It is very common to see people speak with their family members or friends through a live video call rather than an ordinary phone call. It is so popular that you can see live video call features everywhere on internet platforms or apps. For example, you are allowed to do live video chat with a newly known cyber-friend, or you can do a live video chat with your colleagues just through a video conference. Most of the live video chat apps are free.
You may not be able to recall just since when it has been so popular, and wonder just why it has been so common in our daily lives. There are various reasons behind this phenomenon. However, the major reasons are as follows:
Firstly, the upgrade of the telecommunication data network from 3G to 4G makes the network infrastructure ready for live video chat. There are two technical thresholds: network bandwidth must be more than 2M bps or even more to accommodate 720P high definition; the one-way latency must be lower than 300 ms to make people feel the video chat is live or real-time. 4G network just be adequate to satisfy the two requirements.
Secondly, the high penetration of smartphones has enabled almost everyone to access the Internet. Smartphones are more powerful even than ordinary PCs in terms of computational power. It is powerful enough to adopt H.264 or even H.265 to encode and decode video streaming data in the blink of an eye.
Thirdly, the prosperity of the mobile internet has fostered various applications with live video chat features, and in turn, these applications have educated internet users in forming the habit of using live video calls.
In this article, we are going to introduce the common live video chat scenarios, and how the apps with the scenarios make money. In addition, we will dive into a bit of the real-time communication technology behind a live video chat or conversation. Moreover, we will share how to build a live video chat feature into your own app. It feels so fantastic to make a live video call with your friends through your own app rather than popular apps like Whatsapp or Facebook. We will use the technology knowledge base of ZEGOCLOUD since it provides live video chat SDK for building live video call apps.
The Live Video Chat Scenarios
Social Networking
It is a very broad scope. It includes social apps, online dating, video chat rooms, etc. These apps allow users to make cyber-friends or find a partner for an intimate partner on the internet. Normally, they offer full features for real-time interactions, including instant messaging, voice, video, etc.
Video Conference
It is a very standard scenario for professional or office communication. It allows users to communicate or collaborate in real-time through all means, including text messaging, voice and video chat, whiteboard, screen sharing, document sharing, etc. It is made for a business scenario, thus its style is more professional, neat, and clean.
Online Education
It is quite like a video conference, but it is different in a few ways. It may include certain features designed for educational purposes. For example, it offers an intelligent feature to detect the status of students and find out who is sleepy or daydreaming. Moreover, there are various class modes including small class and large class. In a large class, it will transform into live streaming, and a live video chat offers real-time interaction between the teacher and students.
It is designed for entertainment or training purposes. For example, in a streaming program, the host may invite guests to speak and perform together live. During the live video chat session, the host and the guests can interact in a live video chat with background music. The audience can watch the live streaming, and send text messages or even virtual gifts to interact.
The Business Models of Live Video Chat
The business model refers to the way how the platforms make money in live video chat scenarios. Most of the social platforms or live video streaming offer free live video chat, while online education and video conference offer premium modes.
There are a few ways for the social networking scenario to make money. For example, users can pay to learn more personal information about the ones who they are interested in, and they can send virtual gifts during the live video chat session to delight their partners.
The business model of video conferences is quite straightforward and mature, users subscribe to the video conference service and pay monthly. Normally, enterprises will pay for their employees.
The business model of online education is quite standard too. Students prepay to buy course packages and then consume them. Since it is prepaid mode, the platforms secure their cash flow at the very beginning.
The business model of the live streaming scenario is almost the same as social networking. Users pay to purchase virtual gifts and send them to the hosts who they like. To encourage users to buy and send more virtual gifts, platforms organize events or competitions to drive revenue growth.
The Technology Behind Live Video Chat
Technical Logics
It seems pretty simple and straightforward to use a live video chat app, however, the real-time communication technology behind it is complicated. Basically, it requires that the sender end can capture voice and video data, then send it to the receiver end, and finally, it is rendered on the screen to display. The whole system won’t work in this simple way. There are a lot of factors to work on to make it a real-time framework.
