Show HN: Boost Your Voice AI Agents with Open-Source Ten VAD

https://github.com/TEN-framework/ten-vad

8•Jingyi0321•7mo ago

Voice Activity Detection (VAD) is a crucial component for Voice AI, enabling more natural and efficient interactions. TEN VAD is an open-source solution designed to supercharge your Voice AI Agents with lightning-fast, human-like conversations! TEN VAD offers some key advantages:

ONNX Support: Deploy on virtually any platform or hardware architecture! This means greater flexibility and easier integration into your existing systems. Superior Detection Accuracy: Experience noticeable improvements in voice detection, leading to fewer errors and more reliable performance. Smaller & Faster: Enjoy a 32% reduction in Real-Time Factor (RTF) and an 86% size reduction compared to Silero VAD! This translates to lower resource consumption and faster processing.

Get the code: https://github.com/ten-framework/ten-vad

Comments

JuneWW•7mo ago

This looks really promising! VAD is such a critical piece of the puzzle for voice AI. Definitely going to check this out. Thanks for sharing!

fm100•7mo ago

What are the differences between TEN VAD and WebRTC VAD?

Jingyi0321•7mo ago

In general, WebRTC VAD uses pitch information for VAD. Note that pitch only appears in voiced speech, but not in unvoiced speech. With this characteristic, WebRTC VAD may fails in detecting the start of a word, losing the unvoiced start, which will then result in e.g. increased WER in ASR system. On the other hand, noise whose spectrum is similar to voiced speech, e.g. music, may be extracted a non-zero pitch by WebRTC VAD pitch detection system.

Our model incorporates fbank and the pitch information together, and can analyse the input pattern deeply, therefore has better performance than WebRTC VAD.

rambo11•7mo ago

Thanks for sharing this awesome VAD model. A high-performance and low latency VAD is very helpfull in Conversation AI agents.

strassenbahn•7mo ago

It seems the performance is much better than the existing VAD SOTA Silero VAD and the size is much smaller. Good to see this new SOTA VAD model!

Feedback on a client-side, privacy-first PDF editor I built

Clay Christensen's Milkshake Marketing (2011)

Show HN: WeaveMind – AI Workflows with human-in-the-loop

Show HN: Seedream 5.0: free AI image generator that claims strong text rendering

A contributor trust management system based on explicit vouches

Show HN: Analyzing 9 years of HN side projects that reached $500/month

The Floating Dock for Developers

Arcan Explained – A browser for different webs

We are not scared of AI, we are scared of irrelevance

Quartz Crystals

Show HN: I built a free dictionary API to avoid API keys

Show HN: Kybera – Agentic Smart Wallet with AI Osint and Reputation Tracking

Show HN: brew changelog – find upstream changelogs for Homebrew packages

Any chess position with 8 pieces on board and one pair of pawns has been solved

LLMs as Language Compilers: Lessons from Fortran for the Future of Coding

Projecting high-dimensional tensor/matrix/vect GPT–>ML

Show HN: Free Bank Statement Analyzer to Find Spending Leaks and Save Money

Our Stolen Light

Matchlock: Linux-based sandboxing for AI agents

Show HN: A2A Protocol – Infrastructure for an Agent-to-Agent Economy

Drinking More Water Can Boost Your Energy

Proving Laderman's 3x3 Matrix Multiplication Is Locally Optimal via SMT Solvers

Fire may have altered human DNA

"Compiled" Specs

The Next Big Language (2007) by Steve Yegge

Open-Weight Models Are Getting Serious: GLM 4.7 vs. MiniMax M2.1

Using AI for Code Reviews: What Works, What Doesn't, and Why

Show HN: Solnix – an early-stage experimental programming language

DoNotNotify is now Open Source

The British Empire's Brothels