Techniques are described herein that are capable of adaptively modulating audio content based on background noise. For instance, a loudness difference between a loudness of the audio content and a loudness of the background noise is determined. The loudness of the audio content is compared to an upper loudness threshold. A modulation is selected from multiple modulations based on the loudness difference and/or the comparison. Subsequent outputted audio content is modulated using the selected modulation.