For example, the bandwidth sources are limited and will incur costs, thus we have to compress the media data to make it consume less bandwidth. As a norm, we use H.264 to encode the media data. If the terminal devices support H.265, we can use H.265 to achieve better video and voice quality at the same level of bandwidth. At the receiver end, we have to decode the streaming data for rendering purposes. We will explore more at a bird’s view in the below paragraphs.
Front End
The frontends include the sender end and the receiver end. A live video stream starts from the sender’s end and is rendered to display at the receiver’s end. The whole process includes capturing, pre-processing, encoding, transmission, decoding, post-processing, and finally rendering. There is a real-time feedback mechanism to drive the process to work. The sender’s end will adjust the sending strategy to keep a nice balance between real-time and smoothness. The receiver’s end will maintain a jitter buffer to ensure smoothness, and also provide real-time feedback to the sender’s end through TCP-based signals.
The frontends can be native apps or web browsers. Normally, the performance of a live video chat is better on native apps than that on web browsers. A live video chat application can manipulate system-level APIs to achieve native-level performance. A web browser uses WebRTC to enable real-time video communication. Your web application has to access system-level APIs through the web browser and WebRTC, and compromises performance and quality.
Back End
There may not be backend servers for live video chat. You can transmit live video data with server acceleration, and you can also choose to transmit it directly in a peer-to-peer manner. For example, WebRTC doesn’t provide media servers for data transition, instead, it transmits video and voice data directly between two users in a live video chat. It provides ICE, STUNT, and TRUN servers, but these servers are designed for firewall transverse and transmission relays.
Live video chat SDK vendors like ZEGOCLOUD take another course. ZEGOCLOUD builds a cloud-based data transmission network for media acceleration. ZEGOCLOUD’s data network guarantees inter-datacenter bandwidth and computational resources. More importantly, ZEGOCLOUD adopts an intelligent strategy to ensure nearby access and dynamic smart routing.
Live video chat includes the features of live video and voice. They are discussed together very often though, they are so different that the algorithms used to proceed with them are totally different.
Real-time Voice
To improve the quality of real-time voice, you have to ensure the bit rate of the voice stream is higher than a certain level, let’s say 100k bps. If the voice bit rate is higher enough, it will give leeway to increase sampling frequency and bit depth. In addition, you need to remove negative factors such as noise and echoes. More specifically, you need to develop algorithms for acoustic echo cancellation and acoustic noise suppression. Moreover, to mitigate the negative effect created when users change the distance between their mouth and the microphone, you need to develop an algorithm for automatic gain control.
Real-time Video
A video stream consists of a stream of video frames ordered chronologically. Video information can be arranged in two dimensions,i.e., time dimension and space dimension. We use frame rate to measure the number of frames that are rendered per second, namely FPS (frame per second). The higher the FPS is, the smoother the video will look. The FPS of movies is 24, and the FPS of a live streaming show is about 15 FPS. Normally we use the resolution to define the space dimension of a video frame. People often say that a video is 720P or 1080P. A 720P video frame contains 921,600 pixels ( width = 1280, height=720). A 1080P video frame contains 2,073,600 pixels (width = 1920, height = 1080).
Media Data Transmission
The whole process of media data transmission is complicated. However, it can be decomposed into data transmission of each hop between two data centers. Data transmission between two data centers is conducted by a UDP-based transportation protocol. There is a feedback mechanism. The receiver informs the jittering condition to the sender, and the sender can predict how the bandwidth will change, and decide how much bandwidth its data package will consume. If the network condition is good, the sender will encode to produce data packages that consume more bandwidth. If the network condition is predicted to be worsening, the sender will encode to produce data packages that consume less bandwidth.
Nearby Access and Smart Routing
To accelerate transmission, ZEGOCLOUD adopts smart strategies for nearby access and routing. We inquire and find out a user’s location and telecom operator by the user’s smartphone IP address, and direct the user’s app to contact a nearby data center. Being directed by a scheduling center, the smart routing algorithm will find an optimal route for the transmission. Normally, a one-way transmission can be done within 2-3 hops.
