[Subject Prev][Subject Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:03745] Re: HTS-patched HVite


HTS-patched HVite is not so different from the HTK one; its additional functionality is 1) can handle stream-level parameter sharing structure which is introduced in HTS, 2) stream-level caching of computed log likelihoods in addition to state and mixture-level.  So if HTK can do what you want, HTS can do it too.  If HTK cannot do it, HTS cannot do it neither.

Heiga
  


2013/4/30 Georg Schlunz <gschlunz@xxxxxxxxxx>

Hi


Does the HTS patch to HVite allow you to perform recognition with it using bigger (read here: customized beyond triphone) context labels? (Of course, once you have trained these bigger context models on appropriate speech)


If so, can it do only free phone recognition (i.e. a string of acoustically most probable labels), or does it allow you to incorporate some enhanced form of language model as well?


Thank you

Georg


--
This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard.
The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html.


This message has been scanned for viruses and dangerous content by MailScanner,
and is believed to be clean.


Please consider the environment before printing this email.




--
Heiga ZEN (in Japanese)
Byung Ha CHUN (in Korean)
<heigazen@xxxxxxxxxx>

Follow-Ups
[hts-users:03863] MLPG Matlab, David Klein
References
[hts-users:03744] HTS-patched HVite, Georg Schlunz