I'm writing a script for a voice actor and I'd like to distinguish vocalizations that are things like laughs, cries, grunts and other sounds of that sort. What is the formal term that could separate those types of vocals from actual coherent sentences?
"Vocalization" rarely if ever refers to "actual coherent sentences": those would be called "singing" (admittedly, singing includes coloratura and other stuff somewhat remote from coherent sentences). However, vocalizations usually feature at least syllables in style of singing rather than "laughs, cries, grunts and other sounds of that sort". I don't know a better term than "vocal exercises" even though this includes vocalizations and proper singing.
Now either vocalizations and vocal exercises are usually, well, exercises. You are talking about a "script for a voice actor" instead, namely actual productive output. Also it does not appear like you are actually talking about any form of singing at all, so one would actually not even use "vocalization" in this context.
To me the distinction seems to be "text" and "non-textual".