How to Build a Live Video Chat Feature
Build from The Ground Up
This is a really hard mode. You will find yourself lost in nowhere if you don’t have experience and expertise in it at all. It requires professionals with expertise in a few major modules, including voice engine, video engine, and transport.
The voice engine consists of a few sub-modules, i.e., encoding/decoding, anti-jitter for voice, echo cancellation, noise reduction, automatic gain control, audio capturing, and rendering. The video engine is composed of several sub-modules too, i.e., encoding/decoding, jitter buffering for video, image enhancements, video capturing, and video rendering. The transport module consists of two parts, a UDP-based real-time media data transport protocol, and a TCP-based signaling protocol. You will have to set up a minimal 6-member team. The leader should have full-stack expertise in the aforementioned technology. The other 5 engineers will cover the voice engine, the video engine, transport, development on Android, and development on iOS.
Build Based on WebRTC or FFmpeg
This is a mild approach. Basically, you can study how to build from the source code of open-source projects. You can even build your live video chat technology based on the code base of these two projects. However, the difficulty level is not less than building it from scratch since you have to dig into the source code of the projects, which is very difficult and obscure. In the field, there is a saying believed by experts, i.e., “if you develop a deep understanding of WebRTC, you can build the technology from scratch without WebRTC.” If I flip it the other way around, it would become “if you don’t understand WebRTC with insights, you will be stuck in nowhere.” Even if you create something that allows you to make a working live video call, you will be in a nightmare when there is a troubleshooting task or new requirements coming. The code base will become unmaintainable gradually, Let’s say, one day the key engineer leaves you.
Build with ZEGOCLOUD Live Video SDK
Is there an easy and effective way? Surely yes! With the development of RTC technology, cloud-based real-time communication technology vendors like ZEGOCLOUD emerged to resolve the industrial problem. With ZEGOCLOUD’s live video SDK, you will be guaranteed to complete integration and start live video chat in your own app within a couple of hours.
ZEGOCLOUD’s live video SDK has encapsulated the aforementioned key RTC modules, including the voice engine, the video engine, and the transport module. You don’t have to build the technology on your own and can reduce technology investment significantly.
To ensure QoS of transmission, ZEGOCLOUD has built a global coverage data network called MSDN (Massive Serial Data Network). The MSDN possesses more than 500 BGP data centers in more than 212 countries or territories. With ZEGOCLOUD’s live video SDK and services, you don’t need to construct the data transmission network, which requires a huge amount of upright investment and technology expertise.
The ZEGOCLOUD MSDN allows you to grow your business scale at your own pace. It can accommodate a sharp increment or reduction of usage volume. You won’t feel any pressure or technical issues on your side even if you expand or contract your user base rapidly. If you build the data network infrastructure on your own, you will face a dramatic loss in investment if there is a contraction of the business scale.
To help you integrate ZEGOCLOUD live video SDK rapidly, the ZEGOCLOUD team has offered an express version of its live video SDK called ZEGOCLOUD live video express SDK. As its name shows, it allows you to integrate the SDK in an express manner. It is great for developers, but it is not good enough. The express SDK doesn’t offer standard GUI components, and developers still need to spend some time developing the GUIs of their apps. To resolve this issue, ZEGOCLOUD launches a UIKit edition of the live video SDK. On top of being fast integrated, it offers modulized GUIs components, and developers can construct the GUIs of their app in a way like building a model car with lego blocks.
We fully understand that there is a requirement for professional knowledge of real-time voice and video. We set up a technical service team who will be available conveniently 24 hours x 7 days. The experts and professionals of the technical support team will guide you through the integration process smoothly. What you need to do is to “Sign up” or leave your contact information at the “Talk to us” button on ZEGOCLOUD’s official website.
Summary
There are various ways to build live video chat apps. There are easy ones and tough ones. The common practice of the industry is to use vendor solutions and focus on your own core business. Anyway, you don’t have to invent the wheel from the ground up nowadays. ZEGOCLOUD has been in the field for 7 years and has served and been trusted by more than 4000 clients globally. If you want to build a live video chat APP, ZEGOCLOUD will be your best choice.
WhatsApp
Wechat
Book a Meeting
Submit a Request
global@zegocloud.com
