Objective Assessment
Conclusion, Acknowledgment and References VI. CONCLUSION We presented the first ever study on comparing HDR and SDR videos of the same content encoded at different bitrates and resolutions on different display devices. Our study shows that despite HDR’s theoretical capabilities over SDR, its perceptual quality depends heavily on the display device used in practice. We also evaluated several NR and FR VQA algorithms on the new database, and presented a novel NR VQA algorithm called HDRPatchMAX that exceeds the current state-of-the-art on this database. We hope that this spurs research on the modelling of display devices in VQA algorithms as well as for optimal bitrate ladders for streaming. ACKNOWLEDGMENT This research was sponsored by a grant from Amazon.com, Inc., and by grant number 2019844 for the National Science Foundation AI Institute for Foundations of Machine Learning (IFML). The authors also thank the Texas Advanced Computing Center (TACC) at The University of Texas at Austin for providing HPC resources that have contributed to the research results reported in this paper. URL: http://www.tacc.utexas.edu. REFERENCES [1] ITU, “BT.709 : Parameter values for the HDTV standards for production and international programme exchange,” Intl. Telecomm. Union, Tech. Rep., 2011. [2] Z. Shang, J. P. Ebenezer, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “Study of the subjective and objective quality of high motion live streaming videos,” IEEE Trans. Image Process., vol. 31, pp. 1027– 1041, 2022. [3] D. Y. Lee, S. Paul, C. G. Bampis, H. Ko, J. Kim, S. Y. Jeong, B. Homan, and A. C. Bovik, “A subjective and objective study of space-time subsampled video quality,” arXiv preprint arXiv:2102.00088, 2021. [4] P. C. Madhusudana, X. Yu, N. Birkbeck, Y. Wang, B. Adsumilli, and A. C. Bovik, “Subjective and objective quality assessment of high frame rate videos,” IEEE Access, vol. 9, pp. 108 069–108 082, 2021. [5] R. R. R. Rao, S. Goring, W. Robitza, B. Feiten, and A. Raake, “Avt- ¨ vqdb-uhd-1: A large scale video quality database for uhd-1,” in 2019 IEEE Intl. Symposium on Multimedia (ISM). IEEE, 2019, pp. 17–177. [6] J. P. Ebenezer, Y. Chen, Y. Wu, H. Wei, and S. Sethuraman, “Subjective and objective quality assessment of high-motion sports videos at lowbitrates,” in 2022 IEEE Intl. Conf. on Image Process. (ICIP), 2022, pp. 521–525. [7] V. Hosu, F. Hahn, M. Jenadeleh, H. Lin, H. Men, T. Sziranyi, S. Li, and ´ D. Saupe, “The Konstanz natural video database (Konvid-1k),” in Int. Conf. Quality of Multimedia Experience, 2017, pp. 1–6. [8] Y. Wang, S. Inguva, and B. Adsumilli, “Youtube UGC dataset for video compression research,” in IEEE Int. Workshop Multimed. Signal Process. IEEE, 2019, pp. 1–5. [9] Z. Ying, M. Mandal, D. Ghadiyaram, and A. Bovik, “Patch-VQ: ‘patching up’ the video quality problem,” in IEEE Conf. Comp. Vision Pattern Recognit., 2021, pp. 14 014–14 024. [10] Z. Shang, J. P. Ebenezer, A. C. Bovik, Y. Wu, H. Wei, and S. Sethuraman, “Subjective assessment of high dynamic range videos under different ambient conditions,” in IEEE Intl. Conf. Image Process., 2022, pp. 786–790. [11] Z. Shang, J. P. Ebenezer, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “A study of subjective and objective quality assessment of HDR videos,” submitted to IEEE Trans. Image Process., 2022. [12] Z. Shang, Y. Chen, Y. Wu, H. Wei, and S. Sethuraman, “Subjective and objective video quality assessment of high dynamic range sports content,” in Proceedings of the IEEE/CVF Winter Conf. on Applications of Computer Vision (WACV) Workshops, January 2023, pp. 556–564. [13] Z. Wang, A. Bovik, H. Sheikh, and E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Trans. Image Process., vol. 13, no. 4, pp. 600–612, 2004. [14] Netflix, VMAF: The Journey Continues, 2018 (accessed December 28, 2021). [Online]. Available: https://netflixtechblog. com/vmaf-the-journey-continues-44b51ee9ed12 [15] C. G. Bampis, P. Gupta, R. Soundararajan, and A. C. Bovik, “Speedqa: Spatial efficient entropic differencing for image and video quality,” IEEE Signal Process. Letters, vol. 24, no. 9, pp. 1333–1337, 2017. [16] R. Soundararajan and A. C. Bovik, “Video quality assessment by reduced reference spatio-temporal entropic differencing,” IEEE Trans. Circuits Syst. Video Technol., vol. 23, no. 4, pp. 684–694, 2012. [17] P. C. Madhusudana, N. Birkbeck, Y. Wang, B. Adsumilli, and A. C. Bovik, “St-greed: Space-time generalized entropic differences for frame rate dependent video quality prediction,” IEEE Trans. Image Process., vol. 30, pp. 7446–7457, 2021. [18] A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-reference image quality assessment in the spatial domain,” IEEE Trans. Image Process., vol. 21, no. 12, pp. 4695–4708, 2012. [19] M. A. Saad, A. C. Bovik, and C. Charrier, “Blind prediction of natural video quality,” IEEE Trans. Image Process., vol. 23, no. 3, pp. 1352– 1365, 2014. [20] Z. Tu, Y. Wang, N. Birkbeck, B. Adsumilli, and A. C. Bovik, “UGCVQA: Benchmarking blind video quality assessment for user generated content,” IEEE Trans. Image Process., vol. 30, pp. 4449–4464, 2021. [21] Z. Tu, X. Yu, Y. Wang, N. Birkbeck, B. Adsumilli, and A. C. Bovik, “RAPIQUE: Rapid and accurate video quality prediction of user generated content,” arXiv preprint arXiv:2101.10955, 2021. [22] J. P. Ebenezer, Z. Shang, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “ChipQA: No-reference video quality prediction via space-time chips,” IEEE Trans. Image Process., vol. 30, pp. 8059–8074, 2021. [23] ——, “HDR-ChipQA: No-reference quality assessment for high dynamic range videos,” submitted to IEEE Trans. Image Process., 2023. [24] A. Mittal, R. Soundararajan, and A. C. Bovik, “Making a “completely blind” image quality analyzer,” IEEE Signal Process. Lett., vol. 20, no. 3, pp. 209–212, 2012. [25] J. Korhonen, “Two-level approach for no-reference consumer video quality assessment,” IEEE Trans. Image Process., vol. 28, no. 12, pp. 5923–5938, 2019. [26] ITU, “ BT.2020 : Parameter values for ultra-high definition television systems for production and international programme exchange,” Intl. Telecomm. Union, Tech. Rep., 2015. [27] S. Miller, M. Nezamabadi, and S. Daly, “Perceptual signal coding for more efficient usage of bit codes,” SMPTE Motion Imaging J., vol. 122, no. 4, pp. 52–59, 2013. [28] Rtings, Peak Brightness Measurement, 2023 (accessed February 28, 2023). [Online]. Available: https://www.rtings.com/tv/tests/ picture-quality/sdr-peak-brightness [29] ITU, “BT.500 : Methodologies for the subjective assessment of the quality of television images,” Intl. Telecomm. Union, Tech. Rep., 2019. [30] ——, “ITU. 910 : Subjective video quality assessment methods for multimedia applications,” Intl. Telecomm. Union, Tech. Rep., 2008. [31] Z. Wang, E. Simoncelli, and A. Bovik, “Multiscale structural similarity for image quality assessment,” in Asilomar Conf. Signals, Syst., Comput., vol. 2, 2003, pp. 1398–1402 Vol.2. [32] H. R. Sheikh, M. F. Sabir, and A. C. Bovik, “A statistical evaluation of recent full reference image quality assessment algorithms,” IEEE Trans. Image Process., vol. 15, no. 11, pp. 3440–3451, 2006. [33] J. P. Ebenezer, Z. Shang, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “Making video quality assessment models robust to bit depth,” submitted to IEEE Signal Process. Letters, 2023. [34] J. P. Ebenezer, Z. Shang, Y. Wu, H. Wei, and A. C. Bovik, “Noreference video quality assessment using space-time chips,” in IEEE Intl. Workshop Multimedia Signal Process., 2020, pp. 1–6. [35] D. W. Dong and J. J. Atick, “Temporal decorrelation: a theory of lagged and nonlagged responses in the lateral geniculate nucleus,” Netw.: Comput. Neural Syst., vol. 6, no. 2, pp. 159–178, 1995. This paper is available on arxiv under CC 4.0 license. Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Authors: Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction Related Work Details of Subjective Study Subjective Analysis Objective Assessment Conclusion, Acknowledgment and References Abstract and Introduction Abstract and Introduction Related Work Related Work Details of Subjective Study Details of Subjective Study Subjective Analysis Subjective Analysis Objective Assessment Objective Assessment Conclusion, Acknowledgment and References Conclusion, Acknowledgment and References VI. CONCLUSION We presented the first ever study on comparing HDR and SDR videos of the same content encoded at different bitrates and resolutions on different display devices. Our study shows that despite HDR’s theoretical capabilities over SDR, its perceptual quality depends heavily on the display device used in practice. We also evaluated several NR and FR VQA algorithms on the new database, and presented a novel NR VQA algorithm called HDRPatchMAX that exceeds the current state-of-the-art on this database. We hope that this spurs research on the modelling of display devices in VQA algorithms as well as for optimal bitrate ladders for streaming. ACKNOWLEDGMENT This research was sponsored by a grant from Amazon.com, Inc., and by grant number 2019844 for the National Science Foundation AI Institute for Foundations of Machine Learning (IFML). The authors also thank the Texas Advanced Computing Center (TACC) at The University of Texas at Austin for providing HPC resources that have contributed to the research results reported in this paper. URL: http://www.tacc.utexas.edu. REFERENCES [1] ITU, “BT.709 : Parameter values for the HDTV standards for production and international programme exchange,” Intl. Telecomm. Union, Tech. Rep., 2011. [2] Z. Shang, J. P. Ebenezer, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “Study of the subjective and objective quality of high motion live streaming videos,” IEEE Trans. Image Process., vol. 31, pp. 1027– 1041, 2022. [3] D. Y. Lee, S. Paul, C. G. Bampis, H. Ko, J. Kim, S. Y. Jeong, B. Homan, and A. C. Bovik, “A subjective and objective study of space-time subsampled video quality,” arXiv preprint arXiv:2102.00088, 2021. [4] P. C. Madhusudana, X. Yu, N. Birkbeck, Y. Wang, B. Adsumilli, and A. C. Bovik, “Subjective and objective quality assessment of high frame rate videos,” IEEE Access, vol. 9, pp. 108 069–108 082, 2021. [5] R. R. R. Rao, S. Goring, W. Robitza, B. Feiten, and A. Raake, “Avt- ¨ vqdb-uhd-1: A large scale video quality database for uhd-1,” in 2019 IEEE Intl. Symposium on Multimedia (ISM). IEEE, 2019, pp. 17–177. [6] J. P. Ebenezer, Y. Chen, Y. Wu, H. Wei, and S. Sethuraman, “Subjective and objective quality assessment of high-motion sports videos at lowbitrates,” in 2022 IEEE Intl. Conf. on Image Process. (ICIP), 2022, pp. 521–525. [7] V. Hosu, F. Hahn, M. Jenadeleh, H. Lin, H. Men, T. Sziranyi, S. Li, and ´ D. Saupe, “The Konstanz natural video database (Konvid-1k),” in Int. Conf. Quality of Multimedia Experience, 2017, pp. 1–6. [8] Y. Wang, S. Inguva, and B. Adsumilli, “Youtube UGC dataset for video compression research,” in IEEE Int. Workshop Multimed. Signal Process. IEEE, 2019, pp. 1–5. [9] Z. Ying, M. Mandal, D. Ghadiyaram, and A. Bovik, “Patch-VQ: ‘patching up’ the video quality problem,” in IEEE Conf. Comp. Vision Pattern Recognit., 2021, pp. 14 014–14 024. [10] Z. Shang, J. P. Ebenezer, A. C. Bovik, Y. Wu, H. Wei, and S. Sethuraman, “Subjective assessment of high dynamic range videos under different ambient conditions,” in IEEE Intl. Conf. Image Process., 2022, pp. 786–790. [11] Z. Shang, J. P. Ebenezer, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “A study of subjective and objective quality assessment of HDR videos,” submitted to IEEE Trans. Image Process., 2022. [12] Z. Shang, Y. Chen, Y. Wu, H. Wei, and S. Sethuraman, “Subjective and objective video quality assessment of high dynamic range sports content,” in Proceedings of the IEEE/CVF Winter Conf. on Applications of Computer Vision (WACV) Workshops, January 2023, pp. 556–564. [13] Z. Wang, A. Bovik, H. Sheikh, and E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Trans. Image Process., vol. 13, no. 4, pp. 600–612, 2004. [14] Netflix, VMAF: The Journey Continues, 2018 (accessed December 28, 2021). [Online]. Available: https://netflixtechblog. com/vmaf-the-journey-continues-44b51ee9ed12 [15] C. G. Bampis, P. Gupta, R. Soundararajan, and A. C. Bovik, “Speedqa: Spatial efficient entropic differencing for image and video quality,” IEEE Signal Process. Letters, vol. 24, no. 9, pp. 1333–1337, 2017. [16] R. Soundararajan and A. C. Bovik, “Video quality assessment by reduced reference spatio-temporal entropic differencing,” IEEE Trans. Circuits Syst. Video Technol., vol. 23, no. 4, pp. 684–694, 2012. [17] P. C. Madhusudana, N. Birkbeck, Y. Wang, B. Adsumilli, and A. C. Bovik, “St-greed: Space-time generalized entropic differences for frame rate dependent video quality prediction,” IEEE Trans. Image Process., vol. 30, pp. 7446–7457, 2021. [18] A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-reference image quality assessment in the spatial domain,” IEEE Trans. Image Process., vol. 21, no. 12, pp. 4695–4708, 2012. [19] M. A. Saad, A. C. Bovik, and C. Charrier, “Blind prediction of natural video quality,” IEEE Trans. Image Process., vol. 23, no. 3, pp. 1352– 1365, 2014. [20] Z. Tu, Y. Wang, N. Birkbeck, B. Adsumilli, and A. C. Bovik, “UGCVQA: Benchmarking blind video quality assessment for user generated content,” IEEE Trans. Image Process., vol. 30, pp. 4449–4464, 2021. [21] Z. Tu, X. Yu, Y. Wang, N. Birkbeck, B. Adsumilli, and A. C. Bovik, “RAPIQUE: Rapid and accurate video quality prediction of user generated content,” arXiv preprint arXiv:2101.10955, 2021. [22] J. P. Ebenezer, Z. Shang, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “ChipQA: No-reference video quality prediction via space-time chips,” IEEE Trans. Image Process., vol. 30, pp. 8059–8074, 2021. [23] ——, “HDR-ChipQA: No-reference quality assessment for high dynamic range videos,” submitted to IEEE Trans. Image Process., 2023. [24] A. Mittal, R. Soundararajan, and A. C. Bovik, “Making a “completely blind” image quality analyzer,” IEEE Signal Process. Lett., vol. 20, no. 3, pp. 209–212, 2012. [25] J. Korhonen, “Two-level approach for no-reference consumer video quality assessment,” IEEE Trans. Image Process., vol. 28, no. 12, pp. 5923–5938, 2019. [26] ITU, “ BT.2020 : Parameter values for ultra-high definition television systems for production and international programme exchange,” Intl. Telecomm. Union, Tech. Rep., 2015. [27] S. Miller, M. Nezamabadi, and S. Daly, “Perceptual signal coding for more efficient usage of bit codes,” SMPTE Motion Imaging J., vol. 122, no. 4, pp. 52–59, 2013. [28] Rtings, Peak Brightness Measurement, 2023 (accessed February 28, 2023). [Online]. Available: https://www.rtings.com/tv/tests/ picture-quality/sdr-peak-brightness [29] ITU, “BT.500 : Methodologies for the subjective assessment of the quality of television images,” Intl. Telecomm. Union, Tech. Rep., 2019. [30] ——, “ITU. 910 : Subjective video quality assessment methods for multimedia applications,” Intl. Telecomm. Union, Tech. Rep., 2008. [31] Z. Wang, E. Simoncelli, and A. Bovik, “Multiscale structural similarity for image quality assessment,” in Asilomar Conf. Signals, Syst., Comput., vol. 2, 2003, pp. 1398–1402 Vol.2. [32] H. R. Sheikh, M. F. Sabir, and A. C. Bovik, “A statistical evaluation of recent full reference image quality assessment algorithms,” IEEE Trans. Image Process., vol. 15, no. 11, pp. 3440–3451, 2006. [33] J. P. Ebenezer, Z. Shang, Y. Wu, H. Wei, S. Sethuraman, and A. C. Bovik, “Making video quality assessment models robust to bit depth,” submitted to IEEE Signal Process. Letters, 2023. [34] J. P. Ebenezer, Z. Shang, Y. Wu, H. Wei, and A. C. Bovik, “Noreference video quality assessment using space-time chips,” in IEEE Intl. Workshop Multimedia Signal Process., 2020, pp. 1–6. [35] D. W. Dong and J. J. Atick, “Temporal decorrelation: a theory of lagged and nonlagged responses in the lateral geniculate nucleus,” Netw.: Comput. Neural Syst., vol. 6, no. 2, pp. 159–178, 1995. This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

