TY  - RPRT
TI  - Reliable Neural-Codec Text-to-Speech by ASR Self-Verification and Distillation: Near-Zero Catastrophic Failures Across Models and Codecs
AU  - Ali Asaria
AU  - Tony Salomone
AU  - Deep Gandhi
PY  - 2026
UR  - https://arxiv.org/abs/2606.18323
ID  - 2606.18323
ER  -