Etymologiae, Book III, chapter 20: De prima divisione musicae, quae harmonica dicitur
Chapter XX. On the first division of music, which is called harmonic The first division of Music, which is called harmonic, that is, the modulation of the voice, belongs to comedians, tragedians, or choruses, or to all who sing with their own voice. This makes motion from the mind and the body, and from motion sound, from which Music is gathered, which in man is called voice. Voice is air struck by breath, whence words (verba) also are named. Properly, however, voice belongs to men or to irrational animals. For in other things the sound is called voice by misuse, not properly, as “the voice of the trumpet roared,” and “voices broken from the shore.” For it is proper that the rocks of the shore resound, and: “But the trumpet with its brazen note afar a dreadful sound.” Harmonic is the modulation of the voice and the concord, or fitting together, of very many sounds. Symphony is the temperament of modulation, from low and high sounds in concord, whether in voice, or in breath, or in beat. For through it the higher and lower voices agree, so that whoever sounds out of tune with it offends the sense of hearing. The contrary of this is diaphony, that is, voices that disagree or are discordant.
Euphony is the sweetness of the voice. This is also called melos, from sweetness and honey (mel). A diastem is an interval of the voice fitted from two or more sounds. A diesis is certain spaces and deductions of modulating, inclining from one sound to another. Tone is the sharp utterance of the voice. For it is the difference and quantity of harmony, which consists in the accent or tenor of the voice: the kinds of it the musicians divided into fifteen parts, of which the hyperlydian is last and highest, the hypodorian the lowest of all. Song is the inflection of the voice, for sound is direct; sound precedes song. Arsis is the raising of the voice, that is, the beginning. Thesis is the setting down of the voice, that is, the end. Sweet voices are subtle and dense, clear and high. Clear voices are those that are drawn out farther, so that they fill every place continuously, like the blare of trumpets. Subtle voices are those which have no breath, such as that of infants, or women, or the sick, as in the strings. For the strings that are finest give out subtle and thin sounds. Full voices are those when much breath goes out at once, as of men. A high voice is thin, lofty, as we see in strings. A hard voice is one that sends out sounds violently, like thunder, like the sounds of an anvil, whenever the hammer strikes the hard iron. A rough voice is hoarse, and is scattered through small and dissimilar beats. A blind voice is one which, as soon as it is sent forth, falls silent, and, stifled, is not drawn out farther, as in earthenware. A vinnola voice is soft and flexible. And vinnola is named from vinnus, that is, a softly curled lock. A perfect voice is high, sweet and clear: high, so as to suffice to the heights; clear, so as to fill the ears; sweet, so as to soothe the minds of the hearers. If any of these is lacking, the voice is not perfect.