Date
1 - 3 of 3
Transforming a string into a stylized pitch contour?
Emmett STRICKLAND
That's perfect! Thank you!
Le lun. 14 nov. 2022 à 17:35, Piet Mertens <mertens.piet@...> a écrit :
Hi Emmett,For decomposing strings you can use strings functions such as: left$ (label$, 1), mid$ (label$, 2, 1), right$ (label$, 1), etc.As for the mapping from symbolic pitch levels (targets) to raw F0, taking into account speaker pitch range, pitch span, declination phenomena, and so on, you can find inspiration in:Mertens, Piet; Goldman, Jean-Philippe; Wehrli, Eric; Gaudinat, Arnaud (2001) La synthèse de l'intonation à partir de structures syntaxiques riches. Traitement Automatique des Langues 42 (1), 142-195.Mertens, Piet (2002) Synthesizing Elaborate Intonation Contours in Text-to-Speech for French in Bernard Bel & Isabelle Marlien (eds.), 2002. Proceedings of the Speech Prosody 2002 conference. 11-13 April 2002. Aix-en-Provence: Laboratoire Parole et Langage. p. 499-502.Best regards,Piet
Piet Mertens
Hi Emmett,
For decomposing strings you can use strings functions such as: left$ (label$, 1), mid$ (label$, 2, 1), right$ (label$, 1), etc.
As for the mapping from symbolic pitch levels (targets) to raw F0, taking into account speaker pitch range, pitch span, declination phenomena, and so on, you can find inspiration in:
Mertens, Piet; Goldman, Jean-Philippe; Wehrli, Eric; Gaudinat, Arnaud (2001) La synthèse de l'intonation à partir de structures syntaxiques riches. Traitement Automatique des Langues 42 (1), 142-195.
Mertens, Piet (2002) Synthesizing Elaborate Intonation Contours in Text-to-Speech for French in Bernard Bel & Isabelle Marlien (eds.), 2002. Proceedings of the Speech Prosody 2002 conference. 11-13 April 2002. Aix-en-Provence: Laboratoire Parole et Langage. p. 499-502.
Best regards,
Piet
Emmett STRICKLAND
I am using Praat to experiment with pitch manipulation using a .TextGrid file and an associated audio file. The TextGrid segments the file into various linguistic units like syllables and utterances. Some tiers contain textual labels generated by a prosodic labeling model, such as mm (flat mid tone) or lh (low-high rising tone). I would like to produce a resynthesizes version of the original audio file which manipulates the pitch according to these textual labels. The most straightforward way to do this would be to decompose these strings into subunits corresponding to individual pitch points to add to the manipulation object. For example, lh would be decomposed as l (low point at the beginning of the interval) and h (high point at the end of the interval) to produce a rise. Some of these labels contain additional characters, so I must also account for longer strings. For now, I have no idea if it is possible to decompose strings in such a way. Does anyone have some suggestions?