I'm not sure if (the latter half of) your post is joking; 16kHz voices are much better on the ears, as they can handle frequencies up to about 8 kHz, while 8kHz can only get a maximum of 4 kHz; while that's enough for recognizable speech, and even speaker recognition, it sounds really bad. IMHO.

I'm much better with 16 kHz, but I suppose they can be packaged separately. Just as long as they don't get left out entirely...