target-speaker-extraction

VoxLock 是一个无需注册步骤的因果性目标说话人提取模型，通过微调 REAL-TSE 实现，在真实录音上达到接近商业级的语音识别精度（Krisp 基准 WER 4.95%），且 CPU 占用和延迟极低，模型与代码均已开源。

Falcon is a causal audio-visual target speaker extraction model that uses lip video to isolate a target speaker's voice in noisy, overlapping speech, cutting parameters by 91.9% and latency by 11.1× versus causal AV-TFGridNet. The paper also introduces RealSSA, a benchmark spanning simulated and real-recorded acoustic scenes with varied interference conditions.