Cadence Tensilica IP Powers Analog Devices’ Next-Generation DSP Architecture

Cadence (Nasdaq: CDNS) today announced that its Tensilica IP Group and Analog Devices, Inc. (ADI) worked together to develop the latest generation of SHARC® automotive audio processors, further strengthening ADI’s industry-leading position in premium in-vehicle audio solutions. The joint effort brings Cadence’s highly optimized Tensilica DSP architecture into the next-generation ADSP product family from ADI, enabling richer audio experiences, enhanced voice interactions, and greater system intelligence for future automotive platforms.

For decades, ADI has been a leader in automotive audio processing, with its ADSP family powering many of the world’s most advanced infotainment and premium audio systems. These processors enable a broad range of applications, including:

• Automated Driving alerts and Acoustic Vehicle Alerting Systems (AVAS)
• Road Noise and Engine Order Cancellation (RNC and EOC)
• Personal Sound Zones (PSZ) and immersive 3D audio rendering
• Voice processing, AI assistance, and hands-free communication

“As the complexity of automotive audio, acoustic and voice applications continues to grow, it became clear that we needed a DSP architecture capable of delivering significantly higher performance and on-chip AI capability while maintaining the deterministic real-time behavior our customers depend on,” said Andrew Lanfear, head of Automotive Audio & Networking, ADI. “Given their leadership in DSP IP, working with Cadence was the natural choice for ADI. The ability to create custom instructions using Tensilica Instruction Extension (TIE) technology allows us to accelerate proprietary audio and voice algorithms while providing dedicated DSP resources for real-time voice offload and next-generation automotive workloads.”

“ADI’s selection of Cadence’s Tensilica DSP technology for the next generation of ADSP processors underscores its focus on high-performance automotive audio and AI,” said Amol Borkar, group director, product management, Tensilica DSPs, Cadence. “The advanced architecture in the SHARC-FX core is optimized for on-chip AI/ML and signal-processing workloads, enabling applications such as music genre identification, music source separation (MSS), and real-time vocal removal to run efficiently on the DSP, while the host CPU remains dedicated to vehicle-domain and operating system functions.”

“DSP Concepts has been a long-time trusted partner of both Cadence Tensilica DSPs and ADI, enabling rich audio applications on both platforms,” said Chin Beckmann, CEO of DSP Concepts. “Through the integration of their technologies in our Audio Weaver (AWE) framework, we were able to witness and validate the winning, best-in-class results of this amazing collaboration between these industry pioneers.”

Building on the success of the ADSP-SC59x/2159x family, ADI is introducing the new ADSP-SC84x/2184x platform, featuring a next-generation architecture and a transition from 28nm to 16nm process technology, delivering significantly improved energy efficiency, higher operating frequencies, and increased integration for modern software-defined vehicles (SDV). Other benefits include an LPDDR4 memory interface, increased on-chip L2 memory, enhanced cybersecurity via a dedicated hardware security module (HSM), and two optional connectivity cores offering enhanced communication and mainline Linux support.

One of the ADSP-SC84x family’s key innovations is the introduction of the new SHARC-FX processor core, designed and built through a joint engagement with Cadence. The SHARC-FX features a custom DSP architecture, built on the same technology platform that delivers best-in-class Tensilica DSPs for audio, vision, and radar segments. As a result, the SHARC-FX delivers a substantial performance uplift, supporting a broader range of data types, enhanced VLIW SIMD architecture, efficient cache operations, and an expanded instruction set optimized for modern signal processing and AI workloads. Together, these improvements, coupled with the higher operating frequency of the SHARC-FX processor core, provide 5X higher performance compared to the previous generation SHARC+ processor used in the ADSP-SC59x family. This paves the way for an array of next-generation applications in the vehicle, including seat-specific in-cabin audio experiences, AI-enhanced noise cancellation, intelligent voice assistants, and generative AI audio.

As automotive audio and voice-enabled systems continue to evolve, many workloads are now transitioning from traditional signal-processing algorithms to AI-driven approaches that deliver higher accuracy, greater adaptability, and more personalized user experiences. The SHARC-FX core’s expanded instruction set enables it to efficiently implement these emerging AI models alongside conventional DSP workloads, providing the computational performance required for real-time neural-network-based audio inference at the edge. By supporting optimized NN kernel routines for popular AI/ML frameworks such as Google’s LiteRT™, the latest AI models can be easily offloaded to the SHARC-FX core for a high performance and energy-efficient implementation compared to running them on the host CPU.

Furthermore, ADI provides a rich toolset and ecosystem consisting of Cadence’s compiler and tools as well as ADI’s EZ-KIT® evaluation system, CrossCore® Embedded Studio, SigmaStudio+TM, and other industry-leading frameworks, including DSP Concepts’ Audio Weaver.

The Cadence Tensilica IP Group is a leading provider of DSP IP, with widespread adoption across its full range of DSP product families. It also provides automotive-ready DSP IP, certified up to ISO 26262 ASIL-B, for ADAS and other safety-critical applications.