Question: should the kvcache be reset when I use another voice?

[libTorrentUser]

Suppose I did this

tts = api.TextToSpeech();
audio = tts.tts_with_preset("this is voice X", voice_samples=voiceX, ...);

and now, still using that same tts object created above

audio = tts.tts_with_preset("this is voice Y", voice_samples=voiceY, ...);

is that OK? I mean, is it OK to change voices without (somehow) reseting the kvcache? Should I create a new TextToSpeech object if I want to use another voice or is it OK to keep using the same one? I understand the concept of a cache, but I am not exactly sure how it is being used here. Do the selected voice and the cached values have any connection?

I'm assuming that whatever it is that is being cached has nothing to do with the selected voice, so the cached values can be reused no matter what voice we use. Is that true?