[hts-users:02754] Re: HTS-demo_CMU-ARCTIC-SLT and hts_engine

[Ramil <ramil2319@xxxxxxxxx>]

thanks for reply.
i found the problem. problem was hts_engine parameter all-pas. so atof function doesn't worked correctly in visual studio 2008. all-pas paremeter must be 0.42 but in visual studio atof function made it 0.000

2011/3/30 Mehmet Uğur Doğan <mugur@xxxxxxxxxxxxxxxxxxxx>

Hi Ramil,
Congrats for your success. The problem seems as if a  frequency resampling  occur somehow. Can you check that the second PC replays your synthesized wav file at correct frequency(i.e. 16kHz file plays at 16kHz not 32kHz or 48 kHz)? Somehow the header of your riff file is not in tact? Hope helps

Regards,
M Ugur


30.03.2011 08:32, Ramil yazmış:

Hello everyone,
i created HTS-demo_CMU-ARCTIC-SLT speakerdependent models successfully for Azerbaijani. i am using hts_engine synthesizin waveforms. and i created context-dependent labels format successfully too. i compiled hts_engine in visual studio 2008. now Text To Speech for Azerbaijani sounds well.
but i have some problems in tts speaking. Problem is in wave quality. Same application is working perfectly in my pc but another pc the wave is very bad (faster and like child voice). I dont know why it happens actually.
Please, can anyone help me to correct this problem?

--
Best Regards,
        Ramil Mammadov

-- 
Best Regards,
        Ramil Mammadov