Search papers, labs, and topics across Lattice.
This paper introduces a smartphone-based protocol for collecting prosodic speech data in natural settings while addressing semantic confounding and privacy concerns. The protocol uses scripted read-aloud sentences to control lexical content and performs on-device prosodic feature extraction before deleting raw audio. A large-scale deployment (N=560) demonstrates the protocol's feasibility and data quality, validated through speaker sex and affective state prediction tasks.
A new smartphone protocol enables large-scale, privacy-preserving collection of prosodic speech data in the wild, opening doors to studying the subtle emotional nuances in everyday communication.
Collecting everyday speech data for prosodic analysis is challenging due to the confounding of prosody and semantics, privacy constraints, and participant compliance. We introduce and empirically evaluate a content-controlled, privacy-first smartphone protocol that uses scripted read-aloud sentences to standardize lexical content (including prompt valence) while capturing natural variation in prosodic delivery. The protocol performs on-device prosodic feature extraction, deletes raw audio immediately, and transmits only derived features for analysis. We deployed the protocol in a large study (N = 560; 9,877 recordings), evaluated compliance and data quality, and conducted diagnostic prediction tasks on the extracted features, predicting speaker sex and concurrently reported momentary affective states (valence, arousal). We discuss implications and directions for advancing and deploying the protocol.
