search

Marc-Bogonovich / Openwords

This is our first Openwords repository on GitHub
0 stars 0 forks source link

Hearing module - IPA transcription #70

Open Marc-Bogonovich opened 9 years ago

Marc-Bogonovich commented 9 years ago

This is not a pressing issue.

The first clear purpose of the IPA (international phonetic alphabet) seems to be within the hearing module.

Homonyms are words with exactly the same pronunciation (including tone). If a learner has two words within their corpus which are mutual homonymns, then they would not be able to distinguish them on the hearing learning module.

We must develop a front end solution and corresponding back-end solution to this problem.

Marc-Bogonovich commented 9 years ago

Being a regular user of Openwords, I am starting to notice that IPA may have a higher priority than I previously considered.

For example, I am repeatedly getting the following two words incorrect, which keeps them at a high frequency in the word selection algorithm. I am getting them incorrect because they are identical. 十 -- shí 石 -- shí

Thus, this problem has tendency to bubble to the surface.

A separate issue is the following. Some words have a very similar pronunciation e.g. 是 (has a different tone). Or 吃 (slightly different first consonant). Having used the hearing module, for a little while, I can say that the minimal pair LM would be extremely valuable. I may create an issue for this LM. We can discuss the issue at our next technical meeting (2014-10-04), and our next design meeting (2014-10-08).