I’ve tested every single ASR alignment solution that mentioned here https://github.com/pettarin/forced-alignment-tools, but they all performed poorly compared to Aeneas, even with good language models (English)