??? 03/29/06 00:44 Read: times |
#113342 - I think so Responding to: ???'s previous message |
All human languges are generated from a finite set of phonemes. Not languges use everyone. Way back I spoke to the guy who coded the old Radio Shack Text to speech chip. He got it into and TMS9000 (??) micro. Each letter sound breaks up into a unique set of phonemes. Read a Letter spit out the sound pattern ( 1 or more phonemes). The rule come in when each letter sounds different in different words. I am told english is the worst in this regard. (long sound, short sounds, silent letters and such). Google say english use 41. I do not know where you get charts for other languges, or the rules.
|
Topic | Author | Date |
Text to Speech different languages appli | 01/01/70 00:00 | |
Availability? | 01/01/70 00:00 | |
Ask | 01/01/70 00:00 | |
languages | 01/01/70 00:00 | |
I think so | 01/01/70 00:00 | |
letters and sounds | 01/01/70 00:00 | |
russian (slavic) are - relatively - easy | 01/01/70 00:00 | |
an interest fact | 01/01/70 00:00 | |
ñ | 01/01/70 00:00 | |
Spansh user says: | 01/01/70 00:00 | |
try feeding it "La Jolla" | 01/01/70 00:00 | |
If it can say 'set' then you are set....![]() | 01/01/70 00:00 | |
a few points | 01/01/70 00:00 | |
Cool, but good luck getting 'em | 01/01/70 00:00 |