HDR or SDR? A Study of Scaled and Compressed Videos: Conclusion, Acknowledgment, and References

HDR or SDR? A Study of Scaled and Compressed Videos: Objective Assessment

Authors:
(1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu);
(2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work;
(3) Yixu Chen, Amazon Prime Video;
(4) Yongjun Wu, Amazon Prime Video;
(5) Hai Wei,  Amazon Prime Video;
(6)Sriram Sethuraman, Amazon Prime Video;
(7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction
Related Work
Details of Subjective Study
Subjective Analysis
Objective Assessment
Conclusion, Acknowledgment and References Abstract—We conducted a large-scale study of human perceptual quality judgments of High Dynamic Range (HDR) and Standard Dynamic Range (SDR) videos subjected to scaling and compression levels and viewed on three different display devices. HDR videos are able to present wider color gamuts, better contrasts, and brighter whites and darker blacks than SDR videos. While conventional expectations are that HDR quality is better than SDR quality, we have found subject preference of HDR versus SDR depends heavily on the display device, as well as on resolution scaling and bitrate. To study this question, we collected more than 23,000 quality ratings from 67 volunteers who watched 356 videos on OLED, QLED, and LCD televisions. Since it is of interest to be able to measure the quality of videos under these scenarios, e.g. to inform decisions regarding scaling, compression, and SDR vs HDR, we tested several well-known full-reference and no-reference video quality models on the new database. Towards advancing progress on this problem, we also developed a novel no-reference model called HDRPatchMAX, that uses both classical and bit-depth sensitive distortion statistics more accurately than existing metrics. Index Terms—High dynamic range, video quality assessment, video compression I. INTRODUCTION High Dynamic Range (HDR) videos have utilize deeper bit-depths to represent brighter and darker luminance with wider color gamuts than Standard Dynamic Range (SDR) videos. To obtain the full benefits of HDR, however, a display must have the technology to accurately represent high contrasts and the extremes of the dynamic range. SDR videos are gamma-encoded using the power law described in BT 709 [1]. When shown on a TV, these values are decoded and adjusted to the TV’s display capabilities using a look-up table. Although SDR standards were originally created for cathode ray televisions having a maximum display brightness of 100 nits, modern display devices can use their entire brightness range (often much greater than 100 nits) to display SDR content, since the digital values of SDR content are relative and not scene-referred. The exact mapping between the SDR digital values and what the television displays are able to display differs across products and is ordinarily proprietary. Videos following the HDR10 standard, on the other hand, are scene-referred and absolute. The PQ EOTF (used in the HDR10 standard) specifies the absolute luminance that the display must show for a particular digital value. If the absolute luminance value that is required to be shown is greater than the highest luminance value that the TV can display, a tonemapping function called the Electrical-Electrical Transfer Function (EETF) is applied on the HDR content so that clipping does not occur at highlights, while ensuring a smooth roll-off of brightness values at the peak. EETFs differ among televisions and are also usually proprietary. Due to differences in how SDR and HDR signals are displayed on HDR-capable displays, an SDR version of a content may have a higher average brightness than the HDR version, depending on how each is graded and displayed. For example, a content having a maximum brightness of 200 nits in HDR may be graded in SDR such that the digital value of the maximum brightness is 255. When the HDR and SDR versions are displayed on an HDR-capable device having a peak brightness of 1000 nits, the SDR version may present a peak brightness much larger than 200 nits while the HDR version will be displayed with a peak brightness of 200 nits. The SDR version may therefore appear brighter, but it may also be washed out or oversaturated. Higher peak brightness or higher average brightness are not the only reasons why HDR content can be more appealing than SDR. Indeed, SDR videos may suffer from defects such as saturation, banding, lowcontrast, etc., which are less likely to occur in HDR videos. In addition to these differences, HDR videos use 10 bit representations (stored in 16 bits), while SDR videos have 8 bit representations. HDR videos require twice the number of bytes that SDR videos of the same content have. Because of this, HDR videos may be more susceptible to compression artifacts. The tradeoffs between compression, contrast-representation, color-representation, and brightness make the perceptual assessment of HDR and SDR video quality content-dependent and display-dependent. Towards better understanding these tradeoffs, we have conducted a detailed subjective and objective assessment of HDR and SDR videos having the same contents. For the subjective study, we recruited 67 participants who viewed and rated the qualities of 356 HDR and SDR videos of 25 source contents, which were processed by various combinations of scaling and compression using the x265 encoder. We also evaluate objective full-reference (FR) and no-reference (NR) video quality assessment (VQA) on the new subjective database. We also present the design of a new NR VQA model for the task of predicting the quality of both HDR and SDR videos. This paper is available on arxiv under CC 4.0 license. Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Authors: Authors: (1) Joshua P. Ebenezer, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work (e-mail: joshuaebenezer@utexas.edu); (2) Zaixi Shang, Student Member, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA, contributed equally to this work; (3) Yixu Chen, Amazon Prime Video; (4) Yongjun Wu, Amazon Prime Video; (5) Hai Wei,  Amazon Prime Video; (6)Sriram Sethuraman, Amazon Prime Video; (7) Alan C. Bovik, Fellow, IEEE, Laboratory for Image and Video Engineering, The University of Texas at Austin, Austin, TX, 78712, USA. Table of Links Abstract and Introduction Related Work Details of Subjective Study Subjective Analysis Objective Assessment Conclusion, Acknowledgment and References Abstract and Introduction Abstract and Introduction Related Work Related Work Details of Subjective Study Details of Subjective Study Subjective Analysis Subjective Analysis Objective Assessment Objective Assessment Conclusion, Acknowledgment and References Conclusion, Acknowledgment and References Abstract—We conducted a large-scale study of human perceptual quality judgments of High Dynamic Range (HDR) and Standard Dynamic Range (SDR) videos subjected to scaling and compression levels and viewed on three different display devices. HDR videos are able to present wider color gamuts, better contrasts, and brighter whites and darker blacks