Voice conversion using structured Gaussian mixture model in cepstrum eigenspace

A new methodology of voice conversion in cepstrum eigenspace based on structured Gaussian mixture model is proposed for non-parallel corpora without joint training. For each speaker, the cepstrum features of speech are extracted, and mapped to the eigenspace which is formed by eigenvectors of its scatter matrix, thereby the Structured Gaussian Mixture Model in the EigenSpace （SGMM-ES） is trained. The source and target speaker＇s SGMM-ES are matched based on Acoustic Universal Structure （AUS） principle to achieve spectrum transform function. Experimental results show the speaker identification rate of conversion speech achieves 95.25%, and the value of average cepstrum distortion is 1.25 which is 0.8% and 7.3% higher than the performance of SGMM method respectively. ABX and MOS evaluations indicate the conversion performance is quite close to the traditional method under the parallel corpora condition. The results show the eigenspace based structured Gaussian mixture model for voice conversion under the non-parallel corpora is effective.

Participatory approach to design social accountability interventions to improve maternal health services:a case study from the Democratic Republic of the Congo

Background:Social accountability(SA)comprises a set of mechanisms aiming to,on the one hand,enable users to raise their concerns about the health services provided to them(voice),and to hold health providers(HPs)accountable for actions and decisions related to the health service provision.On the other hand,they aim to facilitate HPs to take into account users’needs and expectations in providing care.This article describes the development of a SA intervention that aims to improve health services responsiveness in two health zones in the Democratic Republic of the Congo.Methods:Beneficiaries including men,women,community health workers(CHWs),representatives of the health sector and local authorities were purposively selected and involved in an advisory process using the Dialogue Model in the two health zones:(1)Eight focus group discussions(FGDs)were organized separately during consultation aimed at sharing and discussing results from the situation analysis,and collecting suggestions for improvement,(2)Representatives of participants in previous FGDs were involved in dialogue meetings for prioritizing and integrating suggestions from FGDs,and(3)the integrated suggestions were discussed by research partners and set as intervention components.All the processes were audio-taped,transcribed and analysed using inductive content analysis.Results:Overall there were 121 participants involved in the process,51 were female.They provided 48 suggestions.Their suggestions were integrated into six intervention components during dialogue meetings:(1)use CHWs and a health committee for collecting and transmitting community concerns about health services,(2)build the capacity of the community in terms of knowledge and information,(3)involve community leaders through dialogue meetings,(4)improve the attitude of HPs towards voice and the management of voice at health facility level,(5)involve the health service supervisors in community participation and;(6)use other existing interventions.These components were then articulated into three intervention components during programming to:create a formal voice system,introduce dialogue meetings improving enforceability and answerability,and enhance the health providers’responsiveness.Conclusions:The use of the Dialogue Model,a participatory process,allowed beneficiaries to be involved with other community stakeholders having different perspectives and types of knowledge in an advisory process and to articulate their suggestions on a combination of SA intervention components,specific for the two health zones contexts.