Skip to main content
Version: Next

Pause Characteristics

Fundamental frequency is usually zero when there is no voice detected. Using this understanding, frames where voice is or is not present can be determined and used to characterize pauses during speech and silence during the audio file. These metrics are quantified here.

Raw Variables

VariableDescription
aco_pausetimePause time. Length of time with no speech detected.
aco_totaltimeVideo length. The length of the video.
aco_speakingtimeTime spoken. The total length of time with speech detected.
aco_numpausesNumber of pauses. Number of instances with no speech.
aco_pausefracPause time. aco_pausetime divided by aco_totaltime.

Derived Variables

VariableDescription
aco_pausetime_meanPause time. Length of time with no speech detected.
aco_totaltime_meanVideo length. The length of the video.
aco_numpauses_meanNumber of pauses. Number of instances with no speech.
aco_pausefrac_meanFraction of video with pauses. aco_pausetime_mean divided by aco_totaltime_mean.

Note: The overlap between raw and derived variables for this section may be confusing; it’s a leftover effect of how our code is organized in the AiCure product, where several videos from the same individual at the same time point are averaged. For your purposes, simply rely on the derived variables here.