[hts-users:04504] Re: Methods for obtain better alignment
- Subject: [hts-users:04504] Re: Methods for obtain better alignment
- From: Nickolay V.Shmyrev <nshmyrev@xxxxxxxxx>
- Date: Wed, 15 Mar 2017 23:38:05 +0300
- Authentication-results: mailgw.mains.nitech.ac.jp; dkim=pass (1024-bit key) header.d=yandex.ru firstname.lastname@example.org header.b=sDSaR3Up
- Authentication-results: mxback1m.mail.yandex.net; dkim=pass email@example.com
- Delivered-to: hts-users@xxxxxxxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yandex.ru; s=mail; t=1489610285; bh=07YW0xE14F/JCnuVrNnC2ucdy4iTUZD9FlUV0cClAPQ=; h=From:To:In-Reply-To:References:Subject:Message-Id:Date; b=sDSaR3Up5tcxoR93E4Ot+3FHSHMDAlN+8uTV6a/g+onC+XfzU2HBVGeJtEIYPk4DV kTv12L33/KufyR+ZqrNQnEw5OHTGQh8SumINLFcHT/u8oMV7u1PD2CJfq/aOtk+eZB gYHXimFTY2lPsBTTnXlx0+WjOuEv4GU6PeEByieM=
This method is pretty innovative
it is probably not easy to use, but implementation is available in festvox/clustergen/do_move_label script, you can try it.
15.03.2017, 22:16, "Franco Negri" <franconegri2004@xxxxxxxxx>:
> Hi again,
> I'm currently trying to improve the alignment of my lab files and its
> corresponding audio.
> I have a reliable phone transcription I would like to use instead of
> one generated purely by EHMM, so I thought it could be a good idea to
> use EHMM on the reliable transcriptions to obtain full context labs.
> Have you tried something like this? It's there other ways of improving
> the end result?
- [hts-users:04503] Methods for obtain better alignment, Franco Negri