Security Camera Two-Way Audio

Volume I  ·  May 2026  ·  372 words

Two-way audio allows a security camera owner to speak through the camera to a person at the door or in the monitored area. The feature appears in nearly every consumer camera, but the audio quality varies dramatically — from landline-telephone clarity to unintelligible, latency-ridden exchanges that are functionally useless for conversation. The difference is in the audio codec, the speaker and microphone hardware, and the network path the audio takes.

Audio codec. Most security cameras use the G.711 or G.726 codec for audio compression — narrowband codecs designed for telephony that sample at 8 kHz and capture frequencies between 300–3,400 Hz. This is adequate for speech intelligibility ("Who's there?" "Delivery — leave it at the door") but produces the distinctly metallic, low-fidelity sound of a telephone call. Cameras using the AAC or Opus codec at a higher sample rate (16–48 kHz) provide full-bandwidth audio that sounds natural and captures environmental sounds (footsteps, vehicle approach) with useful fidelity. The Google Nest Cam and Eufy SoloCam S340 use wideband audio codecs, though specific codec details are rarely disclosed in consumer documentation.

Speaker and latency. The camera's built-in speaker is a micro-driver — typically 20–30mm diameter, 1–2W output — that is adequate for close-range conversation within 5–10 feet of the camera but insufficient for projecting voice across a yard or driveway. The primary cause of awkward two-way conversations is latency: the delay between speaking into the phone app and the audio playing through the camera's speaker. Cloud-dependent cameras (Nest, Ring) route audio through cloud servers, adding 500–1,500 ms of latency — enough to create conversational overlap where both parties speak simultaneously, then pause, then speak simultaneously again. Locally processed cameras (Eufy, Reolink, Ubiquiti) route audio directly between the app and the camera on the local network, reducing latency to 100–300 ms and enabling more natural conversation.