Unverified50% confidenceBenchmarkExact time

Meta的EnCodec在24kHz音频上仅需1.5kbps即可实现高质量重建，约为传统MP3的十分之一

1

Sources

50%

Confidence

Long-term

Relevance

7/16/2026

First Seen

Sources

ChatGPT Voice Goes Full-Duplex: A Deep Dive into GPT-Live 1.0

Related Claims

VerifiedA typical EnCodec configuration uses 8 layers of RVQ with a codebook size of 1024 per layer, producing 75 frames per second at a 24kHz sampling rate, compressing 1 second of audio into 600 tokens.79% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/528668

API

curl https://kongchang.com/api/v1/knowledge/claims/528668

MCP

get_claim(id=528668)