The speaker is not only speaking but also listening to his own voice, hence he keeps the Volume of voice at suitable hearing level. It means that if the volume is adequate for self-hearing he assumes that it is adequate for the persons he is speaking to. As a result he keeps the volume of his voice at the level he thinks is adequate. This has become instinctive and natural habit so one does not have to make conscious effort.
In the noisy environment of a crowded restaurant everybody speaks somewhat louder because he starts finding his own normal volume rather inadequate. Short sentences in hushed tones falling on the ears in the library are not as much due to discipline as the silence pervading in the reading room. In short, speaker adjusts volume of his speech to the volume of sound falling on his ears. Since headphones covering the ears (even if the music is not playing) obstruct hearing one unconsciously utters loudly.