Accurate human voice reproduction

Vandersteen

Can you give us an example of a speaker and situation where a speaker can sound good with singing but not with speech that is not recording-dependent?

I think Kr4 makes a good point. The production of a spoken voice recording may be completely different than for singing. Singing is done (hopefully), for good sound, and spoken voice for communication. Going on that assumption, its only natural that there's a gap in sound quality.

IMHO, Harbeth SHL5 are magical for singing across the spectrum of genres but when I put on dialogue / NPR, I switch over to a set of Focal Micro Be in a hurry. The Harbeth sound is too intense for spoken words in my findings. I can also make a loose opposite argument for the Focal Micro Be which excels at speech in my setup but while doing very good vocals, does not stand out - perhaps because everything else sounds so good too.

To me, that sounds like an indictment of the Harbeth.