Jill tells us that Mark is recording material to see if his voice is
suitable for all our needs - many thanks. Here is a case for making Mark
our speaker if his voice is amenable to easy acoustic analysis. And a
couple of questions about possible deviations from the "ideal male
southerner" in case Mark's voice turns out to pose problems. It would be
good to get reactions to the questions now so we don't waste time debating
them if we decide Mark's voice is not suitable.
Jill and I have discussed the issue at length, and (especially if we form
the "database" piecemeal as I've suggested we may need to) then it's
imperative that we use a speaker who is not only skilled, but also
amenable, available at relatively short notice, and likely to stay around
for the project and preferably for some time after this current one. Mark
fits this bill perfectly. He is also male and southern (but not high RP),
and apparently the speaker (or one of the speakers?) for PROSICE. I
haven't yet been able to think of anyone at Cambridge who is as suitable.
Any bids from York?
I suggest we try to make Mark our speaker. If we can't, do we want to
consider working with either the other sex (I assume you'll all say no!)
or another accent, e.g. "Educated Northern". Might be best to consider
these questions so that, if Mark's voice turns out to be unsuitable, we
could look at others rather soon. I for one would be happy with a mild
northern accent, amounting primarily to "grass" said in the right (I.e.
CV4 not CV5) way, but all the other vowels, and /h/, sort of
southern-like, including /k@m'pjut@/ not /.kom'pjut@/ etc. Would there be
a problem for intonation?
Re Mark's voice: it's a bit high and quite light, as he pointed out at our
last meeting. If there's no problems in formant tracking, the rest won't
matter from a technical point of view. But low-pitched, low-larynx-type
voices are often preferred in synthesis, presumably becuase they're
thought to carry more authority. (Though I:m told female voices are
preferred in instructing car drivers to belt up). Should this issue affect
the decision? I think probably not, because the advantages of Mark doing
the recording outweigh matters of fashion, and because ultimately our
work is intended to be "device independent". But I wanted to mention it
because I did regret modelling IVX's vowels closely on Andrew's after I
rather belatedly noticed that he had a short vocal tract and, especially
when in front of a microphone, speaks with raised larynx.
(Mark: forgive me for discussing your voice in such excruciating detail:
thought it best to deal with possible problems while there's still time to
avoid them! I personally think its rather a nice voice.)
best,
Sarah
______________________________________________________________________
Dr. Sarah Hawkins Email: sh110@cam.ac.uk
Dept. of Linguistics Phone: +44 1223 33 50 52
University of Cambridge Fax: +44 1223 33 50 53
Sidgwick Avenue or +44 1223 33 50 62
Cambridge CB3 9DA
United Kingdom