A length-tagged end-to-end speech translation model with a single-pass multi-length beam search improves dubbed-video synchronization with little loss in BLEU.
The duration of translated audio is influ- enced by: (a) the length of the translated text, and (b) the dura- tion model within the text-to-speech (TTS) system
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Length Aware Speech Translation for Video Dubbing
A length-tagged end-to-end speech translation model with a single-pass multi-length beam search improves dubbed-video synchronization with little loss in BLEU.