Prior studies have shown that clear speech enhances both intelligibility and recognition memory. The simultaneous intelligibility and memory enhancements are thought to be connected - under the effortfulness hypothesis, the intelligibility-enhancing acoustic properties of clear speech (e.g., reduced speaking rate), free up cognitive resources, leading to a memory benefit. Past research, however, has only tested effects of naturally produced speaking styles for younger listeners. The current study examines the effect of speaking style, voice naturalness, and listener age on intelligibility in noise and recognition memory. Consistent with previous work, clear speech and naturally produced voices are more intelligible in noise than casual speech and text-to-speech (TTS) voices, and younger listeners have higher transcription accuracy than older listeners. However, contra the effortfulness hypothesis, the recognition memory results only partially parallel the intelligibility results. Although both intelligibility and memory benefits are found for naturally produced voices relative to TTS voices, speaking style did not affect memory for either voice type. Furthermore, the older listeners showed greater discrimination sensitivity in the recognition memory task. The current results suggest that the relationship between intelligibility and memory is complex. Implications for both theories of speech perception and TTS voice development are discussed.

This content is only available via PDF.