Inspiration

Inside joke with friends. We always like to make fun of each other and comment on it as an L. One day we decided having an app for this that would say "L Detected" would be really funny but we never got around to making it until this hackathon.

What it does

You press a button and it plays on the speaker using a comedic voice "L Detected" to any situation that is classified as an L. It's purpose is to further augment the "savagery" of the remark and add some comedic relief to the situation. It's an incredibly useful tool to use with friends and can also help kindle new friendships through mutually enjoyable situations and insults.

How we built it

We built it using Swift for iOS. Some of the frameworks we used were AVFoundation which contained most of the speech synthesizer and audio code.

Challenges we ran into

Figuring out how to prevent repetitive tapping on the buttons and also how to make it support multiple screen sizes. It was also hard to get the speech synthesis to work on iOS and Android. Android proved to be especially difficult with the XML code as we were new to it.

Accomplishments that we're proud of

We built a quick hack that did what we wanted completely. We came to BitCamp with no intention to build anything but after getting inspired by all the talks and people around us, we decided to for the spirit of a hackathon. Our idea was essentially stupid but we are happy we put something together that was functional.

What we learned

We learned a lot about the text to speech synthesis and how it works. We also learned how to support multiple screen sizes in android and code an android app from scratch. Another great part of the app was that we learned how to communicate better with teammates and decide on what features are necessary.

What's next for L-Detector

We want to build a classifier where you can input a situation that just occurred and it would be able to classify it as an L or not. I think this would be especially hard just due to the fact that you need training data on this and it would be incredibly hard to produce. A simple sentiment analysis with a negative sentiment may be all we need though which is simple.

Built With

Share this project:

Updates