Coquettish Rebirth

"Background noise reduction?" Ni Guangnan and Luo Feng were startled, they hadn't thought that Jia Hongjian would ask them to study this thing!At that time, Mr. Ni was startled, and hurriedly said, "Mr. Jia, there is no way to quickly produce results after this research!" After finishing speaking, seeing Jia Hongjian's expression of disbelief, he hurriedly explained—"Mr. Jia, really! 。＋＋＋There are companies in the United States that have actually been researching since the 90s. For example, the bone conduction microphone technology researched by the American aliph company relies on a device like an earphone that is attached to the skull behind the ear while acting as a microphone. Accepting the sound from the skull conduction user to reduce noise, like this company, just applied for a patent two years ago, it seems that it is not easy for us to quickly bypass the patent and make the corresponding effect! And even if we don’t use this technology, but if software is used to identify and reduce noise, computer auditory aspects similar to computer vision must also be designed, and noise screening must also be carried out... This is not something that can be done in three or two months!"

That's right, Ni Guangnan knew how difficult it was.Human beings can tell what is noise and what is human speech at once, but machines can't!For the machine, whether it is human voice or noise, it is just the sound they receive and record. Others’ machines are just grabbing eyebrows and beards, and they don’t know what eyebrows and beards are. How do they distinguish?Similarly, computer vision is also difficult. People can tell what it is when they look at something, but if the computer analyzes the picture passed through the camera, the computer can only analyze that the picture is full of small color blocks.How does a computer know what is what?

In this case, how to make the mobile phone perform noise reduction processing?Otherwise, it is necessary to avoid recording noise as much as possible. That is the principle of the bone conduction microphone adopted by Aliph Company, because sound needs to be transmitted through a medium. If it is transmitted from the air, it will cause the microphone to record human voice and noise.But what if the human voice through the bone conduction is received from the surface of the skin next to the bone?That noise will not come in through human bone conduction!Otherwise, it is necessary to find a way from the aspect of computer digital optimization, that is to say, either teach the computer how to distinguish, or find out the difference between noise and human voice, so as to distinguish them!But seriously.If only a computer could be taught how to recognize it.Isn't that artificial intelligence?And how to find the difference between noise and human voice is also a very difficult task in itself, and it may not take several years to produce results!If it is easy to do, it is reasonable that the company in the United States should have the result very early!

Ni Guangnan really felt that this aspect was extremely difficult.But who knew that after he expressed the difficulty of this matter.But he only heard Jia Hongjian's nonchalant question. "Then why don't we use multiple microphones to distinguish noise and human voice by receiving the difference in sound through different microphones?" "Multiple microphones?" At that time, Ni Guangnan was stunned. ?In the end, who knew that Jia Hongjian nodded as a matter of course, "Yes. Multiple microphones, such as a microphone on the top of the phone, and a microphone on the bottom of the traditional position, so that when the phone is normally held to make a call, the upper and lower microphones receive The size of the human voice is different, but the received noise level is basically the same, so with such a simple multi-microphone arrangement, it can be clearly distinguished whether it is the human voice or the noise, right?"

Whoa!When Ni Guangnan heard what Jia Hongjian said, he was stunned!Why are you stunned?He never thought of this method at all!The methods he thought of before were all the methods being researched by companies in the United States.To be honest, even if those methods are useful, the cost is definitely not low!In particular, power consumption is also a big problem—for example, the so-called software identification method, which is currently being used in the United States. It is said that after the dial-up starts, the system chip in the mobile phone will directly start collecting various information through the microphone. an outside sound.At such a time, the person who has just connected has not yet called, so what is actually collected is noise!Then the voiceprint of these voices can be identified as noise.Then recognize the voice of the person who appeared after the conversation as a human voice, so that it can play a role in noise reduction!

But such a problem is that noise reduction is noise reduction, and it is also very smart.But the problem is that the battery power of the mobile phone is not large. If you meet someone who calls a lot all day long, the extra power will actually consume a lot!In addition, because today's digital storage capacity is too small, the recognized noise cannot be maintained for a long time, and it is not very useful if it is kept for a long time!This eventually led to the fact that every call actually requires an additional re-analysis of what is noise and what is human voice. Isn't this a waste of electricity?But what about the method Jia Hongjian just mentioned?That's amazing!Although such a method has not been proved to be feasible, but if you think about it, it seems to be a method that can "steelmaking by local methods"!

