With video information and UHD becoming the dominant trends, 8K UHD technology has showed remarkable improvements in resolution, frame rate, color depth, color gamut, high dynamic range (HDR) and multi-channel audio, delivering a more striking and immersive user experience.
8K is mainly used in the entertainment industry such as live event streaming, digital signs, movie theaters and VR videos/gaming, as well as in vertical industries such as video surveillance, video conferencing, remote health care and industrial inspection. At present, the development of 8K technology can be divided into the plane, VR and holographic stages.
The plane stage: The 8K display approaching the limit of the human eye mainly improves the user experience from the color gamut, frame rate and contrast. The 8K end-to-end industry chain is basically available, but 8K content is still scarce due to limited production time and costs.
The VR stage: When moving from theplane stage to the immersion stage, 8K resolution is just the beginning. 8K VR delivers a similar viewing experience as that of 480p TV. To achieve the same experience as that of a 4K TV, 24K VR is required. To cope with the pressure on transmission bandwidth caused by massive concurrent data streams and to keep the service latency less than 20 ms, content delivery networks (CDNs) need to move closer to the network edge. In addition, VR projection mapping and rendering impose extremely high requirements on computing capability, so cloud VR becomes the best choice for VR.
Holographic display: Currently, VR content is only displayed on the spherical surface, which easily causes visual fatigue. In the future, holographic display will be the ultimate display technology that reproduces realistic 3D images. The emerging light field technology is considered as the ideal solution for a holographic display.
The above-mentioned development trends show that 8K imposes higher requirements on video production, distribution, and display.
8K video production involve key technologies such as projection mapping and encoding & transcoding.
Projection mapping: The current 8K VR content usually adopts equirectangular projection (ERP) or cubemap projection (CMP), both of which have seriously uneven projection saturation. This causes the waste of pixels in some regions and insufficient pixel density in some other regions, which degrades the video quality. Therefore, to enhance video quality, the non-uniform mapping technology, such as equi-angular cubemap (EAC) projection, is preferred, which helps distribute pixels evenly over a sphere.
Encoding & transcoding algorithm: H.265 as the video compression standard is used for 8K video. However, the high bitrates between 80 and 150 Mbps required by 8K video are not beneficial for storage and transmission. AV1 has not been put into large-scale commercial use yet. The H.266 and AVS3 algorithms, with standards still being formulated, are far from mature commercialization. Therefore, ZTE has optimized the mainstream H.265 algorithm and developed the H.265S video encoding solution based on the intelligent analysis framework. Compared to H.265, H.265S can reduce bitrate by more than 30% while offering the same user experience. It is fully compatible with H.265 and is transparent to downstream products such as CDN and terminals. With superior encoding performance, it won Video Compression Great Challenge at the IEEE International Conference on Image Processing (ICIP) recently.
8K distribution involves key technologies such as CDN sinking, low-latency transmission, and transmission bandwidth optimization.
CDN sinking: To avoid congestion of backbone network and guarantee VR service latency, CDN must be deployed at the edge of the metropolitan area network (MAN) close to the users.
Given different network types and sinking locations, the sinking CDNs can be divided into three types: MEC-vCDN, BRAS-CDN, and OLT-CDN. MEC-based vCDN deployment is mainly used for 5G and fixed-mobile converged networks, and it provides services close to users based on technologies such as containerization, edge scheduling, scheduling and orchestration synergy, QoS assurance, hotspot distribution optimization algorithm, and open capabilities. BRAS-CDN refers to the CDN embedded in the BRAS of the fixed network to save space. OLT-CDN refers to the CDN deployed at OLT with no extra space and power required while supporting IPoE and PPPoE user services.
ZTE is the only vendor in the industry that provides BRAS-CDN and OLT-CDN products.
Low latency transmission: Due to defects in the adaptive bitrate (ABR) implementation mechanism and TCP protocol, the end-to-end transmission latency is relatively high, resulting in a poor user experience. The mainstream low-latency technologies being studied include common media application format (CMAF), quick UDP internet connection (QUIC), and low-latency HTTP live streaming (LL-HLS).
Transmission bandwidth optimization: The transmission bandwidth optimization technology reduces the transmission bandwidth for a single user or a group of users.
For single-user VR transmission, the main development trends are field of view (FOV) and FOV+. FOV transmission overlaps a low quality background stream with a FOV stream to ensure high quality within the FOV while reducing the quality outside the FOV to reduce the transmission bitrate and performance requirements on the server and client. In 2018, the MPEG has released the omnidirectional media application format (OMAF) v1.0 specification, which includes FOV. FOV+ transmits images with a slightly larger angle than FOV to cope with the network and processing latency. Compared to FOV, this technology consumes less transmission bandwidth. However, when a user turns the head very quickly, there will be blank areas in an image. The FOV prediction technology is required to predict the user's head movements so as to transmit images in advance.
The technology to reduce multi-user transmission bandwidth is moving towards mABR+M-FOV. The mABR technology converts the unicast ABR protocol into the multicast protocol for multicast transmission. M-FOV is the multicast transmission technology for VR FOV. It transmits the background stream for users watching the same channel via multicast and FOV stream for each user via unicast.
For 8K videos, the ultra-high performance video server is essential to make content distribution more economical. At present, the throughput of a general server is usually 20 Gbps, while the bandwidth required by an 8K user is 150-200 Mbps. For massive concurrent users, this is not economical in terms of construction costs, space of access office, and energy consumption. This year, ZTE is about to launch its ultra-high performance video server specifically developed for 8K/VR, which can provide the industry leading throughput of 200 Gbps.
Besides the above-mentioned key technologies, 8K display also involves image enhancement. This new technology uses frame interpolation and super-resolution to enhance video quality. At this stage when 8K content is limited, terminals can use AI technologies to improve the frame rate and resolution of 4K or even HD videos to bring a better viewing experience.
Based on years of experience in the video field, ZTE has provided domestic and foreign customers with highly competitive Big Video solutions for individuals and vertical industries. These solutions have been widely used in the industry. At the 2018 World Conference on VR Industry, ZTE and China Telecom teamed for live broadcast of the beautiful scenery of Lake Baiyangdian in Xiong'an New Area with 5G+8K VR. In March 2019, ZTE and the Shanghai Branch of China Telecom broadcast the 26th Music Festival in China live with 5G+4K and 8K+VR. In August 2019, ZTE and the Shanxi Branch of China Mobile have collaborated to deliver China's first 5G sporting event with the broadcast of the country's second National Youth Games.
The continuous success of 8K trials will bring about not only a brand new video experience but also more and more business models, thereby promoting the rapid development of 8K services.