|
|
|
|
|
|
|
|
| `K` | Not a vowel (i.e. a consonant or word boundary or non-alphabetic character). | |
|
|
| `K` | Not a vowel (i.e. a consonant or word boundary or non-alphabetic character). | |
|
|
| `X` | There is no vowel until the word boundary. | |
|
|
| `X` | There is no vowel until the word boundary. | |
|
|
| `Z` | A non-alphabetic character. | |
|
|
| `Z` | A non-alphabetic character. | |
|
|
| `%` | Doubled (placed before a character in \<pre\> and after it in \<post\>. | |
|
|
|
|
|
|
|
|
| `%` | Doubled (placed before a character in \<pre\> and after it in \<post\>. (Beware of [bug #195](https://github.com/espeak-ng/espeak-ng/issues/195)) | |
|
|
| `/` | The following character is treated literally. | |
|
|
| `/` | The following character is treated literally. | |
|
|
| `\xxx` | Character is written as by 3 digit octal value of `xxx`| |
|
|
| `\xxx` | Character is written as by 3 digit octal value of `xxx`| |
|
|
| `@` | One syllable (i.e. at least one vowel or diphthong) | |
|
|
| `@` | One syllable (i.e. at least one vowel or diphthong) | |
|
|
|
|
|
|
|
|
``` |
|
|
``` |
|
|
_) a A // "a" at the start of a word |
|
|
_) a A // "a" at the start of a word |
|
|
a (CC A // "a" followed by two consonants |
|
|
a (CC A // "a" followed by two consonants |
|
|
a (C% A // "a" followed by a double consonant (the same letter twice) (befare of [bug #195](https://github.com/espeak-ng/espeak-ng/issues/195)) |
|
|
|
|
|
|
|
|
a (C% A // "a" followed by a double consonant (the same letter twice) |
|
|
a (/% A // "a" followed by a percent sign |
|
|
a (/% A // "a" followed by a percent sign |
|
|
%C) a A // "a" preceded by a double consonants |
|
|
%C) a A // "a" preceded by a double consonants |
|
|
@@) bi bI // "bi" preceded by at least two syllables |
|
|
@@) bi bI // "bi" preceded by at least two syllables |