than SDR videos. While conventional expectations are that HDR quality is better than SDR quality, we have found subject preference of HDR versus SDR depends heavily on the display device, as well as on resolution scaling and bitrate. To study this question, we collected more than 23,000 quality ratings from 67 volunteers who watched 356 videos on OLED, QLED, and LCD televisions. Since it is of interest to be able to measure the quality of videos under these scenarios, e.g. to inform decisions regarding scaling, compression, and SDR vs HDR, we tested several well-known full-reference and no-reference video quality models on the new database. Towards advancing progress on this problem, we also developed a novel no-reference model called HDRPatchMAX, that uses both classical and bit-depth sensitive distortion statistics more accurately than existing metrics. Abstract —We conducted a large-scale study of human perceptual quality judgments of High Dynamic Range (HDR) and Standard Dynamic Range (SDR) videos subjected to scaling and compression levels and viewed on three different display devices. HDR videos are able to present wider color gamuts, better contrasts, and brighter whites and darker blacks than SDR videos. While conventional expectations are that HDR quality is better than SDR quality, we have found subject preference of HDR versus SDR depends heavily on the display device, as well as on resolution scaling and bitrate. To study this question, we collected more than 23,000 quality ratings from 67 volunteers who watched 356 videos on OLED, QLED, and LCD televisions. Since it is of interest to be able to measure the quality of videos under these scenarios, e.g. to inform decisions regarding scaling, compression, and SDR vs HDR, we tested several well-known full-reference and no-reference video quality models on the new database. Towards advancing progress on this problem, we also developed a novel no-reference model called HDRPatchMAX, that uses both classical and bit-depth sensitive distortion statistics more accurately than existing metrics. Abstract Index Terms—High dynamic range, video quality assessment, video compression Index Terms —High dynamic range, video quality assessment, video compression Index Terms I. INTRODUCTION High Dynamic Range (HDR) videos have utilize deeper bit-depths to represent brighter and darker luminance with wider color gamuts than Standard Dynamic Range (SDR) videos. To obtain the full benefits of HDR, however, a display must have the technology to accurately represent high contrasts and the extremes of the dynamic range. SDR videos are gamma-encoded using the power law described in BT 709 [1]. When shown on a TV, these values are decoded and adjusted to the TV’s display capabilities using a look-up table. Although SDR standards were originally created for cathode ray televisions having a maximum display brightness of 100 nits, modern display devices can use their entire brightness range (often much greater than 100 nits) to display SDR content, since the digital values of SDR content are relative and not scene-referred. The exact mapping between the SDR digital values and what the television displays are able to display differs across products and is ordinarily proprietary. Videos following the HDR10 standard, on the other hand, are scene-referred and absolute. The PQ EOTF (used in the HDR10 standard) specifies the absolute luminance that the display must show for a particular digital value. If the absolute luminance value that is required to be shown is greater than the highest luminance value that the TV can display, a tonemapping function called the Electrical-Electrical Transfer Function (EETF) is applied on the HDR content so that clipping does not occur at highlights, while ensuring a smooth roll-off of brightness values at the peak. EETFs differ among televisions and are also usually proprietary. Due to differences in how SDR and HDR signals are displayed on HDR-capable displays, an SDR version of a content may have a higher average brightness than the HDR version, depending on how each is graded and displayed. For example, a content having a maximum brightness of 200 nits in HDR may be graded in SDR such that the digital value of the maximum brightness is 255. When the HDR and SDR versions are displayed on an HDR-capable device having a peak brightness of 1000 nits, the SDR version may present a peak brightness much larger than 200 nits while the HDR version will be displayed with a peak brightness of 200 nits. The SDR version may therefore appear brighter, but it may also be washed out or oversaturated. Higher peak brightness or higher average brightness are not the only reasons why HDR content can be more appealing than SDR. Indeed, SDR videos may suffer from defects such as saturation, banding, lowcontrast, etc., which are less likely to occur in HDR videos. In addition to these differences, HDR videos use 10 bit representations (stored in 16 bits), while SDR videos have 8 bit representations. HDR videos require twice the number of bytes that SDR videos of the same content have. Because of this, HDR videos may be more susceptible to compression artifacts. The tradeoffs between compression, contrast-representation, color-representation, and brightness make the perceptual assessment of HDR and SDR video quality content-dependent and display-dependent. Towards better understanding these tradeoffs, we have conducted a detailed subjective and objective assessment of HDR and SDR videos having the same contents. For the subjective study, we recruited 67 participants who viewed and rated the qualities of 356 HDR and SDR videos of 25 source contents, which were processed by various combinations of scaling and compression using the x265 encoder. We also evaluate objective full-reference (FR) and no-reference (NR) video quality assessment (VQA) on the new subjective database. We also present the design of a new NR VQA model for the task of predicting the quality of both HDR and SDR videos. This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

HDR or SDR? A Study of Scaled and Compressed Videos: Abstract and Introduction

The Rise of Blockchain-based Video Streaming: Exclusive Interview With Replay CEO Michael Jelen

A recent study conducted by SFJESGS found out that “traditional streaming services have disrupted traditional broadcast television by providing a more personalized, on-demand, and flexible viewing experience. The study also found out that they changed how content is created, distributed, and consumed”. However, it is intriguing to note that within a short window of time, what were once hailed as disruptors are themselves increasingly facing disruption, especially with the emergence of blockchain technology. Fueled by opaque revenue models, exorbitant subscription costs, and content access fragmentation, the disruption is already underway. This article examines the future of video streaming as well as the shortcomings of both the traditional and blockchain-based video streaming models. It highlights how decentralized video streaming models leverage blockchain-based approaches to create a more equitable and transparent video economy for creators, viewers, and the entertainment industry as a whole. Challenges Facing Traditional Streaming Services Despite its huge potential, the existing video streaming model is anything but perfect. Limited transparency and accountability, exorbitant subscription fees, and fragmented content access are some of the pressing challenges limiting the growth potential of the traditional video streaming industry. Limited Transparency The lack of transparency in the distribution of revenue presents a disappointing picture of the opaque payment