Speech Recognition with SAPI: Custom Language Support through phonemes

自古美人都是妖i 提交于 2019-12-04 16:13:39

It very strongly depends on what exactly you want to do. You can (more or less) do this in a command & control grammar; doing this in a dictation environment is a lot trickier (and probably not worthwhile).

First some (more) limitations: The English SAPI recognizer is looking for English phonemes. Giving it (e.g.) Italian phonemes can throw off the recognition model, so you're most likely going to get lower recognition quality.

That being said, however, you can include explicit pronunciations in a SAPI grammar using the PRON attribute of the PHRASE element; so that can work in a command & control grammar.

In a dictation environment, however, you're probably limited to setting up an application lexicon and using ISpLexicon::AddPronunciation to build entries with explicit phone IDs.

Again, however, your accuracy may well suffer greatly.

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!