Glossary
Forced alignment is the process of taking a known-correct transcript and an audio file, and producing precise start/end timestamps for each word. It's a separate operation from ASR — ASR produces the text from scratch, alignment refines the timing of an already-known text. Used to make subtitles snap to syllable boundaries, to enable click-to-jump-to-audio in editors, and to build phonetics datasets.
Why alignment matters separately from transcription
ASR systems output text plus segment-level timestamps ("this segment is 4.2 seconds, here's the text"). That's enough for paragraph-level navigation but not for word-level subtitle timing. Word-level timing is what makes captions feel synchronized — each word lights up exactly when spoken. Forced alignment closes that gap, producing per-word start/end times accurate to ~30 ms.
How alignment works
Two main approaches. (1) Viterbi alignment: take the reference transcript, convert each word to a phoneme sequence, and use a Hidden Markov Model to find the most-likely time-alignment of phonemes to the audio. Classic approach (HTK, Kaldi). (2) Attention-based alignment: in transformer ASR systems like Whisper, the cross-attention weights between encoder (audio) and decoder (text) implicitly encode alignment. Tools like WhisperX extract these to produce word-level timestamps without a second pass.
When you'd use forced alignment in practice
Three big use cases. Burn-in karaoke captions for TikTok/Reels — each word highlights as spoken. Click-to-jump editors like Descript — click a word in the transcript, jumps to that moment in the audio. Linguistic research — building corpora of speech with phoneme-level timing for pronunciation studies, dialect mapping, prosody research.
In practice
When you export an SRT file from Transcript.you, the timestamps come from Whisper's segment timing — accurate enough for normal subtitle reading. If you need word-level precision (e.g., for animated captions where each word pops in separately), enable our "word-level timestamps" option which uses WhisperX-style attention-based alignment. Adds ~5 seconds to processing for a 10-minute clip.
Related terms
Further reading
Generate timed subtitles
Get started — freeLast updated: September 20, 2026
Frequently Asked Questions
What does Forced alignment mean in transcription?
Matching a known transcript to its audio to produce per-word timing. Used to improve subtitle precision after the initial ASR pass.
Why does Forced alignment matter when choosing a transcription workflow?
Forced alignment affects how you understand transcript quality, timing, compatibility, or the technology behind speech-to-text results.
Where can I apply Forced alignment on Transcript.you?
Use the related workflow at /audio-to-srt when you want to see how this glossary concept connects to an actual transcription task.
Browse the full transcription glossary or read the complete guide to AI transcription.