SHANGHAI, Aug. 5, 2026 /PRNewswire/ — From voice calls to intelligent interaction, from music enjoyment to vlog creation, audio is no longer mere signal processing — it has become the invisible infrastructure of the digital era. As user expectations for sound that is “clear, authentic, and immersive” continue to rise, audio technology is moving from behind the scenes to center stage.

As a leading force in mobile chipsets, UNISOC is leveraging UNISOC Lyric Audio technology to deepen its capabilities across four major scenarios — calling, interaction, playback, and recording — making sound a digital bond that connects, heals, and empowers creation.

Device-Edge AI Intelligent Voice: A More Natural Voice Interaction Experience

The device-edge AI intelligent voice capability of UNISOC Lyric Audio builds a complete on-device pipeline spanning wake-up, recognition, and understanding. Built on this technology framework, voice capabilities are deployed deeply on the device itself — from voice wake-up and voiceprint recognition to speech recognition and semantic understanding, all completed locally. This reduces reliance on the cloud and delivers an always-responsive, stable, and reliable interaction experience.

Specifically, the full-pipeline low-power voice wake-up technology adopts a software-hardware co-design to achieve ultra-low-power standby wake-up, keeping the device “always ready.”

Voiceprint recognition precisely matches the user’s voiceprint features against the enrolled voiceprint, providing localized protection for identity authentication and effectively reducing the risk of privacy leakage.

The command word recognition engine responds rapidly to basic voice commands, with recognition accuracy at an industry-leading level.

On-device automatic speech recognition (ASR) completes recognition processing without uploading voice data to the cloud, significantly reducing dependence on the network.

Lightweight natural language understanding (NLP) intelligently classifies user voice intents and accurately extracts key information from complex commands. It currently supports 400 common expressions across six core control scenarios, including music, brightness, and volume, helping terminal devices upgrade to a safer, more natural, and lower-latency voice interaction experience.

AI All-Scenario Noise Reduction: Clearer Voice, Purer Communication

AI all-scenario noise reduction covers all service scenarios, including cellular calls, VoIP, and recording. It innovatively offers multi-tier solutions — lightweight AI voice enhancement, flagship AI vocal prominence, and AI voiceprint noise reduction — flexibly matching terminal requirements from mainstream to flagship.

Among them, the lightweight AIECNR 1.0/2.0 solution is deployed on a dedicated audio ADSP chip, tackling everyday environmental noise with extremely low power consumption and latency. Combining AI noise reduction with echo suppression, it effectively overcomes the echo cancellation challenge in double-talk scenarios and has been validated in mass production in multiple leading customer projects.

For more demanding requirements, the flagship AI vocal prominence solution supports both cellular calls and VoIP scenarios with echo and noise reduction capabilities. It delivers stronger suppression of complex noises such as road noise and tapping sounds, significantly improving voice restoration quality in noisy environments such as subways and shopping malls — making voices stand out and communication clearer.

AI voiceprint noise reduction leverages enrolled voiceprint information and neural network capabilities to precisely separate and enhance the target speaker’s voice from mixed audio, effectively filtering out ambient noise and other people’s voices.

Built on UNISOC’s software-hardware co-designed platform capabilities, AI all-scenario noise reduction ensures that users always enjoy clear voice and pure communication across every scenario.

AI Voice Communication Enhancement: Smooth Communication Even in Weak-Network Scenarios

For extreme weak-network scenarios such as satellite communication, off-grid communication, and emergency rescue, UNISOC integrates two core capabilities — AI Codec (AI-powered codec) and AI PLC (AI packet loss concealment) — to ensure smooth and clear calls even in the most challenging environments.

