Peer-reviewed research
CrisperWhisper: Accurate Timestamps on Verbatim Speech Transcriptions
Laurin Wagner et al., Interspeech 2024, 2024
Read the originalRead on October 17, 2026
- What it is
- A 2024 Interspeech paper on getting accurate word timestamps from Whisper. It adjusts the model's tokenizer and fine-tunes it for verbatim transcription, then measures word segmentation against benchmarks.
- How much weight it carries
- Peer-reviewed at Interspeech. Its gains come from a fine-tuned model, so they do not show what the stock Whisper model achieves.
Facts taken from this source
- In a 2024 Interspeech study, adjusting the tokenizer of the Whisper model significantly improved the precision of its word-level timestamps.
“We demonstrate that carefully adjusting the tokenizer of the Whisper speech recognition model significantly improves the precision of word-level timestamps” (Abstract)