[hts-users:03491] Re: Log-durations
- Subject: [hts-users:03491] Re: Log-durations
- From: "Heiga ZEN (Byung Ha CHUN)" <heigazen@xxxxxxxxxx>
- Date: Tue, 27 Nov 2012 10:19:35 +0000
- Delivered-to: hts-users@xxxxxxxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=jLMUYg8HqprbETCVjLlvsaClHeScShuUQNTPJVB3uB4=; b=dNYbQkP/KlYhqJZlHz/OrmyD7YPgcRBwtUX2qEuOwDVZsM2+6cGuzMcD5180O/6s1f WAqk8+K0UT7iBokoWgaxXv0eIzHlX3jCrjf6QOTL2f69EySLI3ILTFfGrebDQKJ+92te lquxgT4HGyA3iLDXlTQSn7EsEMlBAdqPBr5woKq3cnrNV2Lqo8BH3kGDD9pNVW6FZqWH MpoGWS2Npg5gTYITGAj2U7eY+mQAlG1tx6uQs6z9bkU4dhxUx7tb+Cb7It3Skz/dsP1S 5u8AJfJmtPCiCRnLu666DjHzUVfskM79/xiumiBigeILwoXlluItTDcD38ZLCptCQu7G 7WwQ==
Hi,
There is a paper about this topic:
http://jairo.nii.ac.jp/0083/00140243
Heiga
2012/11/27 Daniel Erro <daniel.erro@xxxxxx>:
> Dear HTS developers,
>
> Would not it be more logical to model the logarithm of the durations rather
> than the durations? I expect log-durations to be more compatible with
> Gaussian modeling and I guess this would solve many problems arising when
> durations are forced by the user. Has anybody tried this in the past?
> Sorry if I missed a previous thread on this topic.
>
> Daniel Erro
>
> Ikerbasque Research Fellow @ Aholab, UPV/EHU, Bilbao, Spain
>
>
>
--
Heiga ZEN (in Japanese)
Byung Ha CHUN (in Korean)
<heigazen@xxxxxxxxxx>
- References
-
- [hts-users:03490] Log-durations, Daniel Erro