Media Processing

Wangsu’s intelligent cloud video platform enables AI and big data technologies in all video service processes and implements product intelligence based on operational intelligence. With the support of content delivery networks worldwide, it realizes product intelligence across the entire video chain, from video production to processing, delivery and playback, and provides full support for VOD (short and long videos), live streaming and other services.

Leveraging multi-machine, multi-server farm and cross-regional distributed clusters, Wangsu’s media processing system possesses the characteristics of stability, high efficiency and high scalability. Through Wangsu’s proprietary media processing technology, the system supports the mutual conversion of common video formats and all types of resolutions and code rates, provides offline transcoding, real-time transcoding and other services, and supports intelligent HD, video resolution reconstruction and other effects. It mainly applies to VR videos, ultra HD videos, low picture quality videos and video business scenarios that require transcoding and resolution conversion.

Highlights

Dynamic special effect video cover: Supports common static covers, as well as customized dynamic special effect covers to more effectively and visually display video features that help users to identify video contents more accurately.

Resolution reconstruction: Deep learning-based video ultra-resolution technology which can make in-depth reconstruction of video images on playback devices, break resolution constraints, enriching image details, automatically restoring image quality thus enabling low-resolution videos to deliver a high-resolution viewing experience.

AI video enhancement: Through deep learning-based AI video enhancement technology, intelligently enhance image effects and definition; remove burrs, artifacts, noises and mosaics, to improve the viewing experience.

Intelligently Controlled HD
a) Lossless bit reduction: Through machine learning and business scenario training, optimizing redundant algorithms, removes pixel data which do not affect picture quality and reducing bandwidth costs to provide users with the best experience.
b) Intelligent transcoding: Intelligently identify video business type and image content, selects the best parameters based on the human visual model, and supports rapid transcoding.
c) Renovation of old films: Through intelligent sharpening, adaptive noise reduction, contrast enhancement, ultra-high resolution and picture restoration technologies, old films no longer suffer from poor quality.

H.265 efficient coding: Supports H.264/H.265 coding and enable adaptive matching between H.264 and H.265. H.265 coding can achieve lower bit rates under the same picture quality and better picture quality under the same bit rate.

VR on-demand transmission: Taking advantage of the sensory limitations of the human eye, optimizes model algorithms through machine learning and VR business scenario training, and dynamically adjusts the effective pixel ratio and definition under primary and non-primary visual fields to save transmission bandwidth, improve transmission efficiency and solve the difficult issue of high VR video costs and high network environment requirements without affecting user experience.

Audiovisual unicast: Split audios and videos for the purpose of playing them separately and only playing audios in weak networks to meet user demand. It is applicable to dubbing, lip synchronization and other application scenarios.

Adding subtitles: Add subtitles to videos, which is applicable to interviews, large galas and other scenarios.

Live stream recording: Record wonderful live streaming contents and record them as VOD to expand live streaming resources for later use and attract raise awareness of the platform.


Comparison of video effects between H.265 coding and H.264 coding

H.265 possesses better compression performance. Under the same picture quality and the same bit rate, H.265 takes up less storage space than H.264. Under the same storage space and the same code rate, H.265 achieves better picture quality than H.264.


Customer