Traditions that chant tend to make a strong claim about sound itself: that the syllables are not a description of something sacred but are the sacred thing in audible form.
Vedic recitation is the strongest version of the claim, and it produced the most remarkable transmission system in the history of memory. The Sanskrit hymns were preserved orally for centuries before they were written, and the method was designed to make undetected error impossible: pupils learned each text in multiple interleaved permutations, reciting the words in reversed and alternating orders, so that any corruption would fail to reconcile across the versions. The accuracy achieved is extraordinary and can be checked by comparing lineages that separated a very long time ago. The reason for that effort is doctrinal: the sound is efficacious in itself, and a mispronounced formula does not work.
Buddhist practice sits differently, and the schools differ among themselves on whether the sound acts intrinsically or acts on the mind of the chanter. Tibetan practice is closest to the intrinsic view, and attaches mantras to wheels, flags and carved stones so that they continue to be produced with no person present, which only makes sense if the syllables do the work.
Sikh practice, kirtan, is sung scripture, and the scripture was composed as poetry set to named musical measures, so the singing is the form the text was made in rather than an ornament added to it.
Islamic recitation of the Quran is governed by a formal science of pronunciation, pausing and elongation, taught systematically and examined publicly, and reciters are trained professionals with international reputations. Muslims do not call it music and do not call reciters singers, and that distinction is maintained carefully rather than casually.