As above emulators are not hardware but most of the time something like this that will cause a crash on hardware will also cause the emulator to throw a wobbly.
Anyway VA = voice acting I assume.
This will then depend upon the file formats used to house them
Generally people will either reduce the individual volume of the samples in question to 0 (for the SDAT format* used by the vast majority of DS games then many will have volumes noted internally) or replace the samples themselves with silence (or possibly loop back if it starts silent so you get an infinite loop of silence, though less ideal for samples). The big boy hacker might remove any calls to the sound in the first place, however that is tedious.
*has several sub formats that handle different aspects
SSEQ is the midi a like. Will be flanked by sbnk files that house the instruments. Unlikely to contain voicework but not impossible.
SWAV, occasionally collected into big SWAR archives of lots of them (can be many hundred). Very short clips, though enough for basic shouts/battle cries/catchphrases and whatnot.
STRM, presumably short for stream. This can be minutes of audio and thus for any really long monologues probably what you will see used (though some might abuse SWAV).
To the best of my knowledge there are no nice editors here to do things so you will have to go manual (generate a nice SMAP file with something like vgmtoolbox first). If you are content with a full rebuild then I think tinke has options these days.
There are however other formats
https://gbatemp.net/threads/the-various-audio-formats-of-the-ds.305167/
like a 250 dollar kit back when DDR4 was Intel only
