Atlas record: https://www.aitimeline.dev/en/events/evt-0189/
Research · February 1978
Dynamic time warping
Sakoe and Chiba described how to compare two recordings of the same word spoken at different speeds, by stretching time non-linearly.
Why it matters
The most obvious obstacle to recognition was solved: the same phrase is never spoken twice with the same duration.
The algorithm finds an optimal mapping of one time axis onto another with constraints on slope, so that nonsensical stretches are ruled out. Until hidden Markov models arrived it was the main method in commercial isolated-word systems. It outlived its field: today it is used to compare any time series, from gestures to market prices.
Event record
- Event date
- February 1978
- Timeline date
- Event date
- Verification
- Sources gathered automatically · September 17, 2026
- Lines
- ID
- evt-0189
The February 1978 issue of IEEE Transactions on Acoustics, Speech, and Signal Processing.
Related events
Earlier
- Related Vintsyuk: connected speech from dictionary words 1971links to this record
Sakoe and Chiba aligned single words in time in 1978; Vintsyuk's 1971 article already recognises connected speech from dictionary words, choosing the word boundaries by the same dynamic programming.
- Contrasts Speech discrimination by dynamic programming 1968links to this record
The atlas dated this approach to February 1978. The review says the essence of the same approach was already in the 1968 work but stayed unknown in the West until the early 1980s, long after the same methods were proposed and implemented by others.
- Related Dynamic Programming 1957
The alignment is built by dynamic programming.
Later