Chrome silently ships a 425KB neural echo cancellation model; reverse-engineer finds 5.4dB gain
chadwallacehart · x · 2026-10-08
Philipp Hancke at webrtcHacks discovered that desktop Chrome quietly downloads a 425KB TFLite neural network that takes over part of acoustic echo cancellation:
- Default-on, no opt-out: the model runs in the getUserMedia({audio: true}) pipeline with no flag or constraint, exposed to all web pages in Chrome only — other browsers, even Chromium-based ones, don't get the weights.
- Big gains: a median 5.4dB of extra echo suppression on a hard echo corpus, and it halves the word error rate a transcribing voice agent sees on the near-end talker.
- Trade-off: allocating TFLite structures adds a small getUserMedia initialization delay (Chromium issue b/442444736).
- The post also recaps how AEC3 (launched 2018) works and reverse-engineers the black-box model against it.
More from Infra
- Nebius up 160% vs CoreWeave's 10%: the AI infrastructure stock divergence explained — Beth_Kindig · 2026-10-08
- Microsoft's $5,999 Surface RTX Spark Dev Box preorders open, ships November — tomwarren · 2026-10-08
- Together scales open-source inference with IBM and NVIDIA on B300 cluster — togethercompute · 2026-10-08
- How a Self-Appending Summarizer Quadrupled Token Costs in a 1.4M-Conversation Agent — Good_Education4713 · 2026-10-08
- National Compute gifts $100M in compute credits to White House Genesis Mission — typewriters · 2026-10-08
- Ask: Is Qwen 3.8 Flash-Next Worth It Over 3.6 35B-A3B on 64GB RAM + 16GB VRAM? — msalsas · 2026-10-08