Background
WeSing is the world's first karaoke singing application for social networking between friends. It has a variety of features such as the singing practice mode, competition, group chat, championship, and UGSV. With close to 10 million MAUs and a penetration rate of about 80% of users in the online karaoke market, WeSing is an unrivaled application in the pan-entertainment industry.
Challenges and Objectives
- WeSing has the largest online karaoke user base in China, with features like local recording, queuing up to sing, singing in rotation, and real-time chorus. As users now have higher requirements for real-timeness and interactivity, the traditional serial chorus solution has various drawbacks such as high latency and complexity. Therefore, WeSing is in dire need of a genuinely real-time chorus solution.
- Challenged by the fierce competitions in the pan-entertainment industry, WeSing needs to go beyond its major business of online karaoke to offer a wealth of premium audio/video live streaming services, such as audio chat, show live streaming, and co-anchoring, with the aim to meet the diverse needs of countless users for live streaming.
- Feed stream design is crucial for social products, including WeSing. WeSing faces several challenges, such as challenges in relationship chain diffusion due to the existence of users who have millions of followers and challenges in ensuring the exposure of important feeds by using complex control strategies due to various types of feed services.
- Different strategies are used to control the output of feed streams to implement various features. The strategies include frequency control on the exposure of feeds published by key opinion leaders (KOLs) to prevent click farms, merging feeds of game playing interaction published by friends to avoid spamming, supporting search for feeds in different categories, filtering user feeds that must be removed due to security issues, recommending real-time insertion/mixing of streams, and frequency control on the exposure of low-quality feeds and system-generated feeds.
Tencent Cloud Solutions
Based on WeSing's needs, Tencent Cloud provides WeSing with a one-stop audio/ video solution including products like TRTC, live streaming, and VOD to enable them to deliver a stable, smooth, and interactive live streaming.
- 192 Kbps dual-channel stereo and AI-based noise cancellation: Support from Tencent Ethereal Audio Lab with the same 3A engine used by Tencent Meeting, which delivers a brand-new Hi-Fi online karaoke experience.
- An extensive music library: 200,000 songs in diverse genres and a pay-as-you-go billing mode.
- A rich set of features: Queuing, song request/selection/pinning, vocal/accompaniment switch and volume adjustment, lyric sync, etc.; Accurate clock sync algorithm: Multiple time syncs with NTP, ensuring accurate and synchronous background music playback across terminal devices.
- A genuinely real-time chorus solution with end-to-end ultra low latency down to 70 ms, which perfectly reproduces the physical karaoke experience.
- Cross-application audio/video stream interconnection, enabling cross-application co-anchoring competition.
Furthermore, WeSing has adopted TencentDB for MongoDB as its primary storage service because this database supports high availability, distributed storage, high performance, high compression, schema-free structure, and sound strategies to balance client access load. This greatly facilitates the rapid iteration of WeSing.
Cooperation Achievements
Tencent Cloud's RT-ONE? audiovisual network provides a robust technology base for WeSing, enabling it to deliver HD, smooth, reliable, and low-latency live streaming services in various scenarios such as online karaoke, radio, and show live streaming. This can fully meet the live streaming needs of large numbers of users of WeSing. In addition, innovative features like real-time chorus and cross-server co-anchoring enhances user retention and payment rates.