I am using library OKHttpUtils2 to retrieve HTML from a web page. I parse out some text from the page and feed it to the text-to-speech engine. A problem arises when the text contains Unicode decimal codes where the engine says "hash" and the number. For now when I find this happening I replace the code to something similar in ASCII or change it to its meaning. Of course I will not be able to get everything so I was wondering if there is anyway to address this problem another way. I know that it would be virtually impossible to replace characters to something that makes sense. For instance I replace the Unicode ¼ which is the 1/4 symbol to " one fourth ". Any advice with this issue would be appreciated.