chengmic99.github.io

Hello, I'm Michelle Cheng

🤠 I'm a second year student at the University of Texas at Austin
🌵 Therefore, I'm based in Austin, Texas
👾 I'm studying computer science and pursuing a business minor
🎶 I'm passionate about music and arts & tech
💌 Email me @ chengmic99@gmail.com
PROJECTS

Language Transcender

Inspiration

As an avid consumer of stuff that isn't in my first language (anime, k-pop, etc), I appreciate fluent speakers who subtitle episodes/songs. Because of them, I've seen that humanity, hardship, and other beautiful things that flow through anime and kpop really transcend language. Since human connection transcends language, so too should communication, one of its many vessels.

What it does?

Language Transcender is a video chat room website that subtitles conversations in real time. You first choose your subtitle language and then create a video chat room. The room has a shareable link which you can then share with your foreign friends. You can all have face to face conversations as if you all spoke the same language, thanks to the subtitles!

How did I make it?

Backend: NodeJS
Frontend: HTML, CSS, JS, Bootstrap, JQuery
Other: I used IBM's Watson's speech to text API to convert speech into text and its language translating API to then translate the text into the language the user chose. I used Okta to handle signalling and SimpleWebRTC to handle the media streams.

Challenges?

I'd never used Node or done web dev before this project, so it took a lot of reading to understand the concepts. In addition, since I worked on this project over the summer, the only motivation I had to work on it was my own intrinsic motivation. I had many periods of discouragement, but after taking week long breaks I immediately solved problems that I had previously spent weeks on. Sometimes you just need a break.

What did I learn?

I learned how to use NodeJS things such as event driven programming, callback functions, npm, modules, servers, token authentication, express, and more. I gained more knowledge on how to use APIs through Watson. Most of all, I learned that if you work hard enough (google hard enough), there is an answer to everything. Also, if something else will work better, it's worth it to start all over again sometimes. Don't be afraid to rebuild from scratch.

Now what?

Recently, SimpleWebRTC, the video chat framework I was using, was taken off the web. So I need to learn how to use WebRTC instead. As of now, the subtitles only show up in the language of the room creator. I need to learn how to make the webpage experience unique to each user.
In the future, I think it would be cool to turn this into a social media type of format too, so that social media influencers can livestream to audiences of all languages.

Speaker Detection Through Audiovisual Cues

Inspiration

What it does?

How did I make it?

Challenges?

What did I learn?

Now what?

The Happ

OTHER INTERESTS

Thanks for checking me out!