Vocalizations of full-term newborns occur in a short latency time during the neonatal period. Contingent response time of preterm babies is still unknown. An increase of preterm babies’ vocalizations following exposure to parental speech was also observed. Mothers and babies co-modulate their vocalizations in preterm dyads. Purpose: To observe temporal features of maternal and infants’ vocalizations in speaking and singing conditions in preterm dyads. Methods: In a NICU mothers (N = 36) were invited to speak and to sing to their preterm infants during Kangaroo Care. Microanalysis of temporal units were performed with ELAN Software. Results and conclusions: Preterm infants vocalize less often while their mothers speak and sing than during baseline and their vocalizations tend to be more alternating in the speaking condition and more overlapping in the singing condition. It is also concluded that preterm infants take more time to respond to maternal speaking than to maternal singing.