I agree it gets closer, but as is with much of speech synthesis, it still to me ends up sounding like "a male talking in a high-pitched voice" and not "a female".

In addition, this would imply that only males can talk in a low register, which is patently false. Low register female voices are fairly common.