CMU-TBD / tbd_polly_speech

ROS Wrapper for Amazon Polly Speech with local caching of generated audio.

tbd_polly_speech

License - MIT
Maintainer - [e-mail]

A ROS wrapper for Amazon Polly Text-to-Speech service. We also cache locally each soundfile, so if you ever repeat the same sentence with the same voice, a local copy of the audio will be used instead of sythesizing it again.

Dependencies

ROS Packages:

• tbd_ros_msgs (https://github.com/CMU-TBD/tbd_ros_msgs)
• sound_play (http://wiki.ros.org/sound_play)
• IF uses tbd_audio_common play type:
  • tbd_ros_msgs (https://github.com/CMU-TBD/tbd_ros_msgs)
  • tbd_audio_common (https://github.com/CMU-TBD/tbd_audio_common)

Python Dependencies

• boto3

Usage

Running

1. Make sure you have a AWS credentials file setup on the system. Guide by AWS. Make sure the account has AWS Polly Speech Enabled
2. launch the backend services

roslaunch tbd_polly_speech polly_speech.launch

3. You can access the service either through the Python API

from succes_polly_speech import PollySpeech

ps = PollySpeech()

ps.speak("I am a good robot",voice_id='Joanna')
ps.speak("I am not a scary robot",voice_id='Joanna', block=False)
ps.wait()
ps.speak('Hello World. I will be interrupted',voice_id='Emma', block=False, cancel=True) 
ps.stopAll() #Interrupts the sentence and stop the voice command

OR directly calling the action server at the topic tbd_polly_speech/speak with the action pollySpeechAction.

ROS Parameters

There are three ROS parameters in the launch file

• no_audio, true if you just want to simulate it and not actually running the code.

• play_type, the default is sound_play in the ros-driver/audio_common repository. You can install this with sudo apt install ros-melodic-audio-common. The alternative is TBD's lab own audio stack (tbd_audio_common) that uses actionlib instead of ROS messages and plays faster.

• polly_audio_storage_path, the path to the location you want to store the audio and also the masterlist.txt which stores the coding from phrases/text to filename. the default is PACKAGE_ROOT/audio_storage

Voices

A list of voice ID can be found here: https://docs.aws.amazon.com/polly/latest/dg/voicelist.html

Change logs

Past Contributors

• Joe Connolly - 07/2018