Hi Keiichiro
I did few modifications (recreated the raw files and rather than using 16kHz, I changed the sampling to 48Khz) and it worked okay. I don't know the reason why it was failing before. The INSTALL says the following versions (HTS-demo_CMU-ARCTIC-SLT requires Festival, SPTK-3.6, HTS-2.3 alpha, and hts_engine API-1.07) and those are what I use.
Right now, everything builds and I was able to generate the hts voice - but the audio doesn't sound like the original speaker after synthesis. The words are intelligeble - but the TTS is too buzzy.
I just used default parameters (./configure with only festival/SPTK paths specified). Are there any parameters I could change to improve closeness to speaker ?
Regards
Karthik
On Thursday, November 14, 2013 3:50 AM, Keiichiro Oura <uratec@xxxxxxxxxxxxxxx> wrote:
Hi,
What version of the engine do you use?
The version should be used is
defined in HTS-demo_CMU-ARCTIC-SLT/INSTALL.
Regards,
Keiichiro Oura
2013/11/14 Noah Kityawong <
noahgundam@xxxxxxxxx>:
> help
>
> I'm newbie in TTS.
> I try to run demo which I download from
http://hts.sp.nitech.ac.jp/?Download> I have some problem about HTS-engine that system said Error: HTS voices
> cannot be loaded.
> So I try to check step by step and the problem is " HTS-engine cannot read
> dur.pdf" (All file have been created during process before ).
> Could you mind to show me how to solve this problem??
>
> Best
> Noah