Noise is a proven cause of wakefulness and qualitative sleep disturbance in critically ill patients. A sound pressure level reduction can improve sleep quality, but there are no studies showing the feasibility of such a noise reduction in the intensive care unit (ICU) setting. Considering all available evidence, we redesigned two ICU rooms with the aim of investigating the physiological and clinical impact of a healing environment, including a noise reduction and day-night variations of sound level. Within an experimental design, we recorded 96 h of sound-pressure levels in standard ICU rooms and the modified ICU rooms. In addition, we performed a sound source observation by human observers. Our results show that we reduced A-weighted equivalent sound pressure levels and maximum sound pressure levels with our architectural interventions. During night-time, the modification led to a significant decrease in 50 dB threshold overruns from 65.5% to 39.9% (door side) and from 50% to 10.5% (window side). Sound peaks of more than 60 decibels were significantly reduced from 62.0% to 26.7% (door side) and 59.3% to 30.3% (window side). Time-series analysis of linear trends revealed a significantly more distinct day-night pattern in the modified rooms with lower sound levels during night-times. Observed sound sources during night revealed four times as many talking events in the standard room compared to the modified room. In summary, we show that it is feasible to reduce sound pressure levels using architectural modifications.