Himalaya FM is a well-known audio sharing platform with a total user size of 480 million. The APP was launched in March 2013. In more than two years, the number of users has exceeded 200 million. It has become the fastest growing and largest online mobile audio sharing platform in China. In 2014, two rounds of high-value financing were completed, laying a solid financial strength for further leading the Chinese audio field. As of December 2015, the total amount of Himalayan audio has exceeded 15 million, and the cumulative number of playbacks in a single day has exceeded 50 million. The market share in the mobile audio industry has reached 73%.
The traditional voice live broadcast and text interaction methods have been unable to meet the interaction between the audiences and the Internet Celebrities. The method has evolved from a one-to-one interaction mode to a chat room mode where people interact with each other by voice. The gameplay also needs to develop from a single radio station and listening broadcast to more online interactive gameplay.
Compared with ordinary single live broadcast scenarios, voice-linked live broadcast has two big technical difficulties.
The first is the high latency requirement of the connect scene, which cannot be met by the RTMP protocol suitable for live broadcasting. The second is the higher sound quality requirements for voice live broadcasting. The echo cancellation during connect interaction and the suppression of noise in noisy environments. The volume gain of human voice in near-field and far-field interactions are several major problems in audio processing in real-time speech. Traditional audio processing technology mainly relies on hardware. It is difficult to be compatible with all types of mobile terminals, and the hardware modules are not flexible enough, which can easily lead to distortion of the audio signal and affect the sound quality.
Zego currently provides Himalaya FM with a multi-person voice link solution and an online KTV solution. The pure self-developed private protocol based on UDP guarantees a stream transmission with a delay of about 200 milliseconds. meanwhile, the method of intelligently smoothing network jitter, forward error correction, and frame loss compensation are used to solve the problem of packet loss in the weak network of the UDP protocol. Experience to ensure low latency and clear and smooth voice calls.
At the same time, Zego's self-developed 3A algorithm solves the problem of audio processing in the voice scene. Adaptive echo cancellation, automatic adjustment of microphone volume, combined with psychoacoustic models, and improvement of 20dB + signal-to-noise ratio allow users to have high performance in various complex environments Sound quality experience.
1. More ways of playing games (scenarios show, funny dubbing, singing solo, etc.) derived from the voice of users connecting have increased the fun of live broadcast. Promote the realization of platform traffic and increase competitiveness in the highly competitive live broadcast industry.
2. At present, the users of audiobook platforms generally do not pay much attention to the anchors. Through the way of anchor live broadcast and audience interaction, it brings a large number of fan users and increases the stickiness of the platform users.
3. Anchors don't need to buy professional sound card equipment. They can have good sound quality effects only through mobile phones, which lowers the entry threshold for anchors and attracts more anchors to the platform.
Scenario 1: Two anchors connect to solo , and the audiences can apply for interaction
Scenario 2: Multi-anchor enters the room, then turn around to sing, and other team games, etc.