All annotations are entered either into the Inline Edit box or into the Edit Annotation box, which are accessed via the Timeline, Interlinear, Grid, Text and Subtitle Viewers. ELAN supports the following options:
In all cases, the Inline Edit box appears, e.g.:
It is possible to enter text that contains line breaks. The text entry box automatically displays a scrollbar if necessary.
The Inline Edit box is automatically preconfigured for the default character set of the tier. If you want to use a different character set, do the following:
If you are using a third-party keyboarding solution like Keyman, make sure to select the default system language as input language for the tier to be edited (e.g. Dutch if your system language is set to Dutch).
Only selections on time-alignable tiers can be saved in this way. To save a selection on a referring tier, see Making and saving a selection on a referring tier.
To exit the Inline Edit box without saving, do one of the following:
When annotations are created, they can be aligned with the video frames by setting it as a preference (see Editing preferences)
An alternative method to get an Inline Edit box is:
An Inline Edit box appears on the selected tier. You can now enter an annotation and save it in the way explained above.
The Edit Annotation box differs from the Inline Edit box in that it has a bigger, resizeable, display and supports line wrapping. It is therefore better for entering longer texts.
To enter an annotation into the Edit Annotation box, do the following:
The Edit Annotation box is automatically
preconfigured for the default character set of the tier (see Changing tier attributes). If you want to use a different
character set, do the following:
To exit the Edit Annotation box without saving, do one of the following:
To return to the Inline Edit box, do one of the following:
If the tier type of the selected tier has a controlled vocabulary both the Inline Edit box and the Edit Annotation box show a list of the CV entries from which an entry can be selected. To commit to a selected entry, use Enter or Ctrl+Enter. Double clicking an entry has the same effect.
The use of just a list works well when the number of entries is limited. For larger CVs another method of selecting the correct entry can be used. If you are either in the Inline Edit box or Edit Annotation box you can find and select a CV entry by reducing the selection list as you type the first characters of an entry. To enable this method, do the following:
There are some options you can set for the suggest panel, which can help in searching the CV entries. These options can be set in Preferences (See Editing preferences )
Annotations on some tiers can be subdivided into smaller units. In this way, you can for example break up a sentence into different words (or words into morphemes, etc.), as in the following illustration:
To divide an annotation into smaller units, do the following:
If you click on New annotation before, the original annotation is divided and the new annotation is inserted to its left (as in the illustration below). If you click on New annotation after, it is inserted to its right.
This option is only available for those tiers that are assigned to the stereotypes Time Subdivision and Symbolic Subdivision (see Basic Information: Annotations, tiers and tier types).
An annotation is always subdivided into two units. If you need further subdivisions, repeat the steps above.
To create depending annotations for a current active annotation, select Annotation > Create depending annotations. This will create dependant annotations for the current active annotations on all its child tiers.
Gaps between annotations on a tier can be transformed into annotations on the same tier or on a new tier. To do so, click Tier > Create Annotations from Gaps.... In the dialog window, select the (independent) tier for which you wish to transform the gaps into annotations. Then select whether you wish to put the new annotation on the same tier or on a new tier and specify a tier name in the latter case. Also specify the contents of the new annotations: either a specific value, the duration of a gap or no contents.
It is possible to select multiple tiers when creating annotations from gaps. Selecting multiple tiers can be done by holding the CTRL key while clicking other tiers than the one already selected. The SHIFT key can be used in a similar way to select the range of tiers from the one that is selected to the one that is clicked. The gaps created from multiple tiers are periods where each of the selected tiers has no annotation.
When there is a time interval you wish to divide in a number of adjacent annotations with the same duration you can do that by clicking on Tier > Create Regular annotations. In the dialog window (see Figure 2.82, “Regular Annotations”) select the tiers that must receive the new annotations. Then do one of the following:
Finally, enter a Annotation size and click Apply.
The total duration of the new annotations does not exceed the interval between
start and end time. So if start time is 1.000 seconds and, end time is 4.000 seconds
(interval of 3 seconds) and the annotation size is 2.000 seconds, than only one new
annotations is created because two would make a duration of 4 seconds which exceeds
the
interval defined by the start and end time.
If there is an overlap between the new annotations and one or more existing annotations, these existing annotations will be removed.
If two annotations are adjacent and have the same length (for instance after creating regular annotations, see Creating regular annotations) and if the second annotation is empty, you can copy the value of the first annotation to the second. This is done by selecting the first annotation and pressing CTRL+SHIFT+D. Now the second annotation is selected and you can repeat the process to get a string of adjacent annotations with the same value.
