Cadence introduced the Tensilica HiFi iQ DSP IP, the sixth-generation member of its HiFi DSP family, powered by a new architecture that ramps up processing performance to meet the rapidly growing demands of immersive audio and voice AI applications in devices ranging from vehicles to smart home speakers, wearables, and hearables.
The HiFi iQ DSP offers 2x greater compute performance, 8x higher AI performance and more than 25% energy savings for most workloads than its fifth-generation predecessor, all major improvements that are entirely necessary when one considers the development of automotive entertainment and infotainment systems that could have many more speakers, or the rapid emergence of voice-as-keyboard and voice navigation for AI applications.
Regarding future in-vehicle audio signal and AI processing needs, Amol Borkar, group director of product management and marketing, Tensilica DSPs, at Cadence, told Fierce Sensors, “Right now, many of these mid level cars have 12 to 16 speakers in the car for [audio[ playback. You got 12 channel Dolby Atmos sampling at 48 kilohertz. You got some level of post-processing and speaker mixing that is happening. But, moving forward, you're going to start seeing 30 or 40 speakers in the car, immersive playback, 96 kilohertz, much more realistic channel rendering. Also, with some AI sprinkles in there, as you can imagine… This workload increase is probably going to put the pressure on the processor to have about 2x or higher compute requirement.”
But enhanced processing in automotive audio subsystems goes beyond the number of speakers, as it also enables in-cabin zoned sound bubbles and cancellation of road and engine noise, as well as powering audio warnings and voice processing for in-vehicle hands-free capabilities. While AI increasingly is being used to enhance audio playback, it also is being used to enhance noise cancellation and voice processing workloads whose compute requirements will at least double–and likely increase many times more–in the years to come.
Higher-performance DSPs also will be needed in other cases. “Voice being used as a keyboard seems to be also quite prevalent now,” Borkar said. “A lot of voice-based interaction and more natural language processing when the voice is used as the input for navigation on a lot of applications is picking up a lot of interest. And then, just in general, voice AI-driven applications, classical signal processing solutions like noise cancelation, beam forming, echo cancelation, speaker separation–all of these are starting to all get transformed or become a mix of signal processing and AI based models." This includes integration of small and large language models aimed at using voice inputs on agentic AI applications, and the growing need to process these on-device in low-latency, power-constrained scenarios.
Anshel Sag, VP and principal analyst, Moor Insights and Strategy, backed up that notion in a reference quote supplied by Cadence, stating, “With the increased growth of AI across virtually every facet of the industry, audio and language have become critical interfaces for users. The AI performance and energy savings with Cadence’s new flagship Tensilica HiFi iQ DSP position it to enable future AI-enhanced applications and offload processing from other IPs.”
The Tensilica HiFi iQ DSP will be available to lead customers and partners in the first quarter of 2026, with general availability expected in the following quarter, according to Cadence.