SlackR

Website Home Page

Inspiration

Many student organization or job meetings online tend to have a lot of time when many of us do not participate at all. During those times, we are both expected to stay in the meeting and to not respond too much. Such times are too precious to be used just to listen in, but currently, many of us do listen to the meetings and attempt to multitask while nervously checking when it is our time to speak. Hence, we decided to create a support system that can both let you focus on something other work or activities during the online meeting when it is not your time to participate and also let you still contribute well in the meetings when notified.

What it does

We developed a website that connects the voice call to a program that transcribes the voice call and summarizes the transcription. The program also has a feature that rings a sound of a bell when someone in the call calls the user's name so that the user could not only get notified that it is his/her time to speak but also see the short summaries of transcriptions to get a quick snapshot of the meeting progress before participating.

How we built it

We used Microsoft Azure API for speech recognition and Firestore to store the speech data. For the website development, we used React.

Challenges we ran into

Audio data is extremely hard to work with, especially because of its complex data type. Our initial approach to this project was to store the voice data as binary data and then pass it into the Microsoft Azure to get the text transcription of binary data. However, after many trials, we learned that using actual voice data as the input to Azure works as well. Thus, we were successful at implementing Azure and using audio data as we designed.

Accomplishments that we're proud of

This project gave us the opportunity to learn various sides of software engineering in a collaborative setting in a short time. This includes not only the audio-related APIs, such as Azure speech recognition, web RTC, and web audio APIs, but also Google Cloud functions.

What we learned

We learned how to use cloud functions, develop websites, implement APIs, and to use Firebase.

What's next for SlackR

We could implement additional features to further enhance the website, such as to recognize the specific speaker's name during transcription and to separate potentially overlapping speeches during transcription.