4.4 Using ELAN

As you saw in the lesson on SayMore, transcriptions and translations of a text can be stored in .eaf files. In this lesson, you will learn how to create .eaf annotation directly in ELAN. 

Creating an ELAN project

  1. Install and open ELAN. In the File drop-down menu, click New.
  2. Navigate to the target file or files. This may be a file with paired video and audio, or separate video and audio files. For each file selected, press >> to add it to the Selected Files prompt. Then, click OK.

Setting up Tiers and adding Segmentation

You will likely include translations of phrases and segmented individual words with glossing/translation. ELAN allows you to store these different types of information on separate tiers, which can be independent of the annotations found on other tiers, or dependent on the annotations in a Parent Tier. Independent tiers can be used, for example, to segment words or phrases. Dependent tiers can be used to add additional annotations to those words or phrases, such as their meaning or grammatical function. Other tier information might include the speaker, gestures, vocal characteristics, or other information relevant to your research. To set up tiers:

  1. In the Options drop-down menu, select Segmentation Mode. This brings up the imported video in the upper left panel, a Timeline panel across the bottom of the screen with the audio waveform and annotation tiers, and a Segmentation/Controls panel in the top right. 
  2. In the Tier drop-down menu, select Add Tier. This prompt allows you to add a Tier name, as well as the name of the speaker and the annotator. You can also specify here whether the new tier should be dependent on another tier by selecting the name of the Parent Tier.
  3. Once the tier settings are selected, click Add at the bottom of the prompt to add the new tier to the Current Tiers panel at the top of the prompt. Once you have added all of your tiers, click Close to return. 
  4. In order to enter segmentation, double click the tier you want to segment on the left of the Timeline panel to select it for segmentation. It will appear highlighted at the top of the tiers list. Click the point on the timeline where you want to add a segmentation, then hit Enter. In the Segmentation panel (top right), you have the option to segment with one or two keystrokes. Chose 'one keystroke' if you want your annotations to follow each other without gaps. Chose 'two keystrokes' if you want to annotate only certain pieces of the timeline, leaving gaps between segments. With two keystrokes, the segmentation won't appear until you have clicked Enter both at the beginning and at the endpoint of the selection. 

Adding Transcriptions and Annotations

  1. You can manually add a transcription to a defined segment by opening the Options drop-down menu and selecting Annotation Mode. Now, in the Transcription tier, double click any segmented chunk of speech on any tier and type your annotation (such as the transcription or other information). 
  2. You can automatically segment one tier based on information in another tier, for example, splitting a phrase in one tier into words in a second tier based on word breaks. First, type the transcription with spaces in the first tier. Then, in the Tier drop-down menu, select Tokenize Tier.  In the Tokenize Tier prompt, select the Source Tier (the phrase with spaces between words) and the Destination Tier (the blank tier you want individual words to appear on). Click Start to automatically segment your annotation based on word breaks. 
  3. By setting this Destination Tier as a Parent Tier, you will be able to add additional annotations, such as glosses, to other Dependent Tiers.

More ELAN training is available at: https://www.mpi.nl/corpus/html/elan/index.html