[hts-users:02420] Re: What is the usage of a cascade of CMLLR transforms?
- Subject: [hts-users:02420] Re: What is the usage of a cascade of CMLLR transforms?
- From: Hui LIANG <tshlmail-hts@xxxxxxxxx>
- Date: Fri, 19 Mar 2010 08:06:11 -0700 (PDT)
- Delivered-to: hts-users@xxxxxxxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1269011171; bh=VJifD3zStXMwdYFCi+EH6G2JDzeKbRyBeUHfbdDsULE=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Reply-To:Subject:To:MIME-Version:Content-Type; b=LVIyxUPG8WwAzasdI/rlmUs7EQ5DyemTq2vuMXJ7uKdbPYkHxZrGSERjo9ssq+P8QVIs/FTGN+w4Sk03D0QQfLCx9yoPedE4L44xTxz7vMh+v7GYYoSgEnlhFLvIVkfjlqf60cW9Sh3fSr1GPZEBKCBx+0qzLpffhpI2VlREA80=
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Reply-To:Subject:To:MIME-Version:Content-Type; b=hY6nFoZyw3ns4okPONFc7fJidOouS7LcQAl03aNiKxY7Dr/w0LiZLx8V3yCHVCEKKWY0Ywg7HAwg1PJnDXltrwovMakxfP4oxTZ8qBA0Pu/CELDpY0WfW5Yo1dHLbVEfyuf5xQM+m/Yt9v4SibqWR6jn2BFnDZzg9B5Rni2hw6g=;
Thanks for your quick reply! I enclose herewith the log file of my second iteration.
My understanding is that parent transforms are applied to features (because of CMLLR) first and that input transforms are then estimated over average voice models and the adjusted features. Hope the understanding is correct as my experiments were based on this assumption.
Thank you very much again,
Hui LIANG
--- On Thu, 18/3/10, Heiga ZEN (Byung Ha CHUN) <heiga.zen@xxxxxxxxxxxxxxxxx> wrote:
> Please show us the command line which you used to estimate
> CMLLR transforms with parent transforms, etc.
> Otherwise we cannot say whether you run it correctly or
> not.
Attachment:
spkr-csmaplr-2.log
Description: Binary data