CONNECT WITH US
Sign out

Smart speakers enrich user experiences with AI and hardware optimizations

DIGITIMES staff
0

Sales of smart speakers integrating artificial intelligence (AI) analyses and targeting wide-ranging smart home application scenarios have performed outstandingly in 2017-2018. According to Canalys' smart speaker market research report released in October, 2017, the global smart speaker market is poised to grow to 56.3 million units in 2018. Such explosive growth will ignite fierce competition.

After Amazon launched its Echo smart speakers with smart voice services, enabling users to access its value-added e-commerce services using its voice assistant, Apple followed suit and unveiled its own smart speakers with a home assistant powered by Siri, Apple's voice-controlled personal assistant originally available on iPhones. Leveraging its smart speakers, Apple's smart service ecosystem spreads into the smart home arena.

Rapid developments of smart speakers

With world-class e-commerce giants and smart product developers entering the market space, vendors - large and small - were scrambling to unveil their own AI-enabled smart speakers in 2017. As a result, the market is seeing a wide variety of smart speaker implementations and designs each with different AI value-added services corresponding to different applications and assistance services. CES 2018 even had a slew of smart speakers competing for the spotlight, following the smartphone frenzy. There will be large differences in terms of core technologies, application scenarios, ecosystem integration and product usability as a great number of IT companies and network service operators join the game.

Global smart speaker shipments reached 30 million units in 2017. As big players pour in more marketing resources and additional mid- and large-size vendors bring new products to the competition, 2018 will be a defining year for smart speakers. The outcome will hinge on who has more "smart" integration advantages, how well the smart speaker is suited to diverse application scenarios, how many resources are available for the vendor to provide immense value-added services, and how the vendor can stand out from the competition, leveraging hardware, smart system and ecosystem technologies and resources. This is why the smart speaker contention is attracting market attention.

Maturing smart speaker industry chain

A review of the smart speaker market with respect to industry chain development showed the products on the market still had many problems in 2017, when the storm was still brewing. For example, their AI smart service modules still needed to be validated and enhanced. Their fundamental SoC solutions had room for improvement. Their market remained to be further developed. However, after 2017, smart speaker SoCs, voice/face recognition and other AI modules gradually reached marketable functionality with every element in the industry chain including mechanical designs, user experiences, major chips and overall technologies also becoming mature. As such, the key to product success will be determined by what application scenario the smart speaker is targeted for and how well it addresses users' pain points.

In fact, a vendor attempting to expand into the smart speaker market in 2017 would encounter some major non-technical challenges. First of all, smart speaker system integration had not yet reached completion. It still needed some time for chip yield rates and service consolidation to keep improving. Furthermore, as there were a large number of human machine interface (HMI) designs for voice assistant, biometrics authentication and decision making, it would take users a while to get used to these new generation products. As a result, market awareness of smart speakers was still limited and market development costs were running high. However, big players including large-scale network/cloud service operators and smart hardware developers have been beefing up efforts toward the smart speaker market in 2017-2018, having launched more than 10 AI-enabled smart speakers.

Independently-developed products with limited functionality have difficulty surviving

In 2017, independently-developed smart speakers were the mainstream on the market. They were mostly developed in-house, combining diverse network services such as streaming music, video and data services. With large-scale network operators and hardware developers joining the game, the smart speaker market puts up a high entry barrier with smart applications including noise reduction with microphone array, AI-based voice recognition, AI-based application scenario analysis and automation applications. World-class hardware developers and network service operators like Apple, Google and Amazon have leveraged their corporate resources and advantages to expand into the smart speaker market, wherein the competition will become increasingly intense and the growth will be fueled by these large vendors offering subsidies, promotions or even more advanced smart applications in 2018.

Thanks to large vendors joining the game, smart speakers entering the market in 2017-2018 come with mature functionality and applications. For example, major network service operators Baidu, JD.com and Alibaba as well as home appliance or consumer electronics providers Haier, Lenovo and Xiaomi are entering the market. International players including Sony and sound equipment vendor Sonos have also launched their own smart speakers, and so have Rokid and Mobvoi endeavoring on smart hardware. Even online content provider Ximalaya has also brought a smart speaker under its brand name on the market.

Smart designs in smart speakers are growingly challenging

As speakers become smart, they can be used in wide-ranging scenarios. Smart speakers in their early days worked in conjunction with smartphones through Bluetooth or other communication protocols. Users controlled smart speakers using smartphone apps for media playback or more advanced smart applications. These early-stage products were much similar to Bluetooth speakers in terms of user experiences. Budget restraints also limited how smart they could be so the overall speaker systems provided little usability. New generation smart speakers have transitioned toward independent smart operation with built-in network connectivity and expanded smart functions such as AI-based voice recognition, smart home appliance control, online streaming video/music and human-like conversation. They are a significant upgrade from the early-day products at system, intelligence and hardware levels.

Smart speakers now basically engage in integrations in three aspects – hardware optimization and integration, e.g. microphone arrays with sound pickup and noise reduction instrumental to voice control quality, resonant cavity designs for compact speakers to improve sound quality and embedded system/hardware with smart computation cores; smart integration of software or network services, e.g. voice assistants Apple Siri, Google Assistant and Amazon Alexa as well as Baidu AI combining smart speakers with cloud services; value-added services for different use scenarios, e.g. voice-activated e-commerce services, voice-controlled music selection and playback and voice-based Internet access/search. Among the three, enriched services for different use scenarios play a particularly important role.

HMI designs in smart speakers are critical to user experiences

HMI developments play an influential role on smart speakers' user experiences, which can decide whether they will be able to broaden their application markets. As smart speakers generally feature compact footprint with small displays or even no display and very few hardware control buttons, advanced voice-based HMI becomes a critical control mechanism for smart speakers, such as Xiaomi Xiao AI, Apple Siri and Amazon Alexa. In addition to the reliability of the hardware itself, user experiences with hardware operations are subject to smooth human machine interaction through voice control, which has the final say on smart speaker usability.

Some vendors with budget considerations designed their smart speakers without displays but some vendors chose to put displays on their products. Users of screenless smart speakers may not be able to understand system or voice responses from the smart modules. Back and forth voice communication between the user and the device results in poor efficiency and disrupts the flow of operation. Display screens may be the easiest solution to enhancing user experiences. To take it a step further, by equipping smart speakers with larger screens, vendors can make more applications available on their smart speakers. For example, Amazon Echo Show and Baidu Duer smart speakers with large screens are added with differentiating features such as remote video communication and voice-activated searches for cooking or other video clips on the Internet.

The interior of Apple HomePod

The interior of Apple HomePod
Source: Company

Baidu Duer smart speaker

Baidu Duer smart speaker
Source: Company

Amazon Echo Show

Amazon Echo Show
Source: Company