An additional result of CTRL+SHIFT+D is that if there are annotations on another tier that have the same begin and end time as the annotations you are working on, and the second of those annotations is empty, then the value of the first annotation on that tier is also copied to the second annotation on that tier.
ELAN offers the possibility to create the so-called snap-annotations. If two annotations are adjacent you can snap them by specifying the maximum close-value in ms. You can set the value to snap annotations in the edit preferences (see Editing preferences).
Sometimes you don't want annotations to contain transcriptions of the media you are working on, but you rather want to label and number them for later processing. In ELAN this is accomplished by clicking Tier > Label And Number Annotations.... In the dialog as shown in Figure 2.83, “Label and Numbering annotations” you can select one or more tiers to provide labels and numbering.
The formatting options of the label and numbering include:
The result of the options is shown below the options in a blue box. This result is updated as you change the options.
To modify an annotation, do one of the following:
The Inline Edit box appears (see Entering annotations (into the Inline Edit box)).
The Inline Edit box appears (see Entering annotations (into the Inline Edit box)).
The boundaries or alignment of annotations can be changed at any time, as described in the Changing the boundaries of an existing selection and annotation section.
It can occur that you put an annotation on the wrong tier. If the start and end time of the annotation unit are correct, you can move the complete annotation to a different tier by pressing and holding ALT and clicking and dragging the annotation with the mouse. This feature currently only works with time based annotations.
Dependent or child annotations will only be moved along with the parent annotation unit if it is clear from the tier names to which tier they should be moved. (from tx@A to tx@B for example). If this is not clear, the dependent annotations might be lost. A safer way to move annotations is to use the copy and paste annotation groups. See Copy and paste annotation groups
To change the case of letters of all annotation on a tier, click Tier > Change Case of Annotations.... Then select the tier(s) you wish to change, choose the case the annotations should be converted to and click OK. When changing the text to Upper-case it is possible to change only the first character (initial capital). Similarly when changing to Lower-case, it is possible to specify that each annotation should begin with a capital.
To delete the value of annotation, do the following:
The annotation value is deleted without further warning.
To delete one annotation, do the following:
ELAN also gives the possibility to delete multiple annotations: click on Annotation > Delete in the main menu. Now click one of the five menu items:
To delete a number of specific annotations on more than one tier, select those annotations by holding ALT while clicking them. The annotations get a purple border. Then right click in the Timeline Viewer and select Delete Selected Annotations.
Annotations are deleted without further warning.
If you delete an annotation on a parent tier, the corresponding annotations on all its child tiers will be automatically deleted as well. Please make sure that you do not accidentally delete a child annotation. An annotation on a child tier can be deleted without consequences for the annotation on its parent tier.
To delete annotations or annotation values on multiple tiers, select Tier > Remove Annotations or Values.... This will bring you this dialog:
Select the tiers on which the annotations are to deleted. First select whether to delete Annotations or Annotation Values on the selected tiers and then select All Annotations to delete all the annotations/ values, Empty Annotations for removing annotation units with no values in them or select Annotations where value is... to delete only annotations/ values where annotation value matches the given value and finally click on OK.
Annotations that are on a tier without a parent tier (top level tier) and that have no or only depending annotations on symbolically associated tiers can be split in two ways:
This will split your annotation exactly in the center and both the annotations will have the same values.
This will split the annotation at the point where the right click is made.
You can also split an annotation (in one of the ways described above) which has only time aligned depending annotations (symbolically associated). In this case the annotation will be split together with its depending annotations.
It is possible to merge two annotations on an independent tier into one annotation. To do so, select the first annotation of the two you want to merge. Then either right click in the Timeline Viewer or click Annotation in the ELAN main menu and select Merge with Next Annotation. The resulting annotation starts where the first had started and it ends where the second had ended. Note that the annotations on dependent tiers are also involved in the merging process. The result is that the last annotation on a dependent tier that is part of the first annotation on the independent tier is stretched so that it ends where the second annotation on the independent tier begins. Figure 2.85, “Merge annotations” illustrates this.
It is also possible to merge with the annotation before the selected annotation. It works in the same way as "Merge with next annotation. To do so select Merge with Annotation Before instead of "Merge with next annotation" in the above mentioned options.
ELAN supports different character sets, making use of the following input methods:
Characters can be entered by using a