[hts-users:04018] Re: lf0 CMLLR speaker adaptation
- Subject: [hts-users:04018] Re: lf0 CMLLR speaker adaptation
- From: Keiichiro Oura <uratec@xxxxxxxxxxxxxxx>
- Date: Sun, 2 Mar 2014 22:52:12 +0900
- Cc: Keiichiro Oura <uratec@xxxxxxxxxxxx>
- 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:cc:content-type:content-transfer-encoding; bh=94KYMqpv981nmiKeYQMXe8D6TvGtx1foG+TLf01HxxU=; b=IjWHMQXmdFRk0XLt00jQhew6NdvgLEdnae5+6T744saP/kq8weeBVDquZJFmi5XsMP ZqFeNGFWQP7boqJdY90VF2GLN6Y5ve57pQwDI/8TOQaoHli7GvNJ0vlkDMn+ZnTnPkX2 IK0BJOQI7G2HlQh91BHCnO+7GMvIksgwVWtSxCTqWNAfO/mS8ODPh/hc7VqfCj6B3dDM Eu3DE0u+89PmgDsgEimeAu5LcNFE0fH5P6L8akFf1VrjbOt3mcQOWu0ZIGods/0SmVBL kU0A/oBpa7hG2VLx26PmAY7kD6kM3Z3iskGEa69a/gl4sbfQQ2iHwX8qqhcclX0J0UZp a1JQ==
Hi,
Please, check HTKBook-3.4.
The CMLLR formula is written in section 9.4.3.
Regards,
Keiichiro Oura
2014-02-26 22:12 GMT+09:00 최영호 <choiyh@xxxxxxxxxxx>:
> Hello
>
>
>
> I have a question about lf0 CMLLR adaptation.
>
> I have performed CMLLR speaker adaptation. And I have generated transform
> matrix.(xform)
>
> The following is a lf0 transform matrix.(xform)
>
>
>
> <LINXFORM> 2
>
> <VECSIZE> 1
>
> <LOGDET> 5.104003e-02
>
> <BLOCKINFO> 1 1
>
> <BLOCK> 1
>
> <XFORM> 1 1
>
> 1.025848e+00
>
> <LINXFORM> 3
>
> <VECSIZE> 1
>
> <LOGDET> 2.727593e-01
>
> <BLOCKINFO> 1 1
>
> <BLOCK> 1
>
> <XFORM> 1 1
>
> 1.146117e+00
>
> <LINXFORM> 4
>
> <VECSIZE> 1
>
> <LOGDET> 1.373038e+00
>
> <BLOCKINFO> 1 1
>
> <BLOCK> 1
>
> <XFORM> 1 1
>
> 1.986787e+00
>
>
>
> When the transform matrix value(xform value) is used, I wonder how lf0
> transform.
>
> In other words, I want to know the formula about lf0 transform.
>
> If you know how to solve, please let us know.
>
> Thank you.
>
>
- References
-
- [hts-users:04017] lf0 CMLLR speaker adaptation, 최영호