Qualcomm debuted a new family of chipsets positioned to power a wave of more responsive, intelligent and better-sounding smart speakers.
The new QSC400 system-on-chip (SoC) series is the company’s first platform dedicated entirely to smart speakers and combines a number of key capabilities including enhanced audio, AI, increased compute power and power optimisation.
In a statement, Rahul Patel, Qualcomm SVP and GM of connectivity said the new chips “raise the bar on both feature integration and power performance for smart audio,” adding the technology will “help manufacturers to more easily overcome significant technical challenges and build smarter speakers and assistants with more intuitive voice UI, connected user experiences and exceptional sound quality.”
Packing a punch
Audio improvements were a key focus for Qualcomm: the chips support 32 channels of integrated audio processing, Qualcomm’s new single-chip DDFA amplifier, and Dolby Atmos and DTS:X sound technology.
But the chip also packs a power punch when it comes to processing power and AI, integrating a quadcore CPU, Adreno GPU, and Hexagon digital signal processor with vector extensions alongside Qualcomm’s AI Engine.
At a launch event at Qualcomm headquarters in San Diego, US, director of product management Nishant Mittal explained the increased compute power, AI capabilities and support for multi-keyword recognition will enable faster and more natural interactions with voice interfaces. Enhanced far field pickup and noise cancellation mean smart speakers will be able to keep up performance even in noisy environments.
Mittal added local automated speech recognition will also allow users to take their smart speaker experience offline, which he said could come in handy in instances when power outages prevent cloud connectivity.
The company also promises to deliver massive power savings. Mittal said the chipset offers up to a 25-times improvement in standby time, and will allow smart speakers to operate unplugged from a power source for up to 14 days with no loss of functionality.
Gary Brotman, Qualcomm head of artificial intelligence strategy and product planning, highlighted AI’s ability to enhance user experiences and predicted the technology will become a critical component of all future audio, visual and gaming systems.
“Nothing that we do today in audio, and actually many domains, will move forward without artificial intelligence in some way shape or form backing it.”
Specifically, Brotman said, on-device AI capabilities will help meet demand for low latency responsiveness: “If you have to deal with a roundtrip to the cloud to draw context from a request that you make to your smart speaker, there’s frustration that comes into play. That’s something we’re trying to address with QCS400.”Subscribe to our daily newsletter Back