[url removed, login to view]
One of the highlights of this project is what you will be working with the chief AT&T Developers who designed the Watson API. They have agreed to help review the application as it is being built and provide feedback on how to best incorporate the API. You will also need to use flash to capture the audio clips of the user. We will provide the built-in audio clips of the words being said.
The deadline for it is tight, it needs to be completed by Oct. 30th, so we need the programmer to begin working immediately.
Description the app:
Voice Recognition Vocab Practice: For the voice recognition portion of this app, we will be licensing the Watson API from AT&T. Watson is the voice recognition software that can be linked into an app at the beginning of the program and then be called throughout the program. I can provide more info and tutorials on using the Watson API.
The program itself will be quite simple and will run as follows:
Interface notes: A window (to look like a note/flash card) with a word in the middle of it. One button that is labeled “Play”, another button labeled “Speak”, another button that changes labeled “Next”.
1) First a word will be shown on the screen
2) The user will be prompted to click “Play” to play the audio clip of the word (we will provide the audio clips).
3) The user will then be prompted to click on “Speak” to say/record the word into their microphone.
4) After a 2-second pause from the last audio input that was received, the program will stop recording and check the audio input. [Note if it is too difficult to end the recording with the pause function, another option is to include a “Stop” button to allow the user to indicate they are done recording their input].
5) The Watson API needs to be called in the program and it will check the audio clip and convert it into text.
5) The program needs to check this text with the word on the screen. If it is a match, the program should display correct and record a correct score. Else, the program should display “Sorry please try again”. The user can then click the “Play” button to hear the sound clip again and redo the loop (unlimited attempts) OR they can choose “Next”.
6) At the end of the program (40 words) the program will display the total number correct matches and total number of words not said correctly (please keep track of attempts in the program but do not display them at this time).
This app should also look visually appealing, especially to children. We can provide cut out images of the characters as well if you choose. The app will be initiated once an icon on our website is clicked on.
We would like to choose the programmer immediately so they can begin working on this right away.
The app needs to be completed and fully functional (without bugs) and uploaded onto the website by [url removed, login to view] 11pm (Seattle / Pacific Time zone).
The budget is fixed at $500 which will be paid as a lump-sum payment when the project is finished.