New community and activities of NVDA in Japan http://t.co/cfKNKIIc #nvdajp #nvdasr
posted at 10:43:20
| Stats | Twitter歴 5,214日（2009/02/22より） |
ツイート数 3,632（0.6件/日） |
表示するツイート :
2012年05月28日(月)14 tweetssource
@jcsteh In Chinese, a character corresponds to a braille code. In Japan, Kantenji is such a system, however, very few people use it. #nvdajp
posted at 11:40:59
@jcsteh To support Japanese braille code with liblouis, tons of rules must be developed and maintained. #nvdajp
posted at 11:56:42
@jcsteh I don't say it is impossible. But the text processing engine is high-cost development. #nvdajp
posted at 12:00:50
@jcsteh For Japanese language, text processing for TTS is done by pattern recognition or machine learning approach. #nvdajp
posted at 12:04:07
@jcsteh To keep both Japanese TTS and braille engine maintained, sharing language processor is one of our choice. #nvdajp
posted at 12:07:06
@jcsteh Japanese braille corresponds to pronunciation of words. Set of written characters corresponds to the individual word. #nvdajp
posted at 13:49:11
@jcsteh A single character in Japanese may have different pronunciations depending on the word which contains the letter. #nvdajp
posted at 13:53:49
@jcsteh Japanese for sighted person does not use spaces between words. But Japanese braille code uses spaces between words. #nvdajp
posted at 13:57:12
@jcsteh So Japanese braille translation must separate words and give pronunciations for each words, considering the vocabulary. #nvdajp
posted at 14:01:31
@jcsteh Japanese braille code has many conventions on representing word breaks and pronunciations. Something similar to TTS. #nvdajp
posted at 14:07:26
@jcsteh We are also studying louis. It is well designed, but to obtain good quality is very difficult for Japanese. #nvdajp
posted at 14:13:01
@tspivey Thank you!!
posted at 17:27:53
@tspivey updated JTalk addon for NVDA http://t.co/8KZcQQgw #nvdajp #nvdasr
posted at 17:55:31