VocalTractLab-dev

This is the VocalTractLab repository for on-going development work. This is not the place for official stable releases!

You can find the bleeding edge builds here including new and experimental features, some of which may break your old projects. For official stable releases, please check the VocalTractLab website.

Please feel free to fork this repo and make your own contributions, though!

The main branch of this repo is reviewed on a semi-regular basis for inclusion into the official release.

Clone the repository

This repository includes the VocalTractLabBackend library as a submodule. You therefore need to clone it slightly differently than what you may be used to:

git clone --recurse-submodules https://github.com/TUD-STKS/VocalTractLab-dev.git

Build instructions

VocalTractLab depends on the following additional libraries:

• VocalTractLabBackend
• wxWidgets 3.1.3
• OpenGL
• OpenAL (on Windows or Linux) or WinMM (on Windows)

The VocalTractLabBackend library is included as a submodule in this repo and built as part of the frontend build workflow.

Install dependencies on Linux (tested on Ubuntu)

You can install the dependencies and build VocalTractLab by simply running the script build/linux/build.sh. If for whatever reason you want to walk through the steps manually, here is what you need to do.

First we need to install OpenGL and OpenAL:

sudo apt-get update 
sudo apt-get install libgtk-3-dev libopenal-dev mesa-utils freeglut3-dev

Now you should be ready to install wxWidgets. Unfortunately, the current wxWidgets binaries in the APT repositories are still version 3.0. So we need to download and build wxWidgets from source like so:

wget https://github.com/wxWidgets/wxWidgets/releases/download/v3.1.3/wxWidgets-3.1.3.tar.bz2
tar -xf wxWidgets-3.1.3.tar.bz2
cd wxWidgets-3.1.3
mkdir buildgtk
cd buildgtk
../configure --with-gtk --with-opengl
make
sudo make install
sudo ldconfig

That should be it.

Install dependencies on Mac OS X

First of all, you need to have Xcode installed (including the CLI toolkit). An Apple Developer account is mandatory as well as administrative rights on the system you want to run VocalTractLab on. Next, make sure you have git installed. Mac OS X already comes with git installed normally. If you do not have git installed for whatever reason, please do so. In addition, make sure CMake is installed and added to your PATH. Last thing to check is wheter curl is installed on your system. If not, please do so. You now should be able to install the dependencies and build VocalTractLab by simply running the script build/osx/build.sh. If for whatever reason you want to walk through the steps manually, here is what you need to do.

Mac OS X already ships with OpenGL and OpenAL built-in. You just need to build and install wxWidgets:

curl https://github.com/wxWidgets/wxWidgets/releases/download/v3.1.3/wxWidgets-3.1.3.tar.bz2 -O -L
tar -xf wxWidgets-3.1.3.tar.bz2
cd wxWidgets-3.1.3
mkdir buildosx
cd buildosx
../configure --with-opengl
make
sudo make install

Install dependencies on Windows

You can install the dependencies and build VocalTractLab by simply running the script build/msw/build.ps1. If for whatever reason you want to walk through the steps manually, here is what you need to do. On Windows, OpenGL and WinMM are part of the Windows SDK and should already be available on your system. All that is left to do is build wxWidgets. Bring up a Powershell and run the following commands:

Invoke-WebRequest https://github.com/wxWidgets/wxWidgets/releases/download/v3.1.3/wxWidgets-3.1.3.zip -outfile wxWidgets-3.1.3.zip
Expand-Archive wxWidgets-3.1.3.zip

Now open the solution wx_vc16.sln in the folder wxWidgets-3.1.3\build\msw using Visual Studio 2019 or above and build all projects. Once all libraries are built, copy them to the location where the VocalTractLab build files expect them. Bring up a Powershell again and run (from the parent directory of wxWidgets-3.1.3):

xcopy wxWidgets-3.1.3\* C:\wxwidgets-3.1.3\ /s /h /e

Build VocalTractLab using CMake (Linux, Mac OS X, Windows)

You can build VocalTractLab on Linux and Windows using CMake like so (starting from the repository root folder):

mkdir tmp
cd tmp
cmake .. -DCMAKE_BUILD_TYPE=Release
cmake --build . --target VocalTractLab --config Release

Build VocalTractLab using Visual Studio (Windows)

Alternatively to CMake (see above), you can build VocalTractLab on Windows using the provided Visual Studio solution VocalTractLab2.sln in the directory build/msw. Simply open it with Visual Studio 2019 or above and build all projects.

Running VocalTractLab

After building the program, the executable file can be found in bin/Release/. Before running it, you may want to copy the speaker file JD3.speaker from the folder resources/ to the same folder as the executable. Otherwise you will have to load it at runtime.

Getting started

Check out the official manual for detailed instructions and examples. Note that the manual is only updated with official releases, so not everything may still apply to the most recent dev release from this repository.

You can also find some example gestural scores and segment sequence files in the folder examples/.

Please be aware of the fact that VTL does not support theming currently. So if the layout/designs seems to be off for you, please check if you are using the default (light) theme of your system.

Troubleshooting

If you run into any problems in the build process, make sure all the dependencies are correctly installed. If you still run into any trouble, feel free to open an issue.