[Subject Prev][Subject Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:02812] Re: hello,friends


by the way,兴宇,I have found that the mono/full .lab files are extracted from Utterance files.So I think what I have to do firstly is that creating utt files,However ,I can only create utt files with festival method.Not my segmentation results.Could you give me some advice?

在 2011年4月28日 下午4:28,那兴宇 <nxy-yzqs@xxxxxxx>写道:
If you have mono labels of your own, you have to prepare the corresponding full labels. If you want to use the same format as the demo context format, you just have to adjust the time information in full labels.
When above is done, make all data except label(ie, mgc/lf0/gv/mlf/scp/list) and then make voice, then you will get your own voice.
HTS itself does not care the segmentation method you use. Just make sure the mono/full labels and raws are compatible.

At 2011-04-28 15:10:17,"zhinan zhang" <zhangzhinan1987@xxxxxxxxx> wrote:
I have read the makefile in the HTS demo and I found that the *.lab files in the mono folder are those that were extracted form utterance files.So my question is that if I put the segmentation label files(segmented by sphinx before)  in the mono folder and delete all the orignal files .Can HTS run the demo with the phoneme segmentation of my method?



体验网易邮箱2G超大附件,轻松发优质大电影、大照片,提速3倍!


References
[hts-users:02809] hello,friends, zhinan zhang
[hts-users:02810] Re: hello,friends, 那兴宇