AI Codec intelligent voice coding/decoding technology introduces AI generative reconstruction, breaking through the bandwidth limits of traditional voice codecs and achieving equivalent voice quality at only 30% of the bitrate of conventional solutions. In satellite communication scenarios, it effectively compensates for severe signal attenuation caused by mountain obstruction, cloud cover, and building blockage. In terrestrial off-grid communication scenarios, AI Codec significantly extends communication range at equivalent voice quality compared with traditional codec solutions, and greatly improves connection rates and call quality in extreme long-range conditions.

Meanwhile, AI PLC intelligent packet loss concealment targets all kinds of weak-network scenarios — whether fluctuating satellite links, the edge of off-grid communication range, or weak cellular signals. When network instability causes voice data loss, it predicts and reconstructs the missing voice segments in real time, keeping speech natural and continuous, without stuttering or interruption.

Together, AI Codec and AI PLC form an end-to-end voice communication enhancement technology with dual weak-network resilience — “ultra-low-bitrate transmission plus intelligent packet loss concealment.” It can widely serve critical scenarios such as field exploration, ocean operations, mountain rescue, and extreme-weather emergency response, enabling users in remote or critical environments to enjoy a reliable, clear, and uninterrupted voice communication experience.

AI Ultrasonic Fused Proximity Sensing: Accidental-Touch Prevention During Calls, Redefining Value with Chips

AI ultrasonic fused proximity sensing reuses the phone’s earpiece and microphone to transmit and receive customized high-frequency ultrasound, and fuses real-time sensor data from the accelerometer, gyroscope, and more. During calls and VoIP scenarios, it accurately determines the relative position between the phone and the user’s face, intelligently turning the screen on and off. This solution not only delivers an experience upgrade featuring a hole-free design, strong interference resistance, and high precision, but also further optimizes terminal costs:

Hole-free integrated design: supports hole-free full-screen solutions, increases the screen-to-body ratio, and gives the device a cleaner, purer appearance;Superior environmental interference resistance: operates reliably even under complex conditions such as strong ambient light or a smudged screen;Precise and responsive, free from false triggers: ultrasound combined with motion and posture recognition effectively reduces the probability of false screen on/off judgments;Ultimate cost optimization: hardware reuse reduces dependence on additional components, allows more flexible screen selection, and strengthens the terminal’s cost advantage.

This feature is about to enter mass production on the T7300 platform. Leveraging the platform’s powerful software-hardware co-design capabilities, the solution offers end-to-end advantages of high integration, low power consumption, and easy deployment.

Diverse Audio Scenarios: A Complete All-Scenario Solution

Beyond the core technologies above, UNISOC Lyric Audio also delivers leading capabilities across diverse scenarios, including wireless audio, audio recording, audio playback, and gaming audio.

For wireless audio, the solution is capable of Hi-Res Audio Wireless certification and supports wireless audio transmission, spatial audio characteristic tuning, and Bluetooth LE Audio;For audio recording, in addition to noise reduction, it provides an integrated audio-visual solution that includes HDR recording, Audio Zoom, and Audio Focus;For audio playback, it supports end-to-end playback algorithms covering high-volume output, speaker control, and other technologies;For gaming audio, it supports customer customization needs, offering a complete solution that includes in-game AI voice enhancement, gaming sound effects, a low-latency gaming audio path, and intelligent gaming power management.

Built on UNISOC’s integrated software-hardware audio platform, UNISOC Lyric Audio has established four core technologies — AI all-scenario noise reduction, device-edge AI intelligent voice, AI voice communication enhancement, and AI ultrasonic fused proximity sensing — together with a complete all-scenario solution covering recording, playback, and gaming audio.

Going forward, UNISOC Lyric Audio will continue to advance its AI audio capabilities, evolving the audio experience from “clear and pleasant” toward “intelligent perception, natural interaction, and all-scenario adaptation,” bringing users a smarter and more immersive next-generation audio experience.

Note: The data in this article comes from UNISOC laboratories.

View original content to download multimedia:https://www.prnewswire.com/apac/news-releases/unisoc-lyric-audio-redefining-the-listening-experience-302843303.html

SOURCE UNISOC