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

[hts-users:00367] my speech


Hi ,
I  have used HTS to generate chinese , I use syllables to construct my HMM models, and  in  the  course of synthesizing,
I found some questions.
1. about the list
 in the  directory of "lists", the  full.list allways has some wrong information.
for example
sil-liu+l/A:.....
iu-lei+sil/A:.....
sil-da+l/A.......a-lian+l/A..........
sil-da+l/A.........
in the example,
between  "sil-da+l/A:.."  and "a-lian+l/A"  , there is no "enter" ,
they are in the same line, but they are different syllable models.
I use the "makefile"  in the "HTS-demo_NIT-ATR503-M001" ,
and make no  modification.

2.some puzzles about f0
I use  tcl/snack to get f0s for HTS. I write a tcl script,
and  set   framelength=0.025 that is for mel  to get f0s.
But ,when I use them to generate speech, the results is too bad.
I find int the course of extracting mel, HTS  needs
$sampfreq    = 16000; 
$framelength = 0.025; 
$frameshift  = 0.005; 
$windowtype  = 0;    
$normtype    = 1;     
$FFTLength   = 512;  
$freqwarp    = 0.42;  
$mceporder   = MCEPORDER;   
but my course of extracting f0 only uses the framelength,
and set it the same as mel'framelength.
I want to know  that
is it necessary to set frameshift and other options for geting f0?
but snack only provides
-method m
-start start
-end end
-framelength t
-windowlength length
-maxpitch val
-minpitch val
-progress callback
 
 
I aso use "pda" to get f0.
when I use them to construct HMMs and synthesize speech,
the speech's quality is also low.

3.about my speech
I guess  it is the f0 that cause my speech's unclear,
so I use "pda" and "tcl/snack" to get f0 respectively,
but I don't get better result.
Any other factors can  effect speech' s articulation?

thank you.
Best regards
 


免费下载 MSN Explorer
Follow-Ups
[hts-users:00368] Re: my speech, Heiga ZEN (Byung Ha CHUN)