structures that content creators have to contend with. In most cases, the amount of efforts put into content creation does not commensurate with the rewards that accrue to it. For smaller creators whose livelihood depend on their content, this imbalance can lead to frustration and disappointment. Besides, since creators have no control over the metrics that are used in calculating their reported earnings, they are usually left in the dark, unsure of how their work translates to actual compensation. Exorbitant Subscription Fees Traditional streaming services are often criticized for charging exorbitant subscription fees, leaving viewers with little or no choice. The constant hike in subscriptions costs is majorly fueled by various video streaming platforms competing for dominance. Fragmented Content Access The traditional video streaming model lacks a standardized approach to content syndication. This is due to the fact that licensing agreements are often complex and vary greatly between platforms, leading to limited availability and fragmentation. It also restricts viewers' options and makes it difficult for creators to reach a wider audience. From the above, it is without a doubt to state that current system is built on limited transparency and accountability and fragmentation. These shortcomings provide a justification for the emergence of  decentralized models, which are not only cost-effective but also transparent as well. Blockchain-based Models There is an ongoing paradigm shift from traditional streaming services to decentralized models, which not only offers immediate financial benefits but also increased transparency and accountability. In a decentralized model, blockchain technology gives creators greater control over their work. Through smart contracts (self-executing agreements built on the blockchain), creators can determine how their content is displayed, distributed, and even monetized beyond traditional avenues. The potential of decentralized video streaming model is vast, especially because it fosters a more equitable ecosystem by empowering creators, offering viewers greater choice, and streamlining content distribution. Challenges Facing Blockchain-based Streaming Models Although blockchain-based models present exciting possibilities in the streaming ecosystem, navigating scalability, mass adoption, content security and regulatory hurdles remains crucial for their long-term success. These challenges reflect the current stage of the evolution of blockchain technology and should not be regarded as insurmountable obstacles. Scalability Scaling blockchain to efficiently handle vast amounts of data such as user information, streaming data, etc. remains a major challenge slowing down the market growth potential. Unlike traditional streaming services that can manage data from a large user base, decentralized models are often confronted with this hurdle . Although blockchain protocols like interplanetary file systems (IPFS) are used for data storage and distribution, Mass Adoption Enlightening users about blockchain technology and its advantages within the streaming ecosystem is crucial for its long-term success. Decentralized streaming platforms need to create user-friendly interfaces and educational resources which can streamline .  Without a clear understanding of how this technology works, onboarding new users might be prove to be extremely difficult. Content Security Protecting users against unauthorized access and copyright infringement underlines the need for decentralized streaming platforms to adopt robust security measures in tackling piracy. Without an assurance that contents uploaded to decentralized streaming platforms are secure, potential users might be reluctant to embrace this new paradigm. Regulation Regulatory frameworks for blockchain technology and cryptocurrencies are beginning to take shape in many countries across the world. Decentralized streaming platforms need to stay abreast of these developments and ensure their operations comply with the local regulations of the countries where they are registered. Technical Complexity Blockchain needs to be abstracted away into the background due to its complex nature. Users do not want to be bored with the technical details about it. Building a user-friendly streaming application is key to attracting and retaining viewers. Replay As a Case Study Having established the thesis that traditional video streaming services face challenges like complex licensing, lack of transparency, and limited viewer choice, it is crucial to examine Replay, a decentralized model, as a practical use case. Replay, a platform built on blockchain technology and generative AI, aims to address the lingering issues faced by traditional video streaming platforms. It leverages the power of blockchain to create a decentralized network. This means a secure and transparent digital ledger tracks every view and interaction with content. The blockchain-powered video streaming future it seeks to create is built on decentralized network, real time licensing and payment, pay-per-view-model and wide distribution network. Creators upload content and set their own licensing terms, eliminating the complexities of traditional licensing agreements. Additionally, secure payments are facilitated through RPLAY, Replay's native token, ensuring creators are fairly compensated for every view. For viewers, Replay offers a pay-per-view model, eliminating subscription fatigue. It also fosters a community-driven experience by encouraging interaction between creators and fans. This not only builds trust but also encourages participation in the decentralized ecosystem. The Future of Video Streaming The same also goes for decentralized platforms which may integrate with existing streaming platforms as a way to maximize user reach.  Chances are that the future of video streaming is a convergent one in which traditional players coexist with decentralized video streaming platforms by adopting blockchain elements in their operations. Regardless of what happens , the potential of blockchain-based streaming technology to revolutionize the entertainment landscape remains is enormous. Conclusion It is an indisputable fact that the traditional video streaming model is hampered by challenges that limit its potential for both creators and viewers. However, the emergence of blockchain technology offers a unique solution, which promises increased transparency, fairer revenue distribution, and wider content accessibility. Whether independent platforms are leading the charge or established players incorporating blockchain solutions, a new era in video streaming is already underway. This is the era that promises to usher in a more equitable and innovative entertainment landscape, where both creators and viewers have a greater stake in the success of the content they produce and consume. A recent study conducted by SFJESGS found out that “traditional streaming services have disrupted traditional broadcast television by providing a more personalized, on-demand, and flexible viewing experience. The study also found out that they changed how content is created, distributed, and consumed”. SFJESGS SFJESGS However, it is intriguing to note that within a short window of time, what were once hailed as disruptors are themselves increasingly facing disruption, especially with the emergence of blockchain technology. Fueled by opaque revenue models, exorbitant subscription costs, and content access fragmentation, the disruption is already underway. This article examines the future of video streaming as well as the shortcomings of both the traditional and blockchain-based video streaming models. It highlights how decentralized video streaming