ANC and ENC Are Not the Same Thing
Procurement routinely confuses ANC and ENC. ANC works on the downlink, making what the wearer hears quieter through reference mics, an inverted signal, and cancellation in the ear canal, evaluated by depth and bandwidth. ENC works on the uplink, cleaning what the far end hears through beamforming and spectral subtraction, evaluated by SNR gain and voice distortion. ANC depth does not predict ENC performance, so RFQs must define both separately.
Microphone Architectures: Feedforward, Feedback, and Hybrid
ANC implementation depends on microphone position. Feedforward mics sit outside the shell and catch noise before the ear canal, effective to 3-4 kHz but blind to fit variations. Feedback mics sit inside in front of the driver and correct residuals and unit variation, but limited loop gain invites howling above 1-2 kHz. Hybrid designs use both and dominate TWS.
| Architecture | Mics | Bandwidth | Fit sensitivity | Cost |
|---|---|---|---|---|
| Feedforward | 1 | Up to 3-4 kHz | High | Low |
| Feedback | 1 | Below 1 kHz | Low | Low |
| Hybrid | 2-3 | 20 Hz to 4 kHz | Medium | Medium-high |
Cancellation Depth, Bandwidth, and Listening Trade-offs
Cancellation depth misleads as a single number: a marketed 45 dB is peak depth near 100-200 Hz, so request the full curve. Typical hybrid curves deliver 25-40 dB at 100-500 Hz where engine and HVAC noise live, and under 10 dB above 2 kHz. Over-deep bass brings ear pressure and audible self-noise, so tuning targets the noise spectrum of the intended scenario, not maximum depth.
Transparency Mode and Wind Noise Handling
Transparency mode replays ambient sound captured on feedforward mics with gain control, high-frequency compensation, and limiters against slams, rather than opening a vent. Quality shows in naturalness, especially occlusion of your own voice: good tuning attenuates 200-500 Hz and lifts 2-5 kHz. Wind noise is separate physics, since airflow on diaphragms creates broadband noise that spectral subtraction cannot fix, addressed by meshes, labyrinth inlets, and dual-mic wind detection verified on real rides.
The Uplink Call Noise Reduction Chain
Uplink ENC keeps the target voice in noise, in four steps. Beamforming across two or three mics with 8-20 mm spacing builds directivity toward the mouth with an effective floor near 1.5-2 kHz. Noise estimation updates during speech pauses using minimum statistics or neural masks, gain application via Wiener filtering or spectral subtraction balances residual noise against musical-noise artifacts, and post-processing adds AGC and limiting. Streets need 10-20 dB of improvement; subway calls need three mics and AI noise reduction.
ANC Tuning Workflow for TWS OEM Projects
ANC headphone OEM schedules spend most time on tuning. The flow runs five stages: cavity and mic layout, which caps performance; passive isolation measurement; filter design with artificial-head testing across at least 20 subjects; subjective listening in target-market scenes; and line calibration holding mic sensitivity within plus or minus 1-2 dB. Qualified TWS factories need acoustic chambers, 100 percent ANC retest, and inlet blockage detection.
Why Choose Aurora Sound
Aurora Sound is a Shenzhen source factory for audio OEM/ODM. On ANC and ENC programs, joint design of cavity, mic placement, and wind structures starts at the ID stage, in-house artificial-head benches verify objective and subjective tracks, filters are tuned per target-market noise profiles, and lines retest ANC function with mic binning. The A9 Pro platform runs hybrid ANC with a multi-mic uplink chain and configurable depth and transparency gain.
