A wider range of applications for voice-interactive technology, what's the big thing? __ Voice Interaction

The last time you had such a strong interest in voice technology or a few years ago, the focus was on mobile phone voice assistants, such as Siri, Google now, and so on. At first it was interesting to see the form of the voice conversation, but after a while trying to find out, in addition to let it tell a joke and occasionally flirt with amuse the child, no longer think of what to wake it. Finally, most people are forgetting this virtual voice helper that stays in the corner.

This time, a new product with voice interaction features has attracted people's attention again, and after two years of actual experience, still interest and praise, this product is Amazon Echo. If you want to generalize it in a sentence, it can be said that it is a "can understand what you say to it and to a certain extent to give effective feedback of the intelligent assistant", if a little hype or better understanding, you can say that it is the real version of Jarvis ("Iron Man" JARVIS).

At present, there is not a relatively successful similar products in China, so it is not clear how domestic users think of voice-interactive technology and will not like such products. Some of the analysis of the article is very serious explanation, because of cultural differences between East and West, this kind of products in domestic more difficult to popular, because we are more reserved, do not like this directly with a hardware product for voice dialogue. This is clearly taken for granted, and Western Zuckerberg (Facebook CEO Mark Zuckerberg) says he does not like to use voice to give his own AI assistant, Mr. Jarvis, under certain circumstances, because of the lack of secrecy and the disturbing of others.

It can be seen in the "dialogue with the robot" this matter regardless of eastern and western people are the same psychological "barrier", the key is to correctly recognize that "voice is not omnipotent", to find the application of the voice of the scene and the use of experience to do the best, the application of voice technology popularization will be very helpful.
So how does Echo do it?
According to Bloomberg reports, the Echo team initially did not consider the main music function in the design, but only in the beta, we found that the common function is to use voice search songs, so it strengthens the characteristics of sound quality, the volume of the product bigger. After that, an engineer tries to connect the voice function to the control of the intelligent hardware, which is more popular with the users, and becomes a stimulating factor of the sales and praise.

What do we do with Echo? You can find very detailed comments and usage feelings on Amazon's official web site, such as-

"Just call her name and say what we want to do, Echo can respond with a wonderful voice and help you do it, whether you're sitting in the living room or walking around the house, she's always ready to listen." 」
"I'll ask her to help me buy things on Amazon, check the weather and my schedule, and adjust the temperature in the room."
"alexa, who sang this song? Click here, next, pause, play Awolnation's Sail, add this song to my playlist, and play a play list in my Spotify.
"Especially when you're cooking in the kitchen, you can just let Echo get you the order you want, and never worry about forgetting what you're going to buy."
"I don't have to worry about forgetting something important, she'll remind me to take my medicine and remind me to see a doctor later."
"Wake up in the morning with Alexa said Good morning, she will say, morning, today is Danny's birthday, remember to prepare a gift ~"
"Every day I would read my selected newsletter to me."
"Help me read all the books in my Kindle, not the sound of the machine, it sounds very comfortable."

Echo can do more and more things, including music search and control, control intelligent hardware, set alarm clock, check weather, search encyclopedia, set reminders, read news, schedule reminders, voice shopping One-click order and Support Inquiry logistics status. And so on a variety of rich and practical, the use of a very good experience of the function.
There are several reasons why you should be loved:
1. Synthetic sound is relatively close to the human voice, is not an abrupt machine sound, it sounds very comfortable;
2. Technology is good enough, the system can respond quickly, needless to say a word after silly wait;
3. The data is full and updated instantly, and can always provide an unexpected service.

In fact, two years ago, there are similar attempts at home, such as "small smart speakers", the core function is also music and intelligent home control, but not very successful, its co-founder Li Xuanfong has shared the thought of why the Chinese market could not be born Amazon Echo.
To sum up, there are probably these factors:

1, in the domestic, the price of audio products 80% within 200, more than 500 pieces of product monthly sales more difficult than 10,000 units.
2, voice interaction to achieve good use of experience, rapid response is the most fundamental demand, other such as the texture of synthetic sound, more complete data and algorithms, to achieve these costs are low;
3, the domestic music copyright concentrated in the hands of several giants, for startups, it is difficult to do what users want to hear what, and then affect the user experience;
4, the domestic play intelligent hardware is not much, manufacturers and brands are quite a lot, unless you do enough good, otherwise no one would like to cooperate with you
5, need a long time can not see the harvest of the technical accumulation period (Echo spent more than three years)

Now, two years on, the above mentioned conditions do not seem to have changed much. Although people are gradually accepting the form of voice input, more people are beginning to try to use the voice input method, but in the domestic "voice interaction" and the form of dialogue with the machine seems to be relatively distant.
For example, at present, more suitable for voice interaction two scenarios-in-vehicle navigation and intelligent home field, there are already many support voice interaction products, but the real use of few people.
The design of the product is ideal--sit in the car and fasten the seat belt, by the way, "navigate to xxx", and then drove off, but the fact is, people would rather have their phones typed, and in the hype of smart homes, there's a lot of smart scenes that don't feel good for most people, "these smart scenes look amazing, But I don't know what it means to me "or" these scenes seem far from my life. " For example, "say a word and you can turn on the lights." This will be very attractive to users. And, for most people, it seems like life is not too busy to need a "smart assistant" to help with scheduling.

Therefore, voice interaction technology in the promotion may also be the first to find the most suitable user groups and truly valuable usage scenarios. For example, the propaganda of voice manipulation seems to weaken a little more, on the one hand because the current domestic smart home can be called smart hardware is not very popular at home, on the other hand, voice control in real life may not be the real use of the frequency is not so high (Facebook CEO Zuckerberg has also shared this experience. For example, "adding a sensor to the door and opening the door at home will automatically open" This function is better than "yelling at the air to return home".

It should be a very complicated and huge workload in the provision of voice question and answer. According to Amazon founder Jeff Bezos at the Recode conference in 2016, Alexa and Echo's research and development team has surpassed 1000 people. It can be seen that to achieve a larger range of people satisfied with the effect of different segments of the population in various areas, it will take a long time to accumulate.

Music, radio, audio books, subscribe to the field of news, is the voice interactive products most suitable for output content, but also seems to be able to cultivate user habits. Morning to wash up to go to work and at night before the two periods of time, is the most intelligent speakers play space time. For the form of the Voice search song, a few years ago, there is a product called Jing, through the "Natural language" search song, is a very popular niche products, you can directly use language description to search for songs, such as "The sun is very good today", "Rain Outside", "I am reading" Want to listen to Western classical light music ", and so on, personally feel that this is a reference to a model."

In addition to scenarios and user groups, the most critical is the voice technology itself. For example, to achieve rapid response, far-field precision recognition, which is a good user experience, but should also be a higher threshold of things, the need for adequate technology accumulation. Echo realizes a quick response within a second, the use of hot words to wake up the basic no longer wait for the system to respond directly to the dialogue, which also to some extent to avoid the voice and machine interaction embarrassment.

On the other hand, the smart speakers and the corresponding field of products, is not the technology to win, the test is the team's grasp of the scene and operational capabilities. After all, it's not a single function that can be planned, it is not a company can completely cover, and more like a grand "ecological", relying on the comprehensive development of technology applications, mutual support, such as the Internet of things, the popularity of intelligent home applications, various types of content sources and scene planning cooperation.

Look forward to good products. https://zhuanlan.zhihu.com/p/25279998
https://zhuanlan.zhihu.com/p/25279998