An audio playback device may drive an audio reproduction device at a volume level. An audio capture device may convert sound waves into audio input. An environment sensing device may detect, based on the audio input, environmental conditions surrounding a user of the audio playback device, the environmental conditions including a loudness estimation indicative of a level of background noise included in the audio input and an audio content classification indicative of presence of speech in the audio input. The environment sensing device may also determine, according to the environmental conditions, a playback action to alter the volume level being provided by the audio playback device, and provide, to the audio playback device, an adjustment to the volume level in accordance with the playback action.