mnutsch / talktometeddy

This is the repository for the Talking Teddy android app.

talktometeddy

This is the Talking Teddy android application version 0.5.

talktometeddy's Issues

Upload v0.5 version to android app store.

Need to submit apk, screenshots and description to android market place.

Help menu in text

we need a help context menu that explain the app in clear text. This is important as app doesn't have any other mode of interaction.

In addition, we can add 'help' keyword and recite the menu when user says it.

Adding user logging and signing up functionality - Client side

Add simple user sign up and logging functionality.

False negatives: negative response for some common phrases

What do you know?
What things do you know?
How old are you?
Who are you?

Received "I didn't understand response" on these phrases above. These were tried again after initial opening conversations.

Usage logging is broken with v1.0 update

Users actual and matched prompted logging got broken with v1.0 update.

SEO

to get better visibility in app store and among audience

Reporting usage

Integrating common audio clips or sounds in app

Laughing sounds, beep sounds that can be played along with TTS to add rich auditory experience.

Interaction issue: After 2-3 unrecognizable responses, we should repeat the main menu.

When Teddy responds negatively 2-3 times consecutively, we should keep a counter and repeat the main menu to guide the user in right direction.

This is important because right now our app doesn't have text help menu. In fact, I am creating another issue saying 'we need a help context menu that explain the app in clear text'.

Add user logging and signup functionality - Server Side

Voice only Vs button (Initiate chat)

Voice only, no need to touch the screen

Changing voice to sound more like a kid

Image/story telling exprience

Adding more images/visuals especially for longer answers

Create API Server Caching

The goal is to speed up response time for common user statements when there are a large number of options.

Additionally, figure out how to still allow for randomized "Say" responses with server caching.

Music and tones (voice recording)

Richer auditory experience

TTS does not properly close if there is another click before playback starts

Steps to re-produce:

Click on Teddy, say something (I used random word: "Computer")
As the Google Speech-to-text popup closes, click on the button again and Google Speech-to-text listens to voice input.
Don't say anything and watch Teddy feedback played back and fed into the Speech-to-text.

False positive: For 'Lalalalala' Teddy told a joke!

Might be low priority as this was hard to reproduce.

Creating audio clips for scenario videos

Kid friendly voice

Tweaked pitch and speech rate to sound Teddy more kid friendly.

Checked in the code yesterday. Would be great to see the feedback on this from users

Upload latest v0.5 version to android app store.

Update

• Package name change to com.talkingteddy
• version # update to v0.5
• grammer fix

Change package name

SimpleSpeechActivityDemo to a meaningful name

Code and Test an Expanded Conversation Data File

Add additional scenarios.

Network lag of around 3-5 seconds

I like the idea of 'Thinking...' toast message to handle waiting time, however I did notice considerable network lag in response - more than 5 seconds at times. I tried this on a public WiFi network.

We will probably have to do performance analysis to see if we can improve on this.

'Thinking...' toast shows up for few seconds after answer is started.

For the questions that Teddy understands or even doesn't understand, 'Thinking...' toast appears even after the TTS answer has already started.

The TTS and answer toast should be in sync. This might as well be because of network delays.