Nuance: Car Information Platform Steps into the Voice Age

Nuance: Natural Voice Technology for Human-Computer Interaction
From left to right, Zhang Yaxi, Director of Shanghai R&D Center, Nuance Communications, Arnd Weil, Vice President of Global Automotive Business Unit, and Zheng Yuqing, General Manager of Greater China.

Gasgoo.com: Please tell us about Nuance's automotive business and its development in China.

Arnd Weil: Simply put, Nuance provides mobile hands-free solutions for making calls, sending messages, reading messages, playing music, selecting songs, and navigating through voice operations. In terms of navigation, for example, users report a location and our product can conduct a route search.

The automotive business is a business of Nuance Mobile's Automotive Business Unit. The range of products and services of the automotive business department includes not only in-vehicle application solutions, but also the development of interconnected services and on-board solutions, and the integration of user interface design with server-side in-car applications and interconnection services. At present, more than 35 million vehicles in the front-loading and after-loading markets already use Nuance voice technology.

Nuance has been working with international manufacturers and has entered the Chinese market through companies such as Continental and Bosch. Companies that have cooperated in our automotive business over the past 10 years include companies such as Ford, Daimler, BMW, Audi, General Motors, and Toyota. We have worked with Ford since 10 years ago on voice technology and are now partners with direct suppliers of voice technology and user interface design.

Nuance attaches great importance to the Chinese market. Last year it began to cooperate with Chinese auto manufacturers. We have established a professional team to research and develop new products in the Chinese market, like Shanghai's Voice Technology Development Center. Apart from local engineers participating in the development process, the global professional team is also developing voice solutions that are more suitable for the Chinese market.

Gasgoo.com: What are the technical advantages of Nuance?

Arnd Weil: Nuance provides a very good voice user interface and vehicle human-machine interface, supports more than 30 languages, whether it is a car application connection terminal, server terminal or networking services, we can support. Nuance also offers multi-modal input technology, including handwriting and smart text input for touch screens.

In terms of voice, Nuance voice technology has been certified and tested in many countries, which proves that Nuance's identification technology is quite accurate. Nuance can provide very good support for languages ​​in different countries. The accumulated experience and successful cases over the years are a good proof. This is something other competitors cannot surpass.

Gasgoo.com: How much recognition is Nuance's speech recognition technology for accented Mandarin?

Zhang Yayi : When Nuance started doing Chinese speech recognition from the very beginning, he realized that China has a large area and a variety of accent. However, it is unrealistic to do speech recognition for each local dialect. In addition, the Chinese government has been encouraging the country to implement Putonghua. Therefore, we still strive to improve Putonghua recognition technology. So we started from the stage of collecting data, collecting sound samples from all over China, south to Guangzhou, Fujian, and north to Heilongjiang and even northwest. Therefore, our speech recognition has a high recognition rate for various accented Mandarin speakers.

There were customers who took our engine and compared it with another company. In terms of accent testing, our competitors had very different recognition rates on different people's tests. Most people may have relatively standard Putonghua. The accent is not standard, and the curve fluctuation is obvious. However, Nuance's test curve changes very little, and the difference between individuals is very small, indicating that we have done very well in accent coverage.

Gasgoo.com: Please talk about the difficulties in research and development of speech recognition technology and future trends.

Zhang Yayu: I think the difficulties and trends will be discussed together because the difficulties are what we will overcome in the future and will be the direction we will develop in the future. Traditional speech recognition restricts the command word. For example, if you define “turn on air conditioner” in the command, you can only say “turn on the air conditioner”, say more, say less one word, or, in other words, it cannot be identified. This is actually a restriction on people. The user must remember each command word very accurately. If the definition of the command word is more and more in the future, it is difficult for ordinary users to strictly note each command word.

So in the solution, we use voice technology to allow users to easily and intuitively communicate and interact with each other. We only define tasks and do not define command words, as long as you express the task in your own way. This technology, called natural language understanding, has been applied to Ford vehicles sold in the North American market and is implemented in the SYNC system of the new generation of MyFord Touch technology.

