ALAnimatedSpeech

NAOqi Audio - Overview | API | Advanced


What it does

The ALAnimatedSpeech module allows you to make the robot talk in an expressive way.

By default, when you make the robot say something through this module, he livens the speech up with body language.

Options

  • You can complete this body language or create more complex speech by annotating the text yourself with some instructions.
  • You can also disable the body language so that the robot only executes your instructions and does not move outside of them.

How it works

A - The ALAnimatedSpeech module receives a text that can be annotated with instructions.

B - It splits the text into little chunks.

C - It analyzes the text and annotates the things he recognizes to choose contextual moves.

D - Any part of the text which is not annotated with animations are filled with Body language modes.

E - The module prepares the robot to execute each instruction so that it can be executed as soon as it is needed. This enables speech and instructions to be well synchronized.

F - The ALAnimatedSpeech module makes the robot say the text and keeps an eye on the speech to launch instructions at the right time.

Annotated text

The annotated text is a string combining text to be said and Instructions managing behaviors.

For example, you could send:

"Hello! ^start(animations/Stand/Gestures/Hey_1) Nice to meet you!"

In this case:

  • The robot says “Hello!”.
  • Then, simultaneously, the robot:
    • starts the execution of the animations/Stand/Gestures/Hey_1 animation and,
    • says “Nice to meet you!”.
  • Once the text is said, the robot stops abruptly the execution of the animations/Stand/Gestures/Hey_1 animation.

If you want it to keep running, you can add a ^wait instruction at the end:

"Hello! ^start(animations/Stand/Gestures/Hey_1) Nice to meet you ^wait(animations/Stand/Gestures/Hey_1)"

Instructions

The following instructions allow you to do some actions during the course of a speech.

Body language mode

Use ... To ...
^mode( body_language_mode )

Change the body language mode for the current text.

body_language_mode can be “disabled”, “random” or “contextual”

Example

"Hello. Look I can stop moving ^mode(disabled) and after I can resume moving ^mode(contextual), you see ?"

For further details, see: Body language modes.

Animations

Use ... To ...
^run( animation_full_name ) Suspend the speech, run an animation and resume the speech.
^start( animation_full_name ) Start an animation.
^stop( animation_full_name ) Stop an animation.
^wait( animation_full_name ) Suspend the speech, wait for the end of the animation and resume the speech.

Examples

"^start(animations/Stand/Gestures/Enthusiastic_4) Look what I can do while speaking!"

"^start(animations/Stand/Gestures/Enthusiastic_4) Look what I can do while speaking! ^stop(animations/Stand/Gestures/Enthusiastic_4) Now I use my body language"

"^start(animations/Stand/Gestures/Hey_1)Hi, guys! ^wait(animations/Stand/Gestures/Hey_1)"

For further details, see: Animations.

Tagged animations

Use ... To ...
^runTag( tag_name )

Suspend the speech, run an animation which has the mentioned tag and resume the speech.

Chooses randomly the animation which will be started among all the animations which have the mentioned tag.

^startTag( tag_name )

Start an animation which has the mentioned tag.

Chooses randomly the animation which will be started among all the animations which have the mentioned tag.

^stopTag( tag_name ) Stop an animation which has the mentioned tag.
^waitTag( tag_name ) Suspend the speech, wait for the end of the animation which has the mentioned tag and resume the speech.

Examples

"^startTag(me) My name is Nao."

"^startTag(hello) Hello. ^waitTag(hello)"

"^startTag(hello) Hello Paul, nice to meet you. ^stopTag(hello) My name is Nao."

For further details, see : Tags.

Sounds

Use ... To ...

^runSound( soundSet/soundFile )

^runSound( soundSet/soundFile, soundVolume )

Suspend the speech, play a sound and resume the speech.

soundVolume is an integer between 0 and 100. (by default it’s 100)

^startSound( soundSet/soundFile )

^startSound( soundSet/soundFile, soundVolume )

Play a sound in background.

The volume is an integer between 0 and 100. (by default it’s 100)

^stopSound( soundSet/soundFile ) Stop a sound.
^waitSound( soundSet/soundFile ) Suspend the speech, wait for the end of the sound and resume the speech.

Examples

"^startSound(my_sound_set/my_sound) That's cool."

"^startSound(my_sound_set/my_sound, 50) That's cool. ^waitSound(my_sound_set/my_sound)"

"That's cool. ^runSound(my_sound_set/my_sound) "

For further details, see: Sounds.

NAOqi API

Use ... To ...
^pCall( ALModule.method(args) ) Call asynchronously any NAOqi method.
$ my_event = my_value Raise an ALMemory event.

Examples

"^pCall(ALMotion.wakeUp()) Ok, I wake up."

"While, I am speaking, $my_event=my_value I can raise an event."

For further details, see: NAOqi API.

Body language modes

There is 3 modes available for the body language:

disabled - The body language never launches any animation.

random - The body language launches randomly some short animations which are executed one after another.

contextual - The body language launches some specific animations each time a keyword like ‘I’, ‘you’ or ‘all’ is detected in an appropriate grammatical context (for example, he will point at him if he talks about himself).
Each time no contextual animations has been found, it launches randomly a new animation.

You can add new associations between words and tagged animations using: ALAnimatedSpeechProxy::addTagsToWords()

The body language takes all motor resources if they are available. And it stops as soon as they are asked by another animation.

Animations

In an annotated text, you can use any supplied animation.

For further details, see:

You can also use your own behaviors provided that you follow these requirements:

  • Limit the content of the behavior to an animation (no speech, no complex behaviors), in order to guaranty a good compatibility with the Text-to-Speech.
  • Lock the resources as follow: Wait 1 second at box startup, and lock resources during box execution - in order to secure a full compatibility with the automatic body language process.

For further details, see : Creating a Behavior for annotated text.

Tags

In an annotated text, you can use any tag supplied by default.

For further details, see:

Additionally, you can declare new tagged animations using: ALAnimatedSpeechProxy::declareAnimationsPackage()

You can also declare new tags for some specific animations (without modifying the animations) using: ALAnimatedSpeechProxy::declareTagForAnimations()

Note

The tags used by Animated Speech are nothing more than the generic tags you can associate to any behavior. But for Animated Speech, we do confidence to the developers to use only tags that cover animations.

Sounds

In an annotated text, you can use any sound file contained in a sound set, using the following syntax:

soundSet/soundFile

For further details, see: Soundset.

NAOqi API

You can call any NAOqi method or set any event value, at any time of your speech.

Here is the syntax:

ALModule.method(args)

$event=value

Note

$my_event=my_value is equivalent to ^pCall(ALMemory.raiseEvent("my_event", "my_value")).

Features inherited from other modules

Add your own behaviors

You can install your own behaviors through the Application Manager panel and uses them in your text.

Warning

As you can call your own behaviors in an annotated text, you could call a behavior which is not an animation. But we advise you to only annotate text with animations. Also we strongly recommend you to set the resources of your behaviors before you start using them in an annotated text.

For further details, see: Creating a Behavior for annotated text.

Voice customization

As ALAnimatedSpeech uses ALTextToSpeech module to say the text, you can customize the voice of the robot.

For further details, see: ALTextToSpeech.

Sound managing

ALAnimatedSpeech uses ALAudioPlayer module to play the sounds.

For further details, see: ALAudioPlayer.

Getting started

To discover the basic functions of the animated speech using Choregraphe, see the tutorial: Testing the Body language.