Audio Team Built LiquidCodec, a 74M-Parameter Streaming Audio Codec
helloiamleonie · x · 2026-11-13
An audio team published an engineering blog introducing their in-house audio codec, LiquidCodec.
What an audio codec does: it turns speech into tokens and reconstructs audio from them, in three steps — encode raw audio into feature vectors, quantize them into discrete tokens, then decode the waveform from the tokens.
Why build their own: earlier audio models used a third-party tokenizer, but over time its token quality and design choices became limiting.
LiquidCodec is a 74M-parameter streaming codec offering higher-fidelity reconstruction closer to the original, optimized for speed and fast even on-device.
More from Research
- Qwen3.5-4B fine-tuned in 27.5 min on one GPU lifts MedQA accuracy to 74.5% — MaziyarPanahi · 2026-11-13
- AromaGen transmits smells in real time with AI, hitting 71% in blind tests — pliang279 · 2026-11-13
- Muon shown to be a special case of Shampoo optimizer — _arohan_ · 2026-11-13
- BioSecBench-Structure: a benchmark for agentic molecular structure reasoning — kenbwork · 2026-11-13
- One 8×B200 node serves 3,000 DeepSeek agents, 19x official capacity — james_y_zou · 2026-11-13
- DocFS turns documents into file systems so VLM agents can reason over them — _reachsumit · 2026-11-13