To remain locked in sync requires triggering a midi file. On the other hand, Mp3 audio files simply need to be short. Triggering vocal samples this way works best on songs with a signature vocal hook line/phrase that's between 1-2 measures to keep the timing from drifting out of sync with the arranger.

My friend Manuel Dorantes told me he utilizes multi-pad triggered 'vocal audio samples' (trimmed from MP3 pop tunes) on a few of his tunes,
and garnering a lot of positive audience response using them at his senior tea dance venues in Arizona.
It's entertaining & fun and adds a recognizable vocalist touch to an otherwise instrumental performance.