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

[hts-users:02822] Re: "Failed to load header of pdfs" while using the demo script to obtain a new voice


Hello,

You were correct, solved it now. The version of HTS in MARY requires v1.01 of hts_engine API while I was using v1.04.

Thanks very much for your help.

Best,

Emre

On 05/02/2011 06:57 AM, Keiichiro Oura wrote:
Hi,

Please, check version of HTS, SPTK, and hts_engine API.

Regards,
Keiichiro Oura

2011/5/2 Emre Çelikten<emre.celikten.17th@xxxxxxxxx>:
Hello,

I hope this is the correct place.

I am using MARY TTS system to build an HMM voice using HTS. MARY uses the
standard HTS demo script to create a voice. The process is detailed here:
http://mary.opendfki.de/wiki/HMMVoiceCreation

Everything goes fine until "synthesizing waveforms" step, which says:


Step: Start synthesizing waveforms using hts_engine at Thu Apr 28 02:58:33
EEST 2011
ERR>
ERR>  Error: HTS_Model_load_pdf: Failed to load header of pdfs.
ERR>  Error in /usr/local/bin/hts_engine -td
/media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/tree-dur.inf
-tf
/media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/tree-lf0.inf
-tm
/media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/tree-mgc.inf
-md /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/dur.pdf
-mf /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/lf0.pdf
-mm /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/mgc.pdf
-dm /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/mgc.win1
-dm /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/mgc.win2
-dm /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/mgc.win3
-df /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/lf0.win1
-df /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/lf0.win2
-df /media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/lf0.win3
-s 44100 -p 80 -a 0.42 -g 0  -b 0.4 -cm
/media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/gv-mgc.pdf
-cf
/media/60C83AA4C83A77FC/thesis/database/hts/voices/qst001/ver1/gv-lf0.pdf -b
0.0  -or
/media/60C83AA4C83A77FC/thesis/database/hts/gen/qst001/ver1/hts_engine/gen_00000000.raw
-ot
/media/60C83AA4C83A77FC/thesis/database/hts/gen/qst001/ver1/hts_engine/gen_00000000.trace
/media/60C83AA4C83A77FC/thesis/database/hts/data/labels/gen/gen_00000000.lab



All of the files specified in parameters until "-or" exist and are
non-empty. I could not find a solution to the problem.

Could you possibly help me? What am I doing wrong?

Thanks in advance,

Emre




References
[hts-users:02814] "Failed to load header of pdfs" while using the demo script to obtain a new voice, Emre Çelikten
[hts-users:02815] Re: "Failed to load header of pdfs" while using the demo script to obtain a new voice, Keiichiro Oura