LiveSubtitles

From air
Jump to navigation Jump to search

Project presentation

Transcribe a teacher speech to subtitles and allow students to correct misinterpreted words

Team

  • Supervisors : Jérôme Maisonnasse
  • Members : BUI David / LECHEVALLIER Maxime / OUNISSI Sara


Specifications

Make an app usable in any browser (mainly Google Chrome)


Google API Speech

Key words : new paragraph, comma, dot

Not supporting long speech

Links

GitHub


Documents API specs : https://dvcs.w3.org/hg/speech-api/raw-file/tip/speechapi.html

Progress of the project

The project started January 11th, 2015.

Week 1 (January 11th - January 17th)

First interview with our supervisor Jérôme. We've learned more about our project and what is expected for the next weeks

  • Handling the project
  • testing Google API Speech
  • Making git repository

Week 2 (January 18th - January 24th)

Week 3 (January 25th - January 31th)

  • Microphone works only when a virtual server is installed, we try with apache (Lamp/Xamp)
  • Learning JavaScript
  • Learning HTML/CSS
  • Trying Bootstrap
  • Amara.org is a website to edit youtube subtitles, might help

Week 4 (February 1st - February 7th)

  • Scrum
  • Trello
  • Trying to add grammar and key-words (like "OK Google") => Not possible

Week 5 (February 08th - February 14th)

Design patterns

  • Model-View-Controller (GoF) : This pattern is used to separate application's concerns. Our project is Web oriented program
  • Singleton (GoF) : Ensure a class has only one instance, and provide a global point of access to it.

Example : a teacher is the only one who can launch slides

  • Visitor (GoF) : Represent an operation to be performed on the elements of an object structure. Visitor lets you define a new operation without changing the classes of the elements on which it operates.

Example : students can edit the subtitles

  • State (GoF) : Allow an object to alter its behavior when its internal state changes. The object will appear to change its class.

Example : Microphone detection

Project work

Solving critical problems : the API is not working with ambient noise. When we are talking directly to the microphone the API is working fine.

Tests :

  • Fast talking : Dead after 1 minute
  • Slow talking (with interruptions) with music arround : Dead after 2 minutes
  • Slow talking : Dead after 2 minutes

Meeting with Jérôme to have new directions after a quick demo of the app.

Week 6 (February 15th - February 21st)

Studying Socket.io, trying the demo chat, linking Reveal.js with socket.io

WebStorm is a Javascript IDE but too complicated too use for us

Week 7 (February 29th - March 6st)

Transmitting data from client to server with socket.io Working on adding collaboration part (javascript database?) Working the presentation

Gallery