[hts-users:04163] Re: Error during making global variance step
- Subject: [hts-users:04163] Re: Error during making global variance step
- From: Keiichiro Oura <uratec@xxxxxxxxxxxxxxx>
- Date: Thu, 13 Nov 2014 19:24:29 +0900
- Delivered-to: hts-users@xxxxxxxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:sender:in-reply-to:references:date:message-id:subject :from:to:content-type; bh=dlt04SB0dC7oZul2Q3EH0PQ7DpRQka5dWW8dvJMBzKQ=; b=ATKa/04xv+1ZIGST7Gmcp8BOIEFgrowcITJUcUZALp1arxsasO1uk+PFcH8XtjFwOY 8u0Bfkssvj+YuMwM2Vw4EfUMryCK2Y3xJL1TpfKTf/tBt3w+aSeArsv3g+E+OQjs2tPn G4QlrEylo4O/zyJ2C4wS4GlnmHiL93kOBokM4XbO5+yBFyuFb+1DbRL5Sb7rN4xptBcx BtlkNVtaIZc2kdgrfXjd/w/RQiAsB25FcBKIAUm9kuEmrjW4A27dalVNEHWQJR6jdDfK 50DWsPRfH9nTffjR4BKdaYVwlotofiQo+fFXXsgzEiqSUWO5m70lKo1WBqWhm4tH4WZs X4pw==
Hi,
The option for beam width of HSMMAlign is "-t".
The default beam width of HTS demo described in scripts/Config.pm is
"1500 100 5000".
Could you increase it to "3000 200 10000" ?
Regards,
Keiichiro Oura
2014-10-21 20:29 GMT+09:00 Max Muljono <muljono@xxxxxxxxxxxxxxx>:
> Dear Prof. Keiichiro Oura
>
> I am sorry to bother you.
> Regarding to your advice for me about my question in hts-users-millist
> (error during making global variance step)
> I changed script trainning. pl
> $HSMMAlign = "$HSMMALIGN -A -C $cfg{'trn'} -D -T 1 -S $scp{'trn'} -I
> $mlf{'mon'} -t $beam -w 1.0 ";
> I already increase beam width for HSMMAlign to be 1.25, 1.5, or 2.0.
> (default 1.0). However, the same error occured again. Could you give me
> solution this problem?
> Your help or your guidance would be extremely appreciated.
> Thank you very much.
>
> Best regards,
> Mulyono
>
> ---------- Forwarded message ----------
> From: Keiichiro Oura <uratec@xxxxxxxxxxxxxxx>
> Date: 2014-08-27 10:42 GMT+09:00
> Subject: [hts-users:04103] Re: Error during making global variance step
> To: hts-users <hts-users@xxxxxxxxxxxxxxx>
> Cc: Keiichiro Oura <uratec@xxxxxxxxxxxxxxx>
>
>
> Hi,
>
> I suppose that forced-alignment step is a failure.
> Could you increase beam width for HSMMAlign ?
>
> Regards,
> Keiichiro Oura
>
>
> 2014-08-14 21:28 GMT+09:00 Max Muljono <muljono@xxxxxxxxxxxxxxx>:
>> Hi,
>>
>> I try to create speech synthesis for my language.
>> I changed HTS-demo-CMU-ARCTIC-SLT using speech, utts, question, label for
>> gen according to my language..
>> And then, I run make. During making global variance step, there is error.
>>
>> If I use 100 speech data (raw and utts), the error as below :
>> ...
>> Making data, labels, and scp from cmu_us_arctic_slt_a0029.lab for
>> GV...done
>> Making data, labels, and scp from cmu_us_arctic_slt_a0030.lab for
>> GV...Cannot open No such file or directory at scripts/Training.pl line
>> 1214,
>> <SCP> line 30.
>>
>> if I use 200 speech data (raw and utts), the error as below :
>> ...
>> Making data, labels, and scp from cmu_us_arctic_slt_a0111.lab for
>> GV...done
>> Making data, labels, and scp from cmu_us_arctic_slt_a0112.lab for
>> GV...Cannot open No such file or directory at scripts/Training.pl line
>> 1214,
>> <SCP> line 112.
>>
>> if I use 1529 speech data (raw and utts), the error as below :
>> ...
>> Making data, labels, and scp from cmu_us_arctic_slt_a1181.lab for
>> GV...done
>> Making data, labels, and scp from cmu_us_arctic_slt_a1182.lab for
>> GV...Cannot open No such file or directory at scripts/Training.pl line
>> 1214,
>> <SCP> line 1182.
>>
>> if I use original HTS-demo-CMU-ARCTIC-SLT (English Language) and I run
>> make.
>> There is no problem and no error until synthesizing waveforms finished.
>>
>> Give me a solution for this case?
>> Thank you very much for your help.
>>
>> Best regards,
>> Mulyono
>
>
- Follow-Ups
-
- [hts-users:04164] Re: Error during making global variance step, Max Muljono
- References
-
- [hts-users:04095] Error during making global variance step, Max Muljono
- [hts-users:04103] Re: Error during making global variance step, Keiichiro Oura