Enhanced Audio Quality with AI-Powered Background Noise Suppression from Meeami

Meeami Technologies pioneers Audio AI, unveiling low-footprint noise suppression solutions for Cadence Tensilica HiFi DSPs. Their proprietary edge-based deep learning network identifies speech in noisy environments, suppressing noise and refining speech for distortion-free audio. These solutions, optimized for individual HiFi DSP models, underscore the versatility and superior performance of the HiFi DSP family. Notably, Meeami sets a benchmark for efficiency with Tensilica HiFi Mini, operating at 150 MCPS and a compact 250 KB size. The collaboration leverages Cadence’s DSP strengths—optimized audio processing, lower power consumption—for Meeami’s advanced noise suppression, enhancing user experiences across voice-controlled devices and audio applications.

Meeami Technologies a pioneer and leader in Audio AI, Noise Cancellation, Speaker ID and Spatial Audio, announced the availability of its AI based, and low footprint background noise suppression embedded solutions for the Cadence Tensilica HiFi DSP family.

AiThority.com AI ML Trends: 10 AI In Manufacturing Trends To Look Out For In 2024

Meeami’s low footprint noise suppression uses proprietary edge-based deep learning network to identify speech in noisy conditions, suppress noise and enhance the speech envelope for clean and distortion-free audio experience. Meeami’s speech enhancement and speaker ID solutions are widely deployed and power over 40+ Million devices.

Meeami’s edge AI R&D initiatives have successfully optimized footprints to enhance compatibility with Cadence’s Tensilica HiFi DSPs, leading to the development of a variety of embedded AI solutions. Each solution is uniquely tailored to the specific strengths of the individual HiFi DSP models, ensuring maximized performance across the HiFi DSP family. This tailored approach underscores the distinct advantages of each DSP model, further highlighting the HiFi DSP family’s versatility and superior performance.

In the case of the Tensilica HiFi Mini, Meeami’s technology establishes a new benchmark for efficiency, operating at approximately 150 MCPS with a compact model size of merely 250 KB. This serves as a testament to Meeami’s steadfast commitment to delivering high efficiency in compact form factors, solidifying its position as an industry leader.

Meeami’s collaboration with Cadence was driven by the unmatched performance and efficiency of Cadence’s Tensilica HiFi DSPs. The HiFi DSP family offers specialized capabilities such as optimized audio processing, lower power consumption, and the ability to handle complex audio AI algorithms, making it the ideal choice for Meeami’s advanced noise suppression technology. Tensilica HiFi DSPs are widely adopted by OEMs worldwide, with billions of HiFi DSPs embedded in a vast array of products.

“Partners like Meeami drive innovation in the AI space and highlight the efficiency that our Tensilica HiFi DSPs bring to edge-based and on-device AI applications,” explained Yipeng Liu, product marketing group director for Tensilica audio/voice DSPs at Cadence. “We are excited to work with Meeami to bring ultra-low-power noise suppression to enhance user experiences in next-generation audio device communication.”

AiThority.com News: Stereolabs and Syslogic Team Up to Advance Robotic Applications that Can Withstand Outdoor Condition

“We are excited to announce a low footprint cutting edge AI solution on Cadence Tensilica HiFi DSPs to improve user experience for multiple use cases such as voice controlled smart home appliances, TWS, earbuds and earphones, headsets and hearing aids” said Harish Rajamani, Director of Engineering at Meeami Technologies.

Sasank Kottapalli, AI Lead for Speech Products at Meeami Technologies, emphasized that the substantial reduction in AI model size, coupled with minimal performance loss, was attained through innovative design techniques and quantization methods in deep learning AI models.

Top AI ML News: LILT Announces Enterprise AI Controls and In-App Multimodal Translation

[To share your insights with us, please write to sghosh@martechseries.com]