This　paper　proposes　an　approach　to　improve　both　the　target　speaker＇s　individuality　and　the　quality　of　the　converted　speech　by　preparing　the　training　data.　In　mixture　Gaussian　spectral　mapping　(MGM)　based　voice　conversion,　spectral　features　representations　are　analyzed　to　obtain　the　right　feature　associations　between　the　source　and　target　characteristics.　A　voiced　and　unvoiced　(V/UV)　decision　scheme　for　time-alignment　is　provided　to　obtain　the　right　data　for　training　mixture　Gaussian　spectral　mapping　function　while　removing　the　misaligned　data.　Experiments　are　conducted　in　terms　of　the　applications　of　spectral　representation　methods　and　V/UV　decisions　strategies　to　the　MGM　functions.　When　linear　predictive　cepstral　coefficients　(LPCC)　are　used　for　time-alignment　and　the　V/UV　decisions　are　adopted　for　removing　bad　data,　results　show　that　the　conversion　function　can　get　a　better　accuracy　and　the　proposed　method　can　effectively　improve　the　overall　performance　of　voice　conversion.　©2004　IEEE.