models leverage blockchain-based approaches to create a more equitable and transparent video economy for creators, viewers, and the entertainment industry as a whole. decentralized video streaming Challenges Facing Traditional Streaming Services Despite its huge potential, the existing video streaming model is anything but perfect. Limited transparency and accountability, exorbitant subscription fees, and fragmented content access are some of the pressing challenges limiting the growth potential of the traditional video streaming industry. Limited Transparency The lack of transparency in the distribution of revenue presents a disappointing picture of the opaque payment structures that content creators have to contend with. In most cases, the amount of efforts put into content creation does not commensurate with the rewards that accrue to it. For smaller creators whose livelihood depend on their content, this imbalance can lead to frustration and disappointment. Besides, since creators have no control over the metrics that are used in calculating their reported earnings, they are usually left in the dark, unsure of how their work translates to actual compensation. Exorbitant Subscription Fees Traditional streaming services are often criticized for charging exorbitant subscription fees, leaving viewers with little or no choice. The constant hike in subscriptions costs is majorly fueled by various video streaming platforms competing for dominance. Fragmented Content Access The traditional video streaming model lacks a standardized approach to content syndication. This is due to the fact that licensing agreements are often complex and vary greatly between platforms, leading to limited availability and fragmentation. It also restricts viewers' options and makes it difficult for creators to reach a wider audience. From the above, it is without a doubt to state that current system is built on limited transparency and accountability and fragmentation. These shortcomings provide a justification for the emergence of  decentralized models, which are not only cost-effective but also transparent as well. Blockchain-based Models There is an ongoing paradigm shift from traditional streaming services to decentralized models, which not only offers immediate financial benefits but also increased transparency and accountability. In a decentralized model, blockchain technology gives creators greater control over their work. Through smart contracts (self-executing agreements built on the blockchain), creators can determine how their content is displayed, distributed, and even monetized beyond traditional avenues. The potential of decentralized video streaming model is vast, especially because it fosters a more equitable ecosystem by empowering creators, offering viewers greater choice, and streamlining content distribution. Challenges Facing Blockchain-based Streaming Models Although blockchain-based models present exciting possibilities in the streaming ecosystem, navigating scalability, mass adoption, content security and regulatory hurdles remains crucial for their long-term success. These challenges reflect the current stage of the evolution of blockchain technology and should not be regarded as insurmountable obstacles. Scalability Scalability Scaling blockchain to efficiently handle vast amounts of data such as user information, streaming data, etc. remains a major challenge slowing down the market growth potential. Unlike traditional streaming services that can manage data from a large user base, decentralized models are often confronted with this hurdle . Although blockchain protocols like interplanetary file systems (IPFS) are used for data storage and distribution, Mass Adoption Mass Adoption Enlightening users about blockchain technology and its advantages within the streaming ecosystem is crucial for its long-term success. Decentralized streaming platforms need to create user-friendly interfaces and educational resources which can streamline .  Without a clear understanding of how this technology works, onboarding new users might be prove to be extremely difficult. Content Security Content Security Protecting users against unauthorized access and copyright infringement underlines the need for decentralized streaming platforms to adopt robust security measures in tackling piracy. Without an assurance that contents uploaded to decentralized streaming platforms are secure, potential users might be reluctant to embrace this new paradigm. Regulation Regulation Regulatory frameworks for blockchain technology and cryptocurrencies are beginning to take shape in many countries across the world. Decentralized streaming platforms need to stay abreast of these developments and ensure their operations comply with the local regulations of the countries where they are registered. Technical Complexity Technical Complexity Blockchain needs to be abstracted away into the background due to its complex nature. Users do not want to be bored with the technical details about it. Building a user-friendly streaming application is key to attracting and retaining viewers. Replay As a Case Study Replay As a Case Study Having established the thesis that traditional video streaming services face challenges like complex licensing, lack of transparency, and limited viewer choice, it is crucial to examine Replay , a decentralized model, as a practical use case. Replay Replay , a platform built on blockchain technology and generative AI, aims to address the lingering issues faced by traditional video streaming platforms. It leverages the power of blockchain to create a decentralized network. This means a secure and transparent digital ledger tracks every view and interaction with content. The blockchain-powered video streaming future it seeks to create is built on decentralized network, real time licensing and payment, pay-per-view-model and wide distribution network. Replay Creators upload content and set their own licensing terms, eliminating the complexities of traditional licensing agreements. Additionally, secure payments are facilitated through RPLAY , Replay's native token, ensuring creators are fairly compensated for every view. RPLAY For viewers, Replay offers a pay-per-view model, eliminating subscription fatigue. It also fosters a community-driven experience by encouraging interaction between creators and fans. This not only builds trust but also encourages participation in the decentralized ecosystem. The Future of Video Streaming The same also goes for decentralized platforms which may integrate with existing streaming platforms as a way to maximize user reach.  Chances are that the future of video streaming is a convergent one in which traditional players coexist with decentralized video streaming platforms by adopting blockchain elements in their operations. Regardless of what happens , the potential of blockchain-based streaming technology to revolutionize the entertainment landscape remains is enormous. Conclusion It is an indisputable fact that the traditional video streaming model is hampered by challenges that limit its potential for both creators and viewers. However, the emergence of blockchain technology offers a unique solution, which promises increased transparency, fairer revenue distribution, and wider content accessibility. Whether independent platforms are leading the charge or established players incorporating blockchain solutions, a new era in video streaming is already underway. This is the era that promises to usher in a more equitable and innovative entertainment landscape, where both creators and viewers have a greater stake in the success of the content they produce and consume.

What Does the Future of Video Streaming Look Like?

