One-page reference: VAD generations (energy / G.729-B classical / WebRTC GMM / Silero neural), the two failure modes (false negative clips words, false positive wastes bandwidth), hangover, the Opus DTX settings (usedtx default 0, ~400 ms SID, whole-frame drop, RFC 3389 discouraged), and the DTX-on-vs-off decision rule.
Download free PDF
The book · Volume 4 of 7
Audio for Video: A Complete Guide to Sound in Video Products: From PCM and Codecs to Loudness, Accessibility, and Real-Time Voice Processing
This cheat sheet covers VAD generations from energy detectors to Silero, the two failure modes, hangover, Opus DTX settings and the rule for turning DTX on or off. Volume 4 explains this in depth: real-time voice processing with VAD, AEC, AGC and noise suppression, Opus and speech codecs, and the dynamic range and level issues behind clipped words. Written by Nikolay Sapunov, CEO at Fora Soft.
Clipped words or wasted bandwidth from your VAD and DTX settings?
Send us recordings and your Opus settings: we will tune VAD, hangover and DTX for your calls and show the bandwidth and speech-quality trade-off. 250+ video products shipped since 2005.
You get a realistic estimate, technical recommendations and next steps. Free, no obligations.
