Speakers modulate their voice when talking to infants but we know little about subtle variation in acoustic parameters during normal adult speech, and investigation is impeded by listeners’ understanding of semantic content. Here we circumvent this …