Sprints
https://pycon.jp/2015/ja/events/sprint/KaraOke style Read Aloud system (textKaraok)
p00. Renyuan Lyu [仁園(二円) 呂]
p01. Hisahiro Ohmura (尚弘 大村)
p02. Piotr Dyba (Peter)
# Start
----
The prototype system is here. (in Windows, using GoogleTTS)
https://www.dropbox.com/s/ufpw4rvs50pbuff/ryWork07_GttsHalign1006.zip?dl=0
# Major files (source and data)
----
f01. runMe.bat
f02. ryGttsHalign.py (for english only)
f03. ryGttsHalign_3lang.py (for 3 languages, incl english, japanese, tChinese)
f04. _input/en00_wiki.utf8
f05. _input/jp00_wiki.utf8
f06. _input/tc00_wiki.utf8
# Descriptions of major files
--------
d01. some pre-installation are described in runMe.bat
d02. _HTKTools is only for Windows to use
d02.1 for Mac, please go to http://htk.eng.cam.ac.uk/ ,
register, download, and build up the HTK Tools
This may be challenge, take care and good luck.
d02.2 another outside tool is FFMpeg, https://www.ffmpeg.org/,
this is easier.
d03. after installation, you can do runMe.bat
--------
# Meet and talk with
智子 打田 (Uchida Tomoko)
she presented 'janome' @ PyConJP
I got much help from her experience of Japanese natural language processing.
I know MeCab from her janome project.
about the Chinese language, she suggested 'Class SmartChineseAnalyzer' for me to use.
--------
# a 'Ref' to see the final result
--------
a demo on JavaScript is here (for ref in this sprint)
https://dl.dropboxusercontent.com/u/33089565/ryWWW2015/HanHanPrenstation/index.html
# the development process ....
--------
c01. Initially, we meet some obstacles to install it on Mac, since HTK has not been updated for at least 5 years (lastUpdatedYear==2009).
c02. Having difficulties? Search 'install HTK on Mac OS', >>> http://htk.eng.cam.ac.uk/docs/inst-nix.shtml, it may be useful.
HTK on OS X, https://github.com/JoFrhwld/FAVE/wiki/HTK-on-OS-X
Hisahiro get a success from the above git, 御目出度う御座います
--------
c03. Getting mp3 from google seems a prohibit now. !!
--------
c04. Determine to skip it first.
Try to use this link to have some speech for the next stage processing:
https://www.dropbox.com/s/aq7hy3xmuq1onc7/wavDir.zip?dl=0
Successful, except for those Windows system calls.
--------
c05. 尚弘 suggests to use Mac's TTS directly.
Great idea!
Now it is possible that we do not need Internet (or google) to run this program.
--------
c06. another confusing point is the fileSize / timeDuration in the audio file.
Solved !
--------
c07. Determine to give a presentation no matter how....
--------
c03. Getting mp3 from google seems a prohibit now. !!
--------
c04. Determine to skip it first.
Try to use this link to have some speech for the next stage processing:
https://www.dropbox.com/s/aq7hy3xmuq1onc7/wavDir.zip?dl=0
Successful, except for those Windows system calls.
--------
c05. 尚弘 suggests to use Mac's TTS directly.
Great idea!
Now it is possible that we do not need Internet (or google) to run this program.
--------
c06. another confusing point is the fileSize / timeDuration in the audio file.
Solved !
--------
c07. Determine to give a presentation no matter how....
沒有留言:
張貼留言