Open Audio System Introduces Large AI Models, Bluetooth Audio SOC Moves To High -End

However, the challenge that the smart glasses equipment cannot be ignored is its open audio system. As the most basic function of smart glasses, audio is the technical problem that requires a better audio experience in addition to completing a variety of interactive experiences.

AITechnical blessing, promote the high -end Bluetooth audio SOC

OWSHeadphones, intelligent audio glasses and other products have put forward higher requirements for open audio systems.Facing the open audio system, where is the technical bottleneck and difficulty of the main chip?

Chen Wuqing, the person in charge of the sales leader of Wuqi Micro Bluetooth product line, said in an interview with an electronic enthusiast network that open audio systems have three major challenges. One is the problem of low -frequency leakage, and the other is the call effect and sports wind noise scene under the large echo structure and the sports wind noise scene.The quality of call quality, the third is the low power consumption of the battery requirements for the design of the small size of the light and comfortable structure.

In order to conform to the design of ergonomics, most OWS headphones use the U -shaped acoustic cavity structure, and the inside is equipped with a sound unit, such as the GoRee2 of Soundpeats, and the Verio 200 of BEYERDynamic Baya 200.The larger dynamic unit has better performance in low frequency response and sound pressure level, but it still cannot solve the challenges of low -frequency leakage, call effects, and wind noise.

Chen Wuqing mentioned that the large echo structure of open headphones has a problem with the MIC-SPK spacing, weak sound isolation, and low letter ratio. Relatively ear-in-ear, effective return loss (ERL) small nearly 20DB, call echo dual lecture effectDifference.

There is also openThe smart glasses of the type audio system are not smaller than the OWS headphones in the audio system.Generally speaking, OWS Ear opportunities are designed through different ergonomic structures to letThe front sound holes are close to the ear canal to improve the volume and bass experience of the open audio system. Although these different ergonomic design are not cured, they can improve the audio experience to a certain extent.The audio system of smart glasses is difficult to improve in the position of the mirror legs.

With the development of technology, AI has become one of the key technologies to improve the audio experience of open audio systems.It is understood that Wuqi's multi -mode AI capability has landed in multiple fields, and edge computing AI processing technology has been introduced into Bluetooth audio SOC.

In order to improve the call quality of the open audio system, based on the advanced circulating neural network and convolutional neural network, Wuli passes a large amount of data training model of 40G, plus the LMS+KLM dual filtering structure, and the linear echo suppression capacity is 30DB, nonLinear residual echo is introduced into AI treatment.In terms of wind -cut noise, through the blessing of multi -wheat technology and neural network algorithms, a large amount of wind noise data training AI models has achieved the effects of less wind noise residue, stable voice, less damage, and high recognition effects.

Compared with the past, it is only processed on the algorithm, and the addition of edge calculation of AI processing technology can achieve clearer call quality, which to a large extent improves the audio experience of open audio systems.

In terms of power consumption, the Bluetooth chips, sensors, microphones and other components of smart glasses are basically placed in the mirror legs. Disabled in space restrictions, generally can only use small batteries, so the battery life is poor. At this timeThe Bluetooth chip needs strong low power consumption performance.

Wuli's Bluetooth chip brings innovation and breakthroughs in improving battery life.The first to adopt the design architecture of RISC-V + DSP, and the SDK of multi-modular processing.Take OWS's application as an example. In the case of chip -level power consumption of only 3.xMA, it has added algorithms such as low -frequency compensation/virtual bass/dynamic EQ and multi -MIC AI ENC for the OWS form.The low power consumption level of 4.XMA is maintained. At a 45mAh battery capacity, listening to music can be used for 10 hours.

Under the technology iteration of chip manufacturers, the application of AI technology in the application of audio processing has made new progress. AI technology introduces Bluetooth audio SOC, becoming one of the symbols of Bluetooth audio SOC to high -end. This will also promote the innovation and innovation of the audio market andprogress.

Audio glasses will gradually start in the audio equipment market

In the first quarter of this year, the global personal intelligent audio equipment market ushered in a 6%year -on -year growth, with shipments exceeding 90 million units.In TWS headsets, wireless necksAmong the three major products of hanging and wireless headsets, the TWS headset category has grown weakly. The author believes that the next OWS headset and smart audio glasses will become the smart audio equipment marketThe segment category with obvious growth in the field.

From the perspective of the large category of smart glasses, the category of smart audio glasses will also start in the market first.Chen Wuqing believes that in the three major categories of AI/Smart Audio/AR smart glasses, smart audio glasses will take the lead, followed by AI glasses, and eventually evolve towards the direction of AR smart glasses containing audio/video/content.

Smart audio glasses will start first. In the current technical background, priority solve the problem of lightweight and mild voice processing. Secondly, the evolution of technology can be extended to the AI language.AI glasses with sound treatment capabilities, and continue to evolve into AR smart glasses containing audio and video and content, need to solve a lot of pain points.

Chen Wuqing said that for the chip, the main chip platform needs high integration and rich peripheral interface and AI processing capabilities. With the blessing of various voice/video processing algorithms, maintain low power under high computing powerRequire.

Judging from the long -term performance, letterBetter/Data pass through the base of the terminalAfter the basic processing, it will go to the cloud for further treatment. Therefore, there will be more and more equipment in the terminal processing, and the demand in cloud treatment will also increase.As smart glasses need to be processed, there are more and more scenes, and the cloud also needs stronger ability to deal with these processing tasks.But before that, the main chip platform of the terminal equipment needs to be more powerful to deal with more complex terminal processing tasks. Chip manufacturers also need to continuously iterate products to launch more competitive chips.