Apple has introduced a suite of Audio Intelligence features integrated into iOS 27, macOS Golden Gate, and firmware for AirPods 5 and AirPods Pro. These tools leverage on-device deep learning to handle real-time translation, noise cancellation, and audio processing without cloud connectivity.
The new real-time translation feature uses the device's neural engine to provide bidirectional language interpretation. It allows users to hear translations directly in their AirPods during face-to-face conversations, with an option to output synthesized speech via iPhone speakers.
Voice Isolation 2.0 technology utilizes neural models trained on millions of hours of acoustic data to suppress background noise, such as wind or crowd sounds. The system adapts to speaker orientation and beamforming to isolate voices for calls, FaceTime, and voice memos.
For content creators, new AI tools on M-series Mac and iPhone devices enable the separation of audio tracks into vocals, instruments, and ambient noise. This allows for independent volume and reverb adjustments, simplifying post-production tasks like podcast editing and mixing.
The Adaptive Hearing algorithm adjusts equalization and spatial audio in real-time by monitoring ear canal seal and environmental acoustics. The features will be available globally following the official system updates and hardware releases scheduled for September 14.
