Text-To-Hymn Synthesis for St Yared Hymn Notations
| dc.contributor.advisor | Assabie, Yaregal (PhD) | |
| dc.contributor.author | Zemedu, Girma | |
| dc.date.accessioned | 2018-06-20T07:08:01Z | |
| dc.date.accessioned | 2023-11-04T12:22:23Z | |
| dc.date.available | 2018-06-20T07:08:01Z | |
| dc.date.available | 2023-11-04T12:22:23Z | |
| dc.date.issued | 2014-03 | |
| dc.description.abstract | Speech synthesis, a process of making artificial speech, has been one of the research areas for the last many decades. In addition to this singing voice synthesis has also started and a lot has been done internationally for western songs. But nothing has been done for the St Yared hymn notation which is used for nearly 1500 in most traditional music and inside Ethiopian Orthodox church. On the other hand, the number of scholars decreased dramatically for different reason and it is scared to be lost in the future. Thus, to preserve these notations text-to-speech synthesis for St Yared hymn notation is required. To this effect, we used a unit selection concatenative approach to synthesize the notations. Initially, different hymn lyrics were collected and transcribed properly by marking the notes. After that voice of an expert of the hymn was recorded for the transcribed lyrics. Segmentation of the note from the recorded voice is done manually and saved in a dataset with its context information of the left and right for target note. In the synthesis phase the selected unit waves are synthesized by making certain changes on the pitch and intensity to smoothly concatenate the units. A prototype was developed to implement this model. The system was evaluated using a test data collected from various hymn lyrics. These lyrics were tested with a simple concatenation of notes without note markers and with a note marker. Mean Opinion Score (MOS) for both Naturalness and intelligibility, which was taken after an assessment made by a number of users, shows a promising result. For naturalness and intelligibility of unmarked notes we achieved a performance result of 2.3 and 2.6 respectively. For naturalness and intelligibility of marked notes we achieved a performance result of 3.9 and 3.4 respectively Key words, St Yared note, concatenative speech synthesis, St Yared Hymn synthesis | en_US |
| dc.identifier.uri | http://etd.aau.edu.et/handle/123456789/2030 | |
| dc.language.iso | en | en_US |
| dc.publisher | Addis Ababa University | en_US |
| dc.subject | St Yared Note; Concatenative Speech Synthesis; St Yared Hymn Synthesis | en_US |
| dc.title | Text-To-Hymn Synthesis for St Yared Hymn Notations | en_US |
| dc.type | Thesis | en_US |