TY - RPRT TI - PixIT: Joint Training of Speaker Diarization and Speech Separation from Real-world Multi-speaker Recordings AU - Joonas Kalda AU - Clément Pagés AU - Ricard Marxer AU - Tanel Alumäe AU - Hervé Bredin PY - 2024 DO - 10.21437/odyssey.2024-17 UR - https://arxiv.org/abs/2403.02288 ID - 2403.02288 ER -