Core demands
In two important scenarios, click travel needs to transmit the text order information to the client through voice broadcast (voice synthesis) to improve the interaction convenience and security during driving.
One is the "listening function" of taxi business, which enables taxi drivers to accurately receive new order content through voice broadcast;
One is the "order listening function+order dispatching service function" of the downwind business, in which the order dispatching service will have high requirements for concurrency.
Based on the above business requirements, it needs to rely on the high accuracy of voice broadcast, the guarantee of concurrency, and clear and natural pronunciation.
Solution
[Scenario 1]: Taxi business "listening function"
In the taxi business of Dida Trip, taxi drivers can open the listening function through the client APP. The Dida travel platform will send the appropriate taxi order to the driver according to the current position of the owner and other factors. The client will remind the driver that a new order request has arrived. The driver can respond and choose to grab the order or ignore it. In addition to the traditional visual interaction, the client side reminder also provides voice broadcast reminders. The platform transmits the information that needs voice broadcast to the client through text, and the client calls Baidu voice synthesis function for real-time broadcast.
In this scenario, it is not a novel practice for our product to choose voice broadcast. In the design of traditional travel products, this has become the default and standard practice.
There are two main reasons for using voice broadcast:
One is to improve the interactive experience. Taxi drivers have certain learning costs for their proficiency in using digital equipment; Because of their age, professional habits and other reasons, they have high requirements for the recognition of words and graphics. In terms of product design, designers not only need to introduce special fonts and design elements in UI design to strengthen, but also increase voice as an interactive way to enhance the recognition of order information by drivers.
Another important reason is security. Most drivers actually listen to the list during driving. Visual interaction is not only unsafe, but also not allowed in the safety regulations of many countries.
[Scenario 2]: Shunfengche business order listening function+order dispatching service
In the downwind car business, after the certification of private car owners, they can select the real-time order listening function to receive the latest downwind order requirements. At this time, the core appeal of taxi drivers is almost the same. Dida Chuxing uses voice broadcast for order distribution service, which is one of the services with the highest performance requirements in the whole platform. At the same time, the higher the number of online car owners, the larger the number of concurrent orders, and the higher the number of voice announcements. According to past data, the peak value of voice broadcast will easily exceed 10Kqps, and the peak hourly call volume may reach tens of millions or even hundreds of millions of times. Therefore, performance is an important consideration.
Baidu speech synthesis technology can convert the text input by users into smooth and natural speech output, and can support the setting of speech speed, tone, volume, and audio code rate, breaking the traditional text based human-computer interaction mode, making human-computer communication more natural.
Application examples
The following is the core product process of Baidu Voice Synthesis for Dida Chuxing:
Step 1: After the taxi owner has registered as a certified owner, enter the first screen of the app "Tick Taxi Driver", click the "Depart" button to start listening to orders and receive orders nearby.
Step 2: The car owner can modify the settings of listening at any time. In addition to the settings of listening for travel, the car owner can also turn on and off the voice broadcast function at any time.
Step 3: When the car owner receives a new order nearby, the APP will pop up the details of the order, and at the same time, the APP will call Baidu Voice Service to broadcast the details of the order, including the starting point of the journey and the broadcast.