In just two weeks, the online claw doll has changed from an industry making money silently into an industry with a full range of interpretation from hardware to software architecture, from profit to investment, from operation to promotion. Only H5 claw doll (specifically refers to mobile Web, wechat claw doll), is still very mysterious.
Is the H5 claw doll really that mysterious? Here you can find out all about it.
What is the H5 claw doll that most people say?
H5 claw doll, refers to support HTML5 Web side claw doll. Can be divided into two categories: PC and mobile.
-
H5 grabber on PC, as long as the browser that supports WebRTC, can use Agora Web SDK directly. Of course, you can also use WebRTC self – developed, solve a number of server transmission, device adaptation, echo, availability and other problems, can be used commercially. PC terminal H5 claw doll, has a very mature solution.
-
Mobile terminal H5 grab doll, more complicated. Mobile TERMINAL H5 claw doll refers to two kinds: mobile web browser and wechat web page claw doll. The demand of most claw doll manufacturers is to promote web claw doll on wechat, not mobile browsers other than wechat.
So, the problem is very clear, most people say H5 grab doll refers to wechat built-in browser grab doll, below for the sake of narrative convenience, referred to as WECHAT H5 grab doll.
What is the difference between Wechat H5 doll grabbing and Naive App doll grabbing?
At present, the mature PC end Web claw doll, is through WebRTC to achieve. Most browsers have good support for WebRTC. But the Built-in browser for wechat on iOS does not support WebRTC.
Three. Wechat H5 claw doll how to achieve?
Next, with the voice network Agora’s wechat H5 scheme, to uncover the mystery of wechat H5 claw doll.
As previously said, wechat H5 to solve the last link – the operation end play video. The strategy adopted by the sound network is to use JSMpeg to play in the wechat browser.
JSMpeg is a video player implemented in JavaScript. It includes an MPEG separator, MPEG1 video and MP2 audio decoders, WebGL and Canvas2D rendering and WebAudio sound output. JSMpeg is relatively small and works fine on most browsers, decoding 720P video at 30fps on the iPhone 5S.
Since JSMpeg only supports MPEG1 format, a transcoding Server is added in the decoding process to convert the video format to MPEG1. Then the video is distributed to the wechat browser through the relay Server and played through JSMpeg.
For other links, the wechat H5 scheme is completely consistent with the non-wechat H5 scheme.
- At the doll end, two cameras are connected through the motherboard or PC to collect video data.
- Video streams are optimized with Agora’s patented encoder.
- Real-time video transmission via Agora’s globally deployed real-time virtual communications network, SD-RTN™
- Finally, it reaches the operation end to decode and play
- The operation end sends the control instruction to the doll machine end through the business Server, and obtains real-time feedback through the video stream.
- A high latency RTMP stream can be streamed from SD-RTN™ to the audience
As can be seen from the technical architecture diagram, wechat H5 claw doll is finally realized through mature open source projects, which is also the realization way of most mainstream wechat H5 claw doll. So, what is the difference in deciding the wechat H5 claw doll experience?
Iv. What are the differences in deciding the wechat H5 claw doll experience?
There are three key aspects to the experience:
- Time delay
- availability
- Service stability
- Time delay
The delay includes the instruction delay from the operation end to the doll machine, and the video delay from the doll machine to the operation end.
- Instruction delay: Due to the operation of the instructions from the end to the doll machine, the data packet is very small, so the delay is very low, generally tens of milliseconds.
- Video delay: If the video delay is less than 400ms, the player can accept it.
The key to deciding on the grab doll delay experience is the video delay.
Video clarity and time delay are two antithetical things. In the case of constant bandwidth, the clearer the video, the higher the frame rate, the higher the bit rate, the higher the delay. Therefore, there is a balance to be struck between video clarity and time delay.
The sound network uses proprietary codecs to reduce the bit rate as much as possible while maintaining clarity. Weak network countermeasures, such as packet loss retransmission, FEC, bandwidth detection, and dynamic bit rate adjustment, ensure smooth network experience even when the network quality is poor.
The sD-RTN ™ communication network consists of nearly 100 nodes deployed worldwide. The SD-RTN™ system can automatically allocate the optimal and most smooth transmission path based on the connection and transmission status, load status, distance to the user and response time of each node in real time, achieving the quality assurance level required for real-time transmission.
- availability
According to Callstats. IO, 89 percent of webrTC-based real-time communications have been successfully established in the United States. Soundnet’s SD-RTN™ has increased the login success rate to 99 percent by placing dense sites in each area.
- Service stability
With an average 400ms delay and a stable 400ms delay, the user experience is different. Stable 400ms latency, as the user expects, is a much better experience than 200 and 800.
The key problem after the quick launch of claw doll is retention. Aside from the business model, it’s the quality of the experience that determines retention. Abandoning user experience for speed is putting the cart before the horse. Returning to quality and user experience is the way to long-term development.
Docking document: document. Agora. IO/cn / 1.14 / tut…
If you are interested in us, or encounter development problems, please visit the Agora q&A panel and communicate with the engineers.