Breathing sounds in the database

[Pilar O]

Could it be possible to add breathing sounds to the database?
I’m thinking basically on this: when we are going to say a long sentence, we will take a deeper breath first and then start talking, as part of the planning procedure of the utterance. I’ve noticed also when I’ve heard songs created with synthesizers (singing voice synthesis) that there are some points were I would have expected to hear that breath and as I don’t it is more clear that it is not natural. So I’m thinking this could add, in some situations, naturalness to the speech (but I’m not sure if in general we would notice that). Maybe it could be another unit, that according to the prosodic features could be added, but it should be included in the database too (as part of the diphones combinations?).

[Simon]

Yes, this would be pretty straightforward to do. As you say, you could treat it as a special kind of word (presumably pronounced as a special new phone).

In fact, you might find that – at least in unit selection – you will get these in-breaths ‘for free’ because phrase-initial silence diphones will be chosen to synthesise silence in phrase-initial positions.

[Author]

Posts