Date: 2012-Jun-22
Version: 1.2.1

Table of Contents

What's New

We are including diagnostic reporting. If you want to participate, read the Data Collection Policy.

1.0.0

1.0.1

1.1.0

1.1.1

1.2.0

1.2.1

Notice

Any incidentally-mentioned product names, or images in screen shots, are trademark by their respective owner.

Overview

Thanks for choosing Text Talk!

Text Talk is an application for text messaging, featuring proactive use of Speech Recognition, Text to Speech, and Haptic Feedback. The goal is to provide you with an eyes-free messaging experience!

This Edition supports SMS messages only!

Back to Top

Features

Text Talk uses the following features found on your Android device:

Though intended for eyes-free use, Text Talk contains some visual elements that may not render in a Screen Reader. These are not required for eyes-free usage. In reverse, there are Screen Reader content descriptions those users will appreciate, that non-SR users will not know exist.

Troubleshooting

Text Talk relies on your device's built-in feature support, and any feature may be absent (at the device manufacturer's discretion). If you have other speech-enabled apps that "work" then these are implementing their own TTS and/or Speech Recognition, and they are not using the "system" version of these services.

Text to Speech

If you hear the earcons (tones) but no speech, you most-likely do not have the Text to Speech resources installed. Go to the device Settings for Speech Input and Output, and install them. That screen should lead you to Google Play to download and install that package.

Speech Recognition

If you do not have Speech Recognition installed on your device, Text Talk displays a message to that effect, and you should also hear it via TTS.

Back to Top

Tasks

An important accessibility enhancement is the task, a dialog you operate using Speech Recognition and the Command Mode as inputs, and Text to Speech and Vibrate as outputs.

Tasks are used instead of keyboard input, when this is more convenient. You always have the ability to use the standard data entry methods as well, e.g. Keyboard.

Most tasks begin with Speech Recognition, and the rest is organized to help you process the results for your "final answer".

Command Mode

An important part of providing input to a task is the Command Mode. These modes are supported:

There are 4 input actions:

  1. Command 1: this is usually the Primary Command action. When working with lists, it is the Previous action.
  2. Command 2: this is usually the Secondary Command action. When working with lists, it is the Next action.
  3. Cancel: this will "abort" one or more levels of nested tasks. When the original task exits, you exit the entire screen.
  4. Accept: applies the task results, e.g. enter message text.

The following table has complete information.

Command 1 Command 2 Cancel Accept
Keys Volume Up Volume Down Back Wait 3 Seconds
UI Buttons "1" Button "2" Button "X" Button "Check" Button
Option Menu As Required As Required "Cancel" item As Required

Speech Recognition

An important feature of Speech Recognition is the list of More Results, which often contains the correct utterance, when the Number One result is not a "match" to what you spoke. Depending on the utterance, there may be up to 20 alternative phrases returned.

The Speech Recognition cycle in a task proceeds as follows:

  1. An earcon (tone) plays to alert you.
  2. Audio feedback UI appears.
  3. Device Vibrates when you should start speaking.
  4. Speak.
  5. Audio feedback UI disappears, Progress UI appears.
  6. Speech Recognition processing begins.
  7. Progress UI disappears when results are ready.

This is usually followed by read back of the number one result, and the option to examine More Results.

Text to Speech

The task communicates with you using Text to Speech. The most-important use of this is repeating back to you the Speech Recognition results.

The task also composes utterances to provide you with information, e.g. message length. These "system" utterances are done in a slightly different "voice" than your default.

Common Tasks

There are some tasks that get reused due to their utility.

Select Item

This task leads you through selecting an item from a list.

The task starts with the first item as the current item. Each item is read out in this format: N of M: the item.

After reading out the current item, the following actions are available:

Message

This task leads you through Speech Recognition, with the option to examine the More Results.

  1. Invoke Speech Recognition as detailed in that section.
  2. The Number One result is read back to you.
  3. When read back completes, the following actions are available:
  4. If you return from Select Item with a selection, enter that into the Message text field. If you cancel Select Item, it also cancels the Message task.
  5. If there is existing message text, the text you just entered is appended to it, and the newly-combined message is read back to you.
  6. If the Read Message Count feature is On, read the current message length.

Recipient

This works the same as the Message task, but you are speaking:

A partial match is used to form the list of More Results. The final selection goes into the Recipients text field.

Unread Messages

This task loads all unread messages and presents them to you with a Select Item task. Each item contains the sender and body.

Back to Top

Conversations Screen

This screen displays your active conversations, ordered by most recent first. A Create New Message item appears first.

Select a Conversation from the list to enter the Compose Screen for that Conversation.

Task

The following actions are available as soon as the screen starts:

If the Start Unread Task setting is On, the Unread Messages task starts automatically when you activate via the Notification bar.

Option Menu

Activate with the Menu key.

Context Menu

Activate with a long-click on a Conversation list item.

Back to Top

Compose Screen

This screen displays a specific conversation, ordered by most recent message last, at the bottom. It contains these elements:

Message Length

The message length UI is always visible. It displays the following values, separated by slash "/".

  1. Number of characters Total,
  2. Number of remaining characters in Current Message Unit,
  3. Number of Message Units used.

When Read Message Length setting is On reads the total message length after completing the Message task or using the keyboard.

The Content Description of the Send button is also set to the message length.

Task

The following actions are available as soon as the screen starts:

New Message

When started for a new message, the screen immediately starts the Recipient task.

Option Menu

Activate with the Menu key.

Context Menu

Activate with a long-click on a Message list item.

Back to Top

Notifications

Text Talk uses Text to Speech in addition to the usual device notification mechanisms:

Ignore Ringer Mode

Each TTS-based notification has an option to ignore the current Ringer Mode, so it will still speak if your device is on Vibrate or Silent.

Incoming SMS

When enabled, this feature reads the sender of incoming messages.

Sender and Body

When enabled, this feature reads the sender and message body.

Unread Message Count

When enabled, this feature reads the count of unread messages, when you wake the screen via the Power Key. It has a "silent period" of 30 seconds.

Back to Top

Settings

Text Talk has the following settings.

SMS Settings

Notification Settings

TTS Settings

Screen On Settings

The screen is usually awakened with the Power key.

LED Settings

Vibrate Settings

Conversations Settings

Compose Settings

Task Settings

Extra Settings

Back to Top

Data Collection Policy

You can submit error logs on an opt-in basis, or you may be requested to do so by Support.

We collect information in the device's system logs via the logcat system tool. We restrict collection to:

  1. All messages from our applications,
  2. All device-wide Error class log messages.

We show you all the data that will be sent, via your device's Send Email screen. We send some of your device's information, but no identifying information. We try very hard not to display Personal Information in the logs you send to us. If you detect anything you are not comfortable sending to us, you may elect not to send a report.

We only retain these logs long enough to troubleshoot issues.

Back to Top