When intelligent voice meets car system and kitchen who is fake demand?

Novelty, fun, and impractical-this is the general impression that many people have of intelligent voice products.

From listening to human TV remote control to speak the speaker, more and more products began to play the "intelligent man-machine interaction" flag, however, we really like the ads said that the need for these products?

A living room or a talking kitchen?

In the author's recent random interview, most people are faced with "talking in the living room to talk in the kitchen, you want which one more" this question, nearly all of the people have chosen the latter, the reason is very simple--in cooking or washing clothes, the more need to be obedient "good helper."

So the question is: Now, why smart voice to choose wearable equipment or TV, speakers and other living room appliances as their own point of exerting force?

Regarding this question, Li Qiangjun thinks, the Hkust chooses from the television as the home appliance domain The starting point is conforms to the overall situation, "in the television market, the news flight accumulated the customer resources already rich enough, at the same time, we also with Tcl, Hisense, Changhong and so on domestic 6 major mainstream television manufacturers to establish the So smart voice into the living room is the decision. At the same time, Li Qiang also pointed out that the current voice of home appliances market is still in the early days, "the vast majority of consumers will not take the initiative to be in the cooking time through voice technology to help themselves," and more by the manufacturers to guide their use of new products. In addition, he also believes that in the individual family, human-computer interaction is the most people and television, while the normal family spend a relatively short time in the kitchen, so the choice from the living room is quite reasonable.

However, the idea of flying "product-led market" has also been questioned by other insiders.

For the kitchen interaction time is short this, cloud informed the CEO Huang in the interview with Sohu it said, although the living room human-machine voice interaction time is long, the kitchen is short, but the living room of the interactive demand degree far less than the kitchen, in other words, intelligent voice This "helper" to the user's help greater and more effective. Therefore, "time determinism" is not a certain science.

Conversely, in Huang's view, voice-interactive crowding can be worn with the main reason for the living room, in part by hardware decisions. The pattern of emerging areas has not been finalised, the use of the Internet, decision-making are more flexible, compared to the traditional electrical business, whether the product line or channel are "ship Big turn Slow", "need to coordinate the communication of the place too", thus, the voice of human-computer interaction in the white electricity market, although the high degree of consumer recognition, But the pace of development must be relatively slow.

In addition, there are some people in the industry raised the question: more advanced voice interaction means to have a larger voice data resources, currently can hold live online input database, can also hold live consumer's daily life needs?

Take Apple sir, Google Now for example, whether it is a hardware product or software system, two of companies in the world to accumulate a large number of personal data, but even so, Siri and Google's performance is not satisfactory, not to mention the voice of a single start other companies?

"The data is adequate", the industry is not too worried about this problem, as far as the domestic market is concerned, the service has been blanket more than 70% of the current voice app access products, and cloud knowledge is also said, as long as the voice data can meet the needs of wearable equipment, for other services, the day-to-day voice demand is not a problem.

Second, car intelligent voice say better than sing?

Apple CarPlay, Google Andriodauto, two big guys have sold, it seems that the car "can listen", "will say" not far away from us, but the truth?

In the near future, the North American market has released a report on auto consumption, which shows that new cars manufactured in 2014 have a high pick-up rate for smart voice. In the mainstream car brands, to Honda, Asia, Subaru led the Japanese and Korean brands, the United States, the local brand Tesla's carrying rate of up to 100%, Volkswagen, Ford's pick-up rate of 99%, 96%. However, although the new car carrying rate is quite high, but after the sale of complaints, "speech recognition" to 8.3% of the high complaints, North American consumers are selected as "the most unsatisfactory car function."

At the same time, the safety of car voice assistants has been widely questioned. A survey by the University of Utah in the United States points out that the smart voice system is more than 3 times times more likely to be distracted by drivers than other factors. Deborah Hersman, chairman of the National Security Council and former U.S. National Transportation Safety Board, also pointed out that "80%~90% traffic accidents are thought to be caused by human-computer voice interaction will make things worse". MIT researcher Bryan Reimer also admits there is not enough evidence to suggest that car voice is safe.

In this respect, some people in the industry pointed out that the cause of smart voice in the car market is so embarrassing reasons for two major points:

1. The function is complex, the actual demand for operation is far more than people expected.

Take Cadillac as an example, "Operation is too complex" is the main reason why its cue car entertainment system is criticized. "To do this, we need faster processors and more sensitive microphones," says Glen DeVos, vice president of Delphi Automotive electronics and security. However, most carmakers, including Hyundai, have now opted for a compromise--with smartphones as a bridge, voice-ordered handsets, and mobile apps to transfer cars.

But the fact is that the move is based on a relatively sophisticated smartphone, but the University of Utah in the follow-up survey found that the mobile phone smart voice itself can not "satisfactorily respond", and how to better coordinate the communication between people and cars?

No matter how Apple refutes the doubts about its own CarPlay and Siri, it is believed that consumers must be aware of it: either reply farfetched or just pull out safari and let you check it yourself. What's more, compared with mobile phone networking, the complexity of the car intranet is higher, data transmission is blocked, the network instability these will make "mobile phone + car" this collocation worse.

2. Voice recognition is low, the car to recognize the noise is the most important.

"There is no denying that the noise in the car is a big problem for us."

Dinesh Paliwal, president of Harman International Group, points out that although in the automotive Voice market Harman has a certain scale, but the Harman Acoustic department is still mainly software resistance, the main problem in hardware resistance is: How to make speech recognition is not affected by the normal passenger traffic, anti-"anti" noise- Create a noise shielding "sonic wall" between the front and rear. In other words, it is necessary to keep the passengers talking and not to interfere with the front-row driver's conversation with the car-"We think it should be welcomed in a market like China because there are 2 or even 2 people in the car."

In the domestic market, with the Hkust news fly led by many voice solutions enterprises have made some progress. In this regard, Li Qiangjun, head of the marketing department, said that in the environment larger than 100km/h, the current domestic market only flying voice recognition rate can be more than 90%, which for the noise source and environment are quite complex in the car is very important. In terms of specific products, Li Qiang also pointed out that to promote the real development of car voice, this requires more efforts from carmakers. "(in view of) the car scene is very sensitive to noise, so our cooperation with the car depot most of the hardware and software combined with the main", coupled with off-line voice technology for the network is not smooth in the vehicle environment is "the right remedy", Li Qiangjun that the car voice of the future development is quite bright.

It is understood that the current flying with Mercedes-Benz, BMW and many other mainstream car manufacturers to reach a cooperation, and in the Mercedes-Benz global speech recognition and noise evaluation in the top.

Whether in the living room or the kitchen, whether in the car outdoors or indoors, smart voice as a new technology must have its limitations, but the general trend, the 2015 spark or will open their own prairie.

(Responsible editor: Mengyishan)