Skip to content
MarieMichellePepin edited this page Oct 16, 2018 · 2 revisions

Prerequisites

You will need ROS (indigo or later).

Installing

Clone the project:

git clone https://github.com/introlab/irl_audio.git

If you don't already have one, you need to create a catkin workspace and put the irl_audio project in the /src folder.

  • Run catkin_make inside the catkin workspace to build and compile the project:

    catkin_make
    source devel/setup.bash
    
  • Or you can build and compile it manually by initiating the workspace and creating a folder to build the project:

    cd ~/catkin_ws
    cd src
    catkin_init_workspace
    cd ..
    mkdir build
    cd build
    

    Run CMake (in the build directory):

    cmake ../src -DCMAKE_INSTALL_PREFIX=../install -DCATKIN_DEVEL_PREFIX=../devel
    

    Compile the project:

    sudo make
    source devel/setup.bash
    

You can test the installation with:

rospack find manyears_ros

It should return the path of the manyears_ros package. If not, make sure you have sourced your workspace with:

source devel/setup.bash

Configure ManyEars

  1. Make sure your sound card & microphone are working.

  2. You must configure your microphone array before starting the system.

    • You can use the ManyEars GUI to position your individual microphones and tune the system. In the ManyEars GUI you can save the configuration file for usage with the launch files using the "Parameters->Save the parameters of the system" menu.

    • You can also use your own file based on an existent configuration file in "manyears_ros/data" and modify it with the configuration of your array of microphones.

  3. Modify the launch files to fit your configuration.

Testing the system

  1. Open a new terminal and launch the rt_audio node.

    roslaunch rt_audio_ros rt_audio.launch
    

    The rt_audio node should select the device card and publish the audio stream to the audio_stream topic.

  2. Open a new terminal and launch the many_ears node.

    roslaunch manyears_ros many_ears.launch
    

    The many_ears node should subscribe to the audio_stream topic and publish to the tracked_sources topic.

Separated sources will be stored in the /data directory in the raw format.

ROS topics available

  • /audio_stream: Audio stream from the sound card detected with rt_audio, 48000Hz
  • /manyears/source_pose: Sources positions (using TF)
  • /manyears/sources_info: Sources informations (ids, positions, separated information 48000Hz, normalized floats[-1,1])
  • /manyears/stream: Audio stream from the sound card, 48000Hz, 16 bits, LE.
  • /tracked_sources: Tracked sources
  • /rosout
  • /rosout_agg
  • /tf

Clone this wiki locally