The Bing Speech API has been a real challenge. After I downloaded it from the github repository, I was puzzled what to do with it. The files were in .ts extension, which I later learned after some research online that they stand for Typescript, an open-source programming language developed and maintained by Microsoft.
After some additional research, I learned that Typescript can be compiled to JavaScript by using some command line arguments, but both Node.js and npm were needed for the process.
– I downloaded Node.js and npm. Both very new to me.
– Downloaded the Sample page to work in the browser, compiled the Typescript files, and exported to the web.
– After pushing it to web, the sample page didn’t work. I solved it by manually pushing each single folder, one at a time. I guess some files were missing at first.
P.s: I still need to go back and make sure all the files are necessary. As I am still understanding how the Speech API works and what dependencies are needed. I will clean the web directory later, if performance seems compromised.
– Then the voice didn’t work. The microphone was being blocked. I later learned that in order for the microphone to work I had to type https:// in front of the domain, as although I had installed the SSL, it wasn’t redirecting automatically to it. I then added a 301 redirect to always ensure the users will be accessing the https:// version of the site automatically.
Now the sample page (speech recognition) is FINALLY working! Yay! Now, on to the next challenges:
1. I need of understand how the speech recognition really works, where are the words captured stored (or where/how should I store them) and what components/files I will need to change.
2. After that I will need to do learn the “Text to Speech” part for the book narration and sync with caption/subtitles/words.