Zheng Yuqing: In fact, we have overcome the difficulty of natural language processing. We have an engine located in the call center that can handle whatever you say and can handle it accordingly. However, the engines used in cars and mobile phones are relatively small. Once compressed, the recognition rate will be limited. Therefore, how to further improve the speech recognition performance of the on-board system is one of the goals of our current work.

Gasgoo.com: At what level is the price of natural voice technology?

Zheng Yuqing: For the high-end market, we can provide a so-called "one shot" solution, which is a word input, and the system will analyze what you want to do. This cost is relatively high. If you just call or control music, the price will be much cheaper. So we use different functions to locate market prices.

Gasgoo.com: There are different market positions.

Zheng Yuqing: Right . For example, dialing the phone through voice control can also be implemented in the low-end car, and the price will be relatively low. However, there are some cars that contain the entire system, including air conditioning, music, navigation, and cloud services. The price is relatively high. Of course, these only appear in high-end cars. Because in fact you have to do so many things, in addition to our software technology prices will be high, other things like memory, CPU, the overall hardware configuration will be increased accordingly.

Gasgoo.com: Looking at the current business of Nuance, how much of the three major business segments account for voice, textual intelligence input, and image solutions?

Zheng Yuqing: Voice business is the core, accounting for 85%. You can see that we covered mobile phones and cars. What we can do now is e-books and computers. Next we have to do IPTV, interactive Internet TV. There are also great things in call centers and medical care. The medical aspect is to use voice to enter the patient's case and file it. Therefore, voice is the most important business of Nuance.

Gasgoo.com: What kind of market strategy did Nuance adopt when promoting car audio systems in China?

Zheng Yuqing: Chinese users are pursuing high quality and low price, so we adjust our strategies under the conditions they provide, implement more flexible pricing strategies in China, and price according to market demand. In terms of support services, we will select some relatively good manufacturers, invest in our manpower, and spend time with them to develop corresponding products according to their needs, so as to ensure that products come out with high quality and very high customer satisfaction.

Gasgoo.com: Nuance recently conducted a user survey of car voice applications in China. Can you share some of the survey results?

Zhang Yayi: We recently conducted a survey on car GPS users in China. Several hundred Chinese car drivers participated in our research, mainly focusing on whether there are voice systems and frequency of use in the car. Nearly 30% of the cars have voice control, which exceeds my expectations. In fact, many people are interested in voice control and voice dialing. For example, if the air conditioner rises twice, the radio is tuned to 97.7Hz, or if you call Zhangsan, they are very interested in this order.

Gasgoo.com: Should this be what you expected?

Zhang Yayu: This is what I expected, but one thing was beyond my expectation. 43.5% of people surveyed will send text messages while driving. This is very dangerous. Of these, 15.5% are texting while driving, and 28% are saying that they are texting while waiting for the red light. But you can imagine that if they turn green, they will continue to be unfinished and unsafe.

Zheng Yuqing: So I spent a lot of time waiting for the traffic light. The green light in front of the car has not yet gone. It is very likely that I am texting or calling.

Zhang Yayi: Many foreign countries have already issued laws and regulations that strictly prohibit the use of cell phones and text messages during driving. However, there are so many people in China who drive while text messaging is unthinkable.

There is one more problem than I expected. Car voice recognition because of the special environment inside the car, the noise is relatively large, the recognition rate will be low, the effect is not as quiet, but still 82.3% think it is qualified.

Zheng Yuqing: The current application of car voice recognition technology is basically a high-end car. Because I also know that some domestic manufacturers configure voice recognition on low-end cars, the effect is very bad. In fact, this standard has not been reached. Nowadays, some users cooperate with us because they have used some domestic technology in the past, and then find that it is really not working. Let us find out again. In fact, the threshold is relatively high, and this threshold cannot be met to meet the ultimate needs of users. For example, Ford in the United States advertises its voice control system. By changing the user experience, it feels safer while sitting up and it has more selling points. I believe that in the future, many car companies in China will go in this direction.

Car Washing Machine MachineFor Bussiness

car wash machine, car washing machine for bussiness,car wash, commercial car wash

Zhengzhou Shinewash Technology Co.,Ltd , https://www.shinewashtech.com