libsonic source hosting

[Jonathan Duddington]

On 08 Dec, William Hubbs <w.d.hubbs at gmail.com> wrote:

> Actually I think also that Bill was talking with Jonathan about
> espeak integration, but I don't know what happened there.

The current development version of eSpeak at:
http://espeak.sf.net/test/latest.html

includes Sonic, as a file sonic.cpp compiled with eSpeak.

Some more experimentation is needed to find the best way to use it.  It
seems that using Sonic for speed-up at ratios of less than times-2
gives some noticeable distortions.  So currently (as an experiment)
eSpeak only uses Sonic at speeds greater than 350 wpm, at which point
it drops its own speed to half that (175 wpm).  Probably eSpeak should
drop to half-speed but keep some of the changes which it makes to the
relative lengths of various sounds at high speeds, partly to avoid a
sudden change at the 350 (or whatever) changeover speed, and partly
because it's probably better for intelligibility.  This will take some
work to calibrate correctly.

If a synthesizer can use Sonic intelligently then we wouldn't need
Speech Dispatcher (or NVDA or other higher-level entity) to also apply
a speed-up.  However such a speed-up may be useful for other
synthesizers which people what to use at faster speeds than they
provide.