Indeed, if one microphone is placed on the top of the mobile phone and the other is placed below, then the human voices collected by the two microphones must be different in volume!But the sound of cars or other people talking in the background, the difference in the volume of the collected sound will not be too big, right?In this way, can't the chip in the mobile phone easily distinguish who is the noise and who is the human voice?At that time, we only need to simply keep the one with the big difference in the sound collected by the two microphones, and remove the one with the small sound difference. Isn't that all right?wipe!This idea is so genius, isn't it!The American company spent almost ten years and poured in countless money, but the method they came up with was not as feasible as the one that Mr. Jia casually said!This, this, this... Mr. Jia is too awesome, isn't he?

Ni Guangnan was really shocked at this moment!To be honest, in his previous view, this Jia Hongjian was actually a young man who was interested in technology, but his foundation was not very good, but Jia Hongjian was very talented in business.So Jia Hongjian can give you directional guidance in some fields, but Jia Hongjian can't give you guidance on the specific technical level.Just like driving a car, Jia Hongjian can point everyone in the right direction and tell everyone which direction to go, but he really doesn't know which road to take to get to that direction!But today, Jia Hongjian, who has always been able to point out the direction but not which way to go, suddenly pointed out a way for everyone, and such a way has not yet been developed by foreigners who have spent so much time and money. , how could this not shock him?And the most important thing is that his old Ni is also a double academician of the Academy of Engineering and Academy of Sciences!As a result, he now feels that he is not as deep as Jia Hongjian in this field, and he even feels that he can't come up with such a good idea as Jia Hongjian!How could this not shock him?

What he didn't know was that Jia Hongjian didn't understand technology, but he understood the future!Mobile phones like the future Apple iPhone will actually use this technology!Even later generations of iphone series mobile phones are no longer dual-microphone collection, or even three-microphone collection!In this way, the noise can be eliminated more accurately and the human voice can be preserved!It can be said that people who are more concerned about mobile phone technology in future generations will basically understand it!And the things that people in the future can understand casually, are such unprecedented ideas in such an era!

After hearing this, Luo Feng felt that the method suggested by Jia Hongjian seemed to be really feasible. He turned his head and looked at Ni Guangnan who was in shock, "Master, why don't we try it?" "Well, let's try it!" Ni Guangnan suddenly nodded.At this moment, he felt that he had no time to express his shock to Jia Hongjian. The shock of Jia Hongjian's idea was beyond the need to express shock and admiration. Try the way you said!Soon, the master and apprentice took Jia Hongjian to the laboratory of Huaxia Communications. First, they found a mobile phone, then disassembled the mobile phone, and disconnected the cable of the microphone at the bottom of the usual position. Connect the external cable and get it to a sound recording and analysis device.Then on the other side, a small microphone was manually built into the handset of the mobile phone, and it was also connected to a sound recording and analysis device.

Then they found a bunch of people and various things, and prepared to simulate the noise—this noise includes the background sound when people around are talking loudly, as well as what kind of car is passing by, what else is being heard Play music on a tape recorder or something.Then they talked casually with such a mobile phone veneer, and then looked at the recorded data... Hey!Sure enough, the background noise and the like, the difference in the sound collected by the two microphones is actually not that big, but the difference in the sound of the human voice is very big!If you compare it with the human voice, which is usually said to be not too loud on the phone, the difference between the sound recorded by the microphone below the phone and the sound recorded by the microphone above the phone is as much as 6 decibels !

The difference of 6 decibels is quite big. Generally speaking, the sound difference that can be distinguished by the human ear is about 0.5 decibels. When the difference reaches 3 decibels, it is a very obvious difference. Basically, anyone who comes to listen Those who don't have ear problems can basically hear the difference!If it is 6 decibels, the gap is even bigger!Relatively speaking, even if there is a difference in the background noise collected by the upper and lower microphones, it is not very big, far less than the difference in human voice!what does that mean?This means that the idea proposed by Jia Hongjian is basically combined with a chip and a simple small software in the chip, which can directly achieve the effect of noise reduction!This is too amazing, right?Boss Jia has an idea...is it directly feasible?And still so simple? (To be continued..)