[hts-users:00089] a couple of general HTS questions

["Nicholas Volk" <nvolk@xxxxxxxxxx>]

Hi All,

This mailing list has been quiet for so long that I'll break
the silence with a couple of more general questions.

Heika Zen wrote in
http://kt-lab.ics.nitech.ac.jp/hts-users/spool/2004/msg00063.html:

"I'm planning to release next version of HTS (1.1.2) based on next HTK
release (3.3). I wish I could release it in this winter.
Next release will be bug-fix version.
I found a number of bugs and fixed them already."

Sounds great. When will this be released?
(Is it the official Festival 2.0 we are waiting for?)

Has anybody ever thought about getting rid of the floating
point arithmetics? The current PDAs lack FPU so the calculations
time are ..um.. interesting. (My initial experiment gave something
like 30s of processing for a 1s of speech on a S60 phone.)

The HTS voices sound extremely nice despite being small.
The Unit Selection voices are typically too big (well Cepstral
says they have Unit Selection in 2MB) and diphone and especially
formant synthesizers are just so passé state-of-the-art-wise.

best regards,
  Nicholas