Finite state and Constraint Grammar based Text-to-Speech processing
View the project on GitHub giellalt/speech-sme
This document is a suggestion on how to convert single letters to pronouncable words in North Sami. This suggestion is taken from the Sammallahti/Nickel dictionary. We will have to make som adjustments, additions.
| Letter | Orthographic representation | Phonemic representation |
|---|---|---|
| A, a | a | /aa/ |
| Á, á | akseanta a | /”ak’sean:hta aa/ |
| B, b | bee | /pee/ |
| C, c | cee | /tsee/ |
| Č, č | čee | /tʃee/ |
| D, d | dee | /tee/ |
| Đ, đ | đee | /đee/ |
| E, e | ee | /ee/ |
| F, f | eaffa | /ĕaf:fa/ |
| G, g | gee | /kee/ |
| H, h | hoo | /hoo/ |
| I, i | i | /ii/ |
| J, j | je | /jee/ |
| K, k | koo | /khoo/ |
| L, l | ealla | /ĕal:la/ |
| M, m | eamma | /ĕam:ma/ |
| N, n | eanna | /ĕan:na/ |
| Ŋ, ŋ | eaŋŋa | /eaŋ:ŋa/ |
| O, o | oo | /oo/ |
| P, p | pee | /phee/ |
| R, r | earra | /ĕar:ra/ |
| S, s | eassa | /ĕas:sa/ |
| Š, š | eašša | /ĕaʃ:ʃa/ |
| T, t | tee | /thee/ |
| Ŧ, ŧ | ŧee | /θee/ |
| U, u | uu | /uu/ |
| V, v | vee | /vee/ |
| Z, z | eazza | /ĕad:tsa/ |
| Ž, ž | eažža | /ĕad:tʃa/ |
Other letters:
| Q, q | kʉʉ | /kʉʉ/ |
| W, w | duppal vee | /tup:paal vee/ |
| X, x | eaksa | /eak:sa/ |
| Y, y | yy | /yy/ (if Norwegian) |
| Æ, æ | ææ | /ææ/ |
| Ø, ø | øø | /øø/ |
| Å, å | oo | /oo/ |
| Ö, ö | øø | /øø/ |
| Ä, ä | ææ | /ææ/ |
The orthographic form must, in most cases, have two vowels, to avoid homonymi with ‘ge’, ‘de’, ‘ce’, which are existing words with short vowels. The one exception is ‘i’, which must have a single vowel, to avoid pronunciation /ij/.
The sonorants, nasals and the sibilants have a disyllabic form. This is correct when referring to the actual letter, such as: This is an s - Dá lea eassa. This is a very special use, though, and in other contexts, monosyllabic forms are preferred:
These must have a monosyllabic counterpart as well:
| F, f | eaff | /ĕaff/ |
| L, l | eall | /ĕall/ |
| M, m | eamm | /ĕamm/ |
| N, n | eann | /ĕann/ |
| Ŋ, ŋ | eaŋŋ | /ĕaŋŋ/ |
| R, r | earr | /ĕarr/ |
| S, s | eass | /ĕass/ |
| Š, š | eašš | /ĕaʃʃ/ |
| Z, z | eazz | /ĕadts/ |
| Ž, ž | eažž | /ĕadtʃ/ |
| X, x | eaks | /ĕaks/ |
The /ea/ in the disyllabic forms comes from Norwegian practice, and it is the correct pronunciation for all these letters, exept for ž, z. For these two, the natural pronunciation is with /e/:
| Z, z | ezz /edts/ | ezza /ed:tsa/ |
| Ž, ž | ežž /edtʃ/ | ežža /ed:tʃa/ |
Á, á is problematic, because the well-known words guhkes a and čuvges a refer to phonetic characteristics in two main dialects. guhkes a will mean nothing to people from eastern Finnmark. The akseanta a for á is dialect neutral, although it might be less known and less used than the two others. We will have to look in real texts what kind of realization this letter should have. The akseanta-a seems to be appropriate in texts about IT-stuff. However, in abbreviations, it is possible that only aa is natural.