Hi,
Thank you for your reply.
As you said in the previous mail i have generated the hts_engine format voices,
here i have used STRAIGHT Demo scripts, HTS-2.2, hts_engine-1.06
once speech files used available in STRAIT-Demo tar file and once my recordings, up sampled to 48kHz from 16kHz.
i didn't changed the parameters like sampling frequency, frame shift, MGC order etc. given in demo scripts, while Training, because i used the 48kHz speech files.
in both the cases synthesized speech files in gen folder are some what good in the quality.
when i am using the these models with festival, it was generating speech file but all the samples in file are zeros (nothing to sound),
here i have changed the sampling frequency and frame shift at [line 125 in festvox_cmu_us_slt_arctic_hts/festival/lib/voices/us/cmu_us_slt_arctic_hts/festvox/cmu_us_slt_arctic_hts.scm] while using the generated voice models as below
'("-s" 48000.0)
'("-p" 240.0)
'("-a" 0.45)
'("-g" 0.0)
'("-b" 0.0)
'("-u" 0.5)
please suggest me, how should i get the voices with festival