Pengaruh Penting Keempat Parameter Dalam Metode LILJENCTRANCTS FANT (LF) Pada Transformasi Wicara (SPEECH) Berbasis Turunan Gelombang Glottal

Sriwijanaka Yudi Hartono; Saidah  Suyuti

Sriwijanaka Yudi Hartono Staf Pengajar, Teknik Telekomunikasi Program Studi Teknik Elektro Fakultas Teknik Universitas Muslim Indonesia
Saidah Suyuti Staf Pengajar, Teknik Telekomunikasi Program Studi Teknik Elektro Fakultas Teknik Universitas Muslim Indonesia

Keywords: transformation, speech, formant

Abstract

The accuracy of the results of speech signal conversion (Speech) is carried out for the synthesis of voice conversion into various forms of sound carried out using SQ, OQ, Formant and Pitch parameters. Voice conversion is done with the help of Tcl-based waveurfer snack software. The conversion process is done by changing the color of the sound without changing the sound quality and sound duration at a sampling frequency of 12KHz. The method used is Liljenctrants Fant (LF), all four parameters are set to match the sound color or sound pattern you want to convert with a human voice database with one speaker, for example conversion of female voices, children's voices and elderly voices and voices in the form of robotic sounds . The SQ and OQ parameter values are in the range between 0.8 - 1.9 while the Formant and Pitch parameter values are 0.8 - 1.5. Depending on the color of the sound you want or need. This research is expected to be developed as a basis for sound synthesis, manufacturing of telecommunications information equipment, dubbing equipment in filmmaking or talking robotic.