Authors:
(1) Lukáš Korel, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic;
(2) Petr Pulc, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic;
(3) Jirí Tumpach, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic;
(4) Martin Holena, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Table of Links Abstract and Introduction ANN-Based Scene Classification Methodology Experiments Conclusion and Future Research, Acknowledgments and References 5 Conclusion and Future Research In this paper was provided an insight into the possibility of using artificial neural networks for scene recognition location from a video sequence with a small set of repeated shooting locations (such as in television series) was provided. Our idea was to select more than one frame from each scene and classify the scene using that sequence of frames. We used a pretrained VGG19 network without two last layers. This results were used as an input to the trainable part our neural network architecture. We have designed six neural network models with different layer types. We have investigated different neural network layers to combine video frames, in particular average-pooling, max-pooling, product, flatten, LSTM, and bidirectional LSTM layers. The considered networks have been tested and compared on a dataset obtained from The Big Bang Theory television series. The model with max-pooling layer was not successful, its accuracy was the lowest of all models. The models with a flatten or product layer were very unstable, their standard deviation was very large. The most stable among all models was the one with an average-pooling layer. The models with unidirectional LSTM and bidirectional LSTM had similar standard deviation of the accuracy. The model with a bidirectional LSTM had the highest accuracy among all considered models. In our opinion, this is because its internal memory cells preserve information in both directions. Those results shows, that models with internal memory are able to classify with a higher accuracy than models without internal memory. Our method may have limitations due to the chosen pretrained ANN and the low dimension of some neural layer parts. In future research, it is desirable to achieve higher accuracy in scene location recognition. This task may also need modifying model parameters or using other architectures. It also may need other pretrained models or combining several pretrained models. It is also desirable that, if the ANN detects an unknown scene, it will remember it and next time it will recognize a scene from the same location properly. Acknowledgments The research reported in this paper has been supported by the Czech Science Foundation (GACR) grant 18-18080S. Computational resources were supplied by the project "e-Infrastruktura CZ" (e-INFRA LM2018140) provided within the program Projects of Large Research, Development and Innovations Infrastructures. Computational resources were provided by the ELIXIRCZ project (LM2018131), part of the international ELIXIR infrastructure. References [1] Zhong, W., Kjellström, H.: Movie scene recognition with Convolutional Neural Networks. https://www.diva-portal.org/smash/get/diva2 :859486/FULLTEXT01.pdf KTH ROYAL INSTITUTE OF TECHNOLOGY (2015) 5–36 [2] Simonyan, K., Zisserman, A.: Very Deep Convolutional Networks for Large-Scale Image Recognition. https://arxiv.org/pdf/1409.1556v6.pdf Visual Geometry Group, Department of Engineering Science, University of Oxford (2015) [3] Russakovsky O., Deng J., Hao S., Krause J., Satheesh S., Ma Sean, Huang Z., Karpathy A., Khosla A., Bernstein M., Berg A. C., Fei-Fei L.: ImageNet Large Scale Visual Recognition Challenge. International Journal of Computer Vision 115 (2015), pp. 211–252 [4] Li-jia L., Hao S., Fei-fei L., Xing E.: A High-Level Image Representation for Scene Classification & Semantic Feature Sparsification. https://cs.stanford.edu/groups/vision/pdf/ LiSuXingFeiFeiNIPS2010.pdf NIPS (2010) [5] Felix A. Gers, Schmidhuber J., and Cummins F., Learning to forget: Continual prediction with LSTM, in Proceedings of ICANN. ENNS (1999), pp. 850-–855 [6] Benavoli A., Corani G., Mangili F.: Should We Really Use Post-Hoc Tests Based on Mean-Ranks? Journal of Machine Learning Research 17 (2016), pp. 1–10 [7] García S., Herrera F.: An Extension on "Statistical Comparisons of Classifiers over Multiple Data Sets" for all Pairwise Comparisons Journal of Machine Learning Research 9 (2008), pp. 2677–2694 [8] Graves, A.: Supervised Sequence Labelling with Recurrent Neural Networks. Springer (2012) [9] Kaiming H., Xiangyu Z., Shaoqing R., Jian S.: Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (2016), pp. 770–778 [10] Sudha V., Ganeshbabu T. R.: A Convolutional Neural Network Classifier VGG-19 Architecture for Lesion Detection and Grading in Diabetic Retinopathy Based on Deep Learning. http://www.techscience.com/cmc/v66n1/40483 Computers, Materials & Continua (2021), pp. 827–842 [11] Zhou B., Lapedriza A., Khosla A., Oliva A., Torralba A.: Places: A 10 Million Image Database for Scene Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence (2018), pp. 1452–1464 This paper is available on arxiv under CC0 1.0 DEED license. Authors: (1) Lukáš Korel, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (2) Petr Pulc, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (3) Jirí Tumpach, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic; (4) Martin Holena, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Authors: Authors: (1) Lukáš Korel, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (2) Petr Pulc, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (3) Jirí Tumpach, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic; (4) Martin Holena, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Table of Links Abstract and Introduction Abstract and Introduction ANN-Based Scene Classification ANN-Based Scene Classification Methodology Methodology Experiments Experiments Conclusion and Future Research, Acknowledgments and References Conclusion and Future Research, Acknowledgments and References 5 Conclusion and Future Research In this paper was provided an insight into the possibility of using artificial neural networks for scene recognition location from a video sequence with a small set of repeated shooting locations (such as in television series) was provided. Our idea was to select more than one frame from each scene and classify the scene using that sequence of frames. We used a pretrained VGG19 network without two last layers. This results were used as an input to the trainable part our neural network architecture. We have designed six neural network models with different layer types. We have investigated different neural network layers to combine video frames, in particular average-pooling, max-pooling, product, flatten, LSTM, and bidirectional LSTM layers. The considered networks have been tested and compared on a dataset obtained from The Big Bang Theory television series. The model with max-pooling layer was not successful, its accuracy was the lowest of all models. The models with a flatten or product layer were very unstable, their standard deviation was very large. The most stable among all models was the one with an average-pooling layer. The models with unidirectional LSTM and bidirectional LSTM had similar standard deviation of the accuracy. The model with a bidirectional LSTM had the highest accuracy among all considered models. In our opinion, this is because its internal memory cells preserve information in both directions. Those results shows, that models with internal memory are able to classify with a higher accuracy than models without internal memory. Our method may have limitations due to the chosen pretrained ANN and the low dimension of some neural layer parts. In future research, it is desirable to achieve higher accuracy in scene location recognition. This task may also need modifying model parameters or using other architectures. It also may need other pretrained models or combining several pretrained models. It is also desirable that, if the ANN detects an unknown scene, it will remember it and next time it will recognize a scene from the same location properly. Acknowledgments The research reported in this paper has been supported by the Czech Science Foundation (GACR) grant 18-18080S. Computational resources were supplied by the project "e-Infrastruktura CZ" (e-INFRA LM2018140) provided within the program Projects of Large Research, Development and Innovations Infrastructures. Computational resources were provided by the ELIXIRCZ project (LM2018131), part of the international ELIXIR infrastructure. References [1] Zhong, W., Kjellström, H.: Movie scene recognition with Convolutional Neural Networks. https://www.diva-portal.org/smash/get/diva2 :859486/FULLTEXT01.pdf KTH ROYAL INSTITUTE OF TECHNOLOGY (2015) 5–36 [2] Simonyan, K., Zisserman, A.: Very Deep Convolutional Networks for Large-Scale Image Recognition. https://arxiv.org/pdf/1409.1556v6.pdf Visual Geometry Group, Department of Engineering Science, University of Oxford (2015) [3] Russakovsky O., Deng J., Hao S., Krause J., Satheesh S., Ma Sean, Huang Z., Karpathy A., Khosla A., Bernstein M., Berg A. C., Fei-Fei L.: ImageNet Large Scale Visual Recognition Challenge. International Journal of Computer Vision 115 (2015), pp. 211–252 [4] Li-jia L., Hao S., Fei-fei L., Xing E.: A High-Level Image Representation for Scene Classification & Semantic Feature Sparsification. https://cs.stanford.edu/groups/vision/pdf/ LiSuXingFeiFeiNIPS2010.pdf NIPS (2010) [5] Felix A. Gers, Schmidhuber J., and Cummins F., Learning to forget: Continual prediction with LSTM, in Proceedings of ICANN. ENNS (1999), pp. 850-–855 [6] Benavoli A., Corani G., Mangili F.: Should We Really Use Post-Hoc Tests Based on Mean-Ranks? Journal of Machine Learning Research 17 (2016), pp. 1–10 [7] García S., Herrera F.: An Extension on "Statistical Comparisons of Classifiers over Multiple Data Sets" for all Pairwise Comparisons Journal of Machine Learning Research 9 (2008), pp. 2677–2694 [8] Graves, A.: Supervised Sequence Labelling with Recurrent Neural Networks. Springer (2012) [9] Kaiming H., Xiangyu Z., Shaoqing R., Jian S.: Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (2016), pp. 770–778 [10] Sudha V., Ganeshbabu T. R.: A Convolutional Neural Network Classifier VGG-19 Architecture for Lesion Detection and Grading in Diabetic Retinopathy Based on Deep Learning. http://www.techscience.com/cmc/v66n1/40483 Computers, Materials & Continua (2021), pp. 827–842 [11] Zhou B., Lapedriza A., Khosla A., Oliva A., Torralba A.: Places: A 10 Million Image Database for Scene Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence (2018), pp. 1452–1464 This paper is available on arxiv under CC0 1.0 DEED license. This paper is available on arxiv under CC0 1.0 DEED license. available on arxiv

Video Scene Location Recognition Using AI: Conclusion and Future Research

Authors:
(1) Lukáš Korel, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic;
(2) Petr Pulc, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic;
(3) Jirí Tumpach, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic;
(4) Martin Holena, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Table of Links Abstract and Introduction ANN-Based Scene Classification Methodology Experiments Conclusion and Future Research, Acknowledgments and References 4 Experiments 4.1 Experimental Setup The ANNs for scene location classification were implemented in the libraries Python language using TensorFlow and Keras. Neural network training was accelerated using a NVIDIA GPU. The versions of the employed hardware and software are listed in Table 1. For image preparation, OpenCV and Numpy were used. The routine for preparing frames is a generator. It has lower capacity requirements, because data are loaded just in time when they are needed and memory is released after the data have been used for ANN. All non-image information about inputs (video location, scenes information, etc.) are processed in text format by Pandas. We have 17 independent datasets prepared by ourselves from proprietary videos of the The Big Bang Theory series, thus the datasets can’t be public. Each dataset originates from one episode of the series. Each experiment was trained with one dataset, so results are independent as well. So we can compare behavior of the models with different datasets. Our algorithm to select data in training routine is based on oversampling. It randomly selects target class and from the whole training dataset is randomly select source scene with replacement. This algorithm is applied due to an unbalanced proportion of different target classes. Thanks to this method, all targets are distributed equally and the network does not overfit a highly represented class. 4.2 Results Summary statistics of the predictive accuracy of classification all 17 episode datasets are in Table 3. Every experiment was performed on every dataset at least 7 times. The table is complemented with results for individual episodes, depicted in box plots. The model with a max-pooling layer had the worst results (Figure 12) of all experiments. Its overall mean accuracy was around 10 %. This is only slighty higher than random choice which is 1/12. The model was not able to achieve better accuracy than 20 %. Its results were stable and standard deviation was very low. Slightly better results (Figure 10) had the model with the a flatten layer, it was sometimes able to achieve a high accuracy, but its standard deviation was very high. On the other hand, results for some other episodes were not better than those of the max-pooling model. A better solution is the product model, whose predictive accuracy (Figure 9) was for several episodes higher than 80 %. On the other hand, other episodes had only slightly better results than the flatten model. And it had the highest standard deviation among all considered models. The most stable results (Figure 11) with good accuracy had the model based on average-pooling layer. Its mean accuracy was 32 % and for no episode, the accuracy was substantially different. The model with unidirectional LSTM layer had the second mean accuracy of considered our models (Figure 13). Its internal memory brings advantage in compare over the previous approaches, over 40 %, though also a comparatively high standard deviation. The highest mean accuracy had the model with a bidirectional LSTM layer (Figure 14). It had a similar standard deviation as the one with a unidirectional LSTM, but an accuracy mean nearly 50 %. This paper is available on arxiv under CC0 1.0 DEED license. Authors: (1) Lukáš Korel, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (2) Petr Pulc, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (3) Jirí Tumpach, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic; (4) Martin Holena, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Authors: Authors: (1) Lukáš Korel, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (2) Petr Pulc, Faculty of Information Technology, Czech Technical University, Prague, Czech Republic; (3) Jirí Tumpach, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic; (4) Martin Holena, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Table of Links Abstract and Introduction Abstract and Introduction ANN-Based Scene Classification ANN-Based Scene Classification Methodology Methodology Experiments Experiments Conclusion and Future Research, Acknowledgments and References Conclusion and Future Research, Acknowledgments and References 4 Experiments 4.1 Experimental Setup The ANNs for scene location classification were implemented in the libraries Python language using TensorFlow and Keras. Neural network training was accelerated using a NVIDIA GPU. The versions of the employed hardware and software are listed in Table 1. For image preparation, OpenCV and Numpy were used. The routine for preparing frames is a generator. It has lower capacity requirements, because data are loaded just in time when they are needed and memory is released after the data have been used for ANN. All non-image information about inputs (video location, scenes information, etc.) are processed in text format by Pandas. We have 17 independent datasets prepared by ourselves from proprietary videos of the The Big Bang Theory series, thus the datasets can’t be public. Each dataset originates from one episode of the series. Each experiment was trained with one dataset, so results are independent as well. So we can compare behavior of the models with different datasets. Our algorithm to select data in training routine is based on oversampling. It randomly selects target class and from the whole training dataset is randomly select source scene with replacement. This algorithm is applied due to an unbalanced proportion of different target classes. Thanks to this method, all targets are distributed equally and the network does not overfit a highly represented class. 4.2 Results Summary statistics of the predictive accuracy of classification all 17 episode datasets are in Table 3. Every experiment was performed on every dataset at least 7 times. The table is complemented with results for individual episodes, depicted in box plots. The model with a max-pooling layer had the worst results (Figure 12) of all experiments. Its overall mean accuracy was around 10 %. This is only slighty higher than random choice which is 1/12. The model was not able to achieve better accuracy than 20 %. Its results were stable and standard deviation was very low. Slightly better results (Figure 10) had the model with the a flatten layer, it was sometimes able to achieve a high accuracy, but its standard deviation was very high. On the other hand, results for some other episodes were not better than those of the max-pooling model. A better solution is the product model, whose predictive accuracy (Figure 9) was for several episodes higher than 80 %. On the other hand, other episodes had only slightly better results than the flatten model. And it had the highest standard deviation among all considered models. The most stable results (Figure 11) with good accuracy had the model based on average-pooling layer. Its mean accuracy was 32 % and for no episode, the accuracy was substantially different. The model with unidirectional LSTM layer had the second mean accuracy of considered our models (Figure 13). Its internal memory brings advantage in compare over the previous approaches, over 40 %, though also a comparatively high standard deviation. The highest mean accuracy had the model with a bidirectional LSTM layer (Figure 14). It had a similar standard deviation as the one with a unidirectional LSTM, but an accuracy mean nearly 50 %. This paper is available on arxiv under CC0 1.0 DEED license. This paper is available on arxiv under CC0 1.0 DEED license. available on arxiv

Stories, News & Mentions about VIDIO

Vidio’s Stories on HackerNoon

Vidio's News & Mentions