Re: [Espeak-general] Catalan pronunciation issues

[Jonathan D. <jo...@ta...>]

On 12 May, Fergus Henderson <fe...@go...> wrote:

> Forwarding on a bug report from a user of translate.google.com.

> > Unlike Spanish, in Catalan 'ia' is not considered a diphthong when
> > at the end of a word

> I had a quick glance at the sources.
> The file dictsource/ca_rules seems to be the place which defines the
> pronunciation rules for Catalan,
> but the fix is not obvious.  Any suggestions?

Yes.  I have now added a line to ca_rules (line 95):

      i (a_    i

This means that letter 'i' when followed by 'a' and end-of-word
translates to vowel phoneme [i] rather than phoneme [j] which was given
by the more general rule for when 'i' is followed by any vowel:

      i (A     j 

The command:

espeak -X -v ca "secretaria"

shows which rules in ca_rules are being considered, together with their
"scores".

> BTW, line 188 seems to be a bit odd:

>         tm         mm	Geminació

> Should Geminació there be commented out?

Yes, although in practice it's ignored by eSpeak because a space
character terminates the phoneme string [mm].  This seems a strange
rule; it translates 'tm' to a double [m], but it's plausible and I
guess the comment confirms that it should be a "geminated" (doubled)
consonant.

Here is the modified  ca_rules  file (attached).  It needs the command:

espeak --compile=ca

to make a new version of espeak-data/ca_dict  which is the compiled
Catalan data which eSpeak uses during speech synthesis.

I will include (within the next couple of days) it in the next
development version of eSpeak after 1.43.24, at:
http://espeak.sf.net/test/latest.html

What is your policy for picking up a new version of eSpeak?  Do you
take the current official release (currently 1.43.03), or are you tied
to a particular OS such as some Linux distribution, or can you take
development versions?