How was ＂Autonavi Gang＂ voice navigation born？

[Editor’s note] Recently, the Autonavi map released a laughing version. In addition to the ability to routinely broadcast basic navigation actions such as "slowing down, turning, and reaching the destination", the biggest feature is that it can prompt real-time traffic information in Guo Degang’s unique teasing language according to the scene.

Many people are curious: why Guo Degang? Is "Turn right at the intersection ahead, you must not go wrong" a complete sentence that has been recorded long ago? The following article is based on the oral content of "Flower Underpants", one of Autonavi’s map product managers.

Driving navigation has now become an important channel for people to interact with mobile maps, and because of the uniqueness of driving scenarios, voice navigation has become an important function for mobile maps to play a guiding role. We signed Lin Zhiling as a spokesperson before and created the concept of "celebrity voice navigation". Lin Zhiling, as the "Muse" spokesperson of Autonavi maps, has a very unique voice navigation to help Autonavi maps enhance the reputation of many brands, but many female friends have expressed the need for a new voice other than Lin Zhiling.

Driving is actually a very boring and boring thing many times, especially in Beijing. If Lin Zhiling’s standard "baby voice" is not your cup of tea, then you may need a new navigation voice to "refresh".

Our cooperation with Guo Degang started in the second half of last year, and the selection of Guo Degang is actually a win-win situation: as a travel map application with more than 300 million users, Guo Degang’s endorsement is a strong addition to its own exposure; and for Autonavi, Guo Degang’s identity as a famous cross talk artist is in line with the product characteristics – professional and funny. Internet tool applications have developed to this day, and there is no much differentiation in basic capabilities. Humanization, entertainment, and personalization have become new ways to differentiate competition. Guo Degang is undoubtedly an excellent choice to generate chemical reactions with Autonavi maps.

I remember that in a navigation voice star selection poll held by Baidu Maps at the end of last year, Lin Zhiling Guo Degang won the first or second place. In fact, before Baidu Maps launched its research, we had already done relevant research. Combined with the product characteristics of navigation voice, we had already foreseen such results. The list of artists who are suitable for navigation and have appeal is actually not many, but it also proves that the two we signed do represent the needs and voices of users.

On December 18 last year, the happy navigation conference of Autonavi Maps was officially held. On that day, we witnessed the DAU of Autonavi Maps hit a record high. The whole city and major social media were talking about Guo Degang providing voice navigation for Autonavi Maps. We knew at that time that we signed Guo Degang correctly, and "Autonavi Gang" became another successful Internet product endorsement case. Since the two parties announced their cooperation on December 18, 2014, the product active level of Autonavi Maps has continued to rise – among them, the rate of new users has increased by 230%, the number of daily active users has increased by 36%, and the highest ranking of Autonavi Maps on the Apple App Store list has also rushed to the second place.

Many users are wondering why Guo Degang and Lin Zhiling can broadcast such complex place names, roads, and navigational movements for us. In fact, we use a TTS Text To Speech technology.

In the early stages of production, we would record some of the star’s fixed sentence pronunciation in the studio. There were about 2,000 such key sentences. Lin Zhiling spent three days recording in Taiwan at that time, and Teacher Guo also recorded it for nearly three days, which was very hard. Then, we will analyze their sentence and voice, extract their respective pronunciation rules, voiceprint details, tone and intonation and other characteristics, and use a series of complex algorithms to carry out Text To Speech, in order to fill in the lack of various "variable words" in the navigation voice, so as to form a complete and smooth star voice navigation voice.

In comparison to Lin Zhiling, Mr. Guo’s voice navigation version also added various funny jokes, such as "There are vehicles in front, don’t touch the cars that shouldn’t be touched!", "There are many accidents in front, and the tanks are all knocked over!" and so on, so Mr. Guo’s workload is also particularly large, and the content of all kinds of jokes is also brainstormed by various departments of our Autonavi map and the majority of netizens. Through the participation of Mr. Guo Degang, determine the jokes to be put into the navigation voice, and reduce them to an appropriate length to decide which navigation actions to appear in.

At the same time, we have also launched several popular dialect navigation voice packages, which are indeed personalized needs for some users. However, according to user feedback, the synthesis effect of Lin Zhiling, Guo Degang, and dialect navigation voice is still very good, especially the star voice, which has star characteristics and a high degree of synthesis reduction. Fans of Lin Zhiling and Guo Degang also like it, and they are very sure of our technical level.

At present, Autonavi is also the only Internet map manufacturer in China that can provide full voice navigation of celebrities. Friends are indeed lacking in this regard. They spend a lot of money to sign celebrities but only use 7 sentences from others, and switch voice guidance back and forth with standard female voices, which sounds very divisive.

There is another episode: this upgrade of Guo Degang’s navigation voice laughter version not only enhances the quantity and quality of Guo Degang’s sections, but also adds a random section function to various important navigation scenarios. For example, in the sharp turn prompt scenario, a total of 5 prompt sections in the voice library are randomly selected and played, which can increase the fun and make users feel happy. But it is such a seemingly simple change that places great pressure on research and development: the addition of random voice functions requires major changes or even rewrites to the traditional architecture and logic of the SDK, and as a PM, it has been under great pressure, but in the end, the R & D team completed the launch of this function under the impetus of cans of X cattle (and PM) – so a good voice navigation also requires a motivated and willing team.

Recently, Lin Chiling was ridiculed as "Navigation Muse" when she participated in the Dragon TV show "Sister Pattern", which surprised us who did not put this show at all. Therefore, celebrity endorsement should be like this: to achieve a win-win situation for the brand power of both stars and products, that is, to establish a strong connection between the public, and become each other’s attribution and label.

But at the same time, I also think that a good voice navigation is not just Lin Zhiling and Guo Degang. When AutoNavi released Guo Degang’s version of the voice navigation laughter version on March 13, Dong Zhenning, vice president of Autonavi, said: "If we want to do a professional voice navigation, we need to pay attention to four points. The first point is precision; the second point is intelligence; the third is humanization; and the fourth is entertainment."

We believe that the "four modernizations" should be a progressive relationship: precision refers to the fact that the underlying data supporting the navigation module must be complete and accurate, which is the foundation of the product. Autonavi has nearly 13 years of professional experience in this field; and on the basis of accurate data, it can use big data technology implementation "intelligence" to provide pain point functions such as avoiding congestion and route selection planning; humanization is the pursuit of the ultimate experience, such as the timing of voice prompts is just right, and the threshold for user understanding is low. And the final entertainment is to introduce various elements such as celebrity fun voice, so that users can enjoy professional services while also feeling happy from the heart.

Leifeng.com’s original articles are prohibited from being reproduced without authorization. For details, please refer to the reprinting instructions.