ToBI Annotation Environment on WaveSurfer

This manual explains how to set up and use the ToBI Annotation Environment on WaveSurfer, both for running the examples and exercises in the "Guidelines for ToBI Labeling", and for doing ToBI labeling using a graphical user interface.

If you find an error or have suggestions, please send me e-mail to agus [at] cs.columbia.edu.

About WaveSurfer: Version 1.8.3 or higher of WaveSurfer required (available at http://www.speech.kth.se/wavesurfer/).


  » 1. Setup
  » 2. Examples and Exercises
  » 3. ToBI Labeling
  » 4. Text Transcription

  « Back

1. Setup

After downloading and uncompressing "ToBI_WaveSurfer_v#.#.zip", you should have these files:
  examples/
  exercises/
  preferences
  README.txt
  ToBI.conf

In order to set things up, you will need to:
  1. Locate the ".wavesurfer" directory created by the program.
  2. Copy the 'preferences' file in .wavesurfer/1.8/
  3. Copy 'ToBI.conf' and 'Games Transcription.conf' in .wavesurfer/1.8/configurations/

Note: "1.8" is the version of WaveSurfer. If you have a more recent version, it might be 1.9, 1.10, 2.0 etc. Use the directory that you find in ".wavesurfer\".



2. Examples and Exercises

To open the examples and exercises in the "Guidelines for ToBI Labeling" (version 3.0, March 1997) follow these steps:

  1. Choose File » Open.
  2. Select the desired wav file from the "examples" or "exercises" directory.
  3. When prompted for a configuration, choose "ToBI". If that option is not available, check that you have correctly copied the 'ToBI.conf' file in the correct directory (see the Setup section).


3. ToBI Labeling

Annotation Files

Each wav file has four annotation files in its same directory, with file extensions 'tones', 'words', 'breaks' and 'misc'. For example, 'sample.wav' would have these four optional associated annotation files:

To open a wav file and its ToBI annotations (if any) follow these steps:

  1. Choose File » Open.
  2. Select the desired wav file.
  3. When prompted for a configuration, choose "ToBI". If that option is not available, check that you have correctly copied the 'ToBI.conf' file in the correct directory (see the Setup section).

If the annotation files are found, they are displayed in the corresponding tiers. Otherwise, the tiers of the missing files are left blank.

Therefore, to start labeling a wav file from scratch, just open the wav file after making sure that none of the annotation files exist in the same directory.

ToBI Labels

ToBI labels for the Tones, Break Indices and Misc tiers are inserted by clicking the right button of the mouse at the desired position, and selecting the corresponding label from the drop-down menu.



4. Text Transcription

To transcribe the text of a wav file, follow these steps:

  1. Transcribe the complete text into a plain text file, using Notepad or a similar text editor. You can use either WaveSurfer or any other sound player to listen to the wav file.
  2. When you are done with the full transcription, open WaveSurfer.
  3. Choose File » Open.
  4. Select the desired wav file.
  5. When prompted for a configuration, choose "Games Transcription". If that option is not available, check that you have correctly copied the 'Games Transcription.conf' file in the correct directory (see the Setup section).
  6. Right-click on the words tier, and choose Load Text Labels... Select the text file containing the complete transcription.
  7. Left-click on the words tier to give it focus. Press the TAB key. This will insert a label with the first word in the transcription, at the current position of the mouse. After dragging the label to its correct position, click on a new position (to the right of the previous one) and press TAB to insert a label for the next word. Repeat this procedure until all the words in the transcription have been inserted.
  8. Saving the resulting transcription will generate a file with the ".words" extension, which will later be the ortographic tier of the ToBI labeling.

Tips

The "Browse" Dialog

WaveSurfer provides another useful feature for doing text transcription. By right-clicking on the transcription tier and choosing the "Browse..." option, you can open a dialog that allows:



« Back last updated