‹ Instruments

Speech timing

Read one line aloud. The page finds the syllables without a transcript.

“The north wind and the sun were disputing which was the stronger.” Or say anything you like for about five seconds. Normal pace, normal voice.

Press record and read the line aloud.

●--syllable nuclei
◉--syllables per second
∿--time spent pausing
≈--rhythm variability
0.10

Every local maximum in the envelope has a prominence: how far you must descend before you can climb to anything higher. Raise the threshold and shallow bumps stop counting as syllables. Nothing is re-recorded, the same envelope is re-segmented.

Syllable nuclei are found as prominent peaks in the amplitude envelope, so no transcript, dictionary or forced aligner is needed. Pauses are stretches below a low-energy floor lasting more than 180 milliseconds. Speech rate counts nuclei against speaking time rather than total time, and rhythm variability is the spread of the intervals between nuclei as a share of their average.

Method from: Kashyap B, Horne M, Pathirana PN, Power L, Szmulewicz D. Automated topographic prominence based quantitative assessment of speech timing in cerebellar ataxia. Biomedical Signal Processing and Control 2020, 57: 101758, doi:10.1016/j.bspc.2019.101758.

Research demonstration, not a medical device. Nothing is stored or sent. Everything is computed on this phone and disappears when you close the page. Built by Bipasha Kashyap, NSBE Lab, Deakin University.