[PD-announce] ratts: realtime analog text-to-speech
Bryan Jurish
moocow at ling.uni-potsdam.de
Sun Jan 12 05:33:05 CET 2003
Hi all,
Anyone who wishes to hear PD speak may wish to check
out my new "ratts" library, based on Nick Ing-Simmons'
"rsynth-2.0" and Jon Iles' "klatt3.04", and on
Travis Newhouse's "mapper", via Orm's "maphash".
This library replaces my "klatt~" external (which it
now includes), and allows full text-to-speech synthesis
within PD without any nasty resampling or shameless
cpu-blocking. There are even some rudiments for a
potential "singing voice" abstraction.
The library probably will not enter CVS until it
reaches version 1.0, when I hope to have it using
all GPL code; for the moment, there are some lurking
question marks. For now, it's available from me at:
http://www.ling.uni-potsdam.de/~moocow/projects/pd
Tested under linux.
Short summary of objects included:
EXTERNALS:
klatt~ : Klatt cascade-parallel formant speech synthesizer
holmes : Holmes/Mattingly/Shearme phonetic interpreter
phones2holmes : phone-string to Holmes-element converter
guessphones : NRL english to phone-string converter
number2text : number to english text converter
rattstok : raw text tokenizer
toupper : upper-case symbol normalizer
rattshash : Auto-resizing hash table
rattshread : Auto-resizing hash table reader
rattshwrite : Auto-resizing hash table writer
ABSTRACTIONS:
ratts~ : full text-to-speech abstraction
rattsdict : shareable dictionary
EXECUTABLE PROGRAMS:
mkdicttxt : create a "rattshash"-readable dictionary
from a raw text source (beep or cmudict).
peace, love, and marmosets,
Bryan
More information about the Pd-announce
mailing list