Unveiled by NXP Semiconductors at the AES Automotive Audio conference in Detroit, the new FRDM-AUD10 development platform enables rapid audio prototyping and the evaluation of real-world audio workflows, including AI- and ML-based processing. Aimed at audio product developers who need to test audio algorithms with real signals before committing to hardware, the FRDM-AUD10 board is centered on NXP’s latest AUD10 processor.
Unveiled by NXP Semiconductors at the 2026 AES International Automotive Audio conference in Detroit, the new FRDM-AUD10 development platform enables rapid audio prototyping and the evaluation of real-world audio workflows, including AI- and ML-based processing. Aimed squarely at audio product developers who need to test audio algorithms with real signals before committing to hardware, the FRDM-AUD10 board is centered on NXP’s AUD10 processor and paired with Audio Weaver, HiFi 5 SDK, and Xtensa Xplorer license to shorten the path to working code.
Nik Dimitrakopoulos, Product Marketing Infotainment from NXP Semiconductors with the brand new FRDM-AUD10 platform introduced at the 2026 AES International Automotive Audio conference in Detroit, MI.
The FRDM-AUD10 development board from NXP is a compact, low-cost platform that combines real audio I/O, a powerful HiFi 5-based processing architecture, an Arm Cortex-M7 control core, and broad digital connectivity so developers can move quickly from algorithm ideas to practical system validation.
NXP designed the FRDM-AUD10 for testing real-world workflows, including AI- and ML-based processing, with built-in MEMS microphones, ADC inputs, and stereo DAC outputs, allowing developers to assess audio quality, latency, and end-to-end performance using actual signals rather than simulated ones. This makes the board especially useful for audio teams who want to catch and track how their processing chains are working right from the start of the design process.
The board supports the standard interfaces, including TDM and I2S, while USB-C power and Ethernet simplify lab setup. The included Audio Weaver board support package (BSP), HiFi 5 SDK, and Xtensa Xplorer license give the platform a software stack that is immediately useful for development work. The Xtensa Xplorer IDE combines software development, processor optimization, and SoC architecture tools designed for Tensilica Xtensa processors.
At the center of the board is NXP’s AUD10 processor, which NXP describes as a scalable audio chip solution that combines HiFi 5 audio DSPs, an Arm Cortex-M7 MCU, and both analog and digital audio interfaces for professional and semi-professional audio, robotics, industrial audio, and consumer electronics applications. The chip is still listed as preproduction, but its feature set is already clearly aimed at developers who need a flexible, software-defined audio engine rather than a fixed-function signal path.
The new AUD10 includes two floating-point HiFi 5 DSPs with neural network accelerators, so it is designed not just for conventional DSP tasks but also for embedded AI and ML workloads. NXP also emphasizes the presence of 12 configurable audio ADCs, 20 integrated asynchronous sample-rate converters, and a fully programmable digital audio I/O port, which tells you the processor is built to handle complicated multi-channel workflows and mixed-format designs. For system architects, that mix of resources is especially useful in audio control, signal routing, and algorithm partitioning between the MCU and DSP domains.
The AUD10 supports PCIe 3.1 with 2.5 Gb/s and 5.0 Gb/s data rates, plus a second PCIe interface or Gigabit Ethernet through SGMII, giving developers a path toward networked and low-latency system designs. NXP also points to onboard SRAM, hardware-based biquad accelerators, and a 9x9mm FC-CSP package, which suggests the part is intended for dense, production-oriented designs where integration matters as much as raw compute.
For audio developers, the FRDM-AUD10 board turns AUD10’s capabilities into something practical to evaluate. The FRDM-AUD10 brings in four MEMS microphones, multiple analog inputs, and dual stereo DAC outputs, making it possible to prototype capture, processing, and playback in one environment. That is particularly helpful when testing beamforming, noise suppression, voice interfaces, or multi-channel playback chains where timing and conversion details can make or break the final result.
Featuring built-in MEMS microphones, sound-source ADC inputs, dual DAC outputs for headphones or speakers, USB-C power, and Ethernet, the NXP FRDM-AUD10 development board allows assessment of audio quality, latency, and performance using actual signals. Development is accelerated with Audio Weaver, support for HiFi 5 DSP SDK, and included Xtensa Xplorer license.
New SAF9800 SoC
In a related announcement, NXP is now shipping the new SAF9800 automotive audio and radio processor previously announced in May 2026. Designed to simplify infotainment architectures, the new device combines analog radio (AM and FM) and AI/ML-enabled audio processing into a highly flexible single-chip solution that meets the needs of a vast range of next-generation vehicles. The solution enables automakers to reduce system complexity and scale features across global vehicle platforms. This empowers OEMs to consolidate functions that have traditionally required multiple discrete radio and audio components into a single, scalable device.
The SAF9800 integrates a HiFi 5 audio DSP with neural-network processing capabilities together with NXP’s hardware-based biquad accelerators, delivering more than 10x audio processing performance compared to previous generations. This enables advanced audio use cases such as real-time noise reduction, voice and sound classification, and more adaptive in-cabin listening experiences — using data-driven algorithms rather than fixed signal processing.
Furthermore, the new SAF9800 processor leverages AI-driven acoustic source separation technology that transforms in-cabin microphone data by isolating critical sounds like vocal commands, mechanical failures, or emergency sirens from overwhelming background noise. Designed for software-defined architectures, the solution allows automakers to configure, upgrade, and differentiate radio and audio capabilities without changing hardware, and add OTA feature upgrades.
The SAF9800 uses the same DSP architecture of the new AUD10 processor specifically for automotive, showing how NXP is building a family of audio platforms that put DSP, control, and connectivity into a tighter, more software-defined system model. The solution also supports the same software and tools, including DSP Concepts Audio Weaver (AWE Lite included), open-core HiFi 5 SDK (Xtensa Xplorer included), and Neural Network Extensions.
The HiFi 5 DSP in NXP’s audio processors is the main engine for real-time sound processing, and also carries AI/ML inference tasks. It handles filters, mixing, noise reduction, voice enhancement, classification, and other compute-heavy audio jobs while the Arm Cortex-M7 manages control and system tasks. In the AUD10, NXP uses two floating-point HiFi 5 DSPs, each with neural-network accelerators, which means the chip can also execute data-driven audio features such as sound classification and adaptive processing. Audio Weaver makes it easier to build complex pipelines without writing everything from scratch and developers can apply their processing architectures to both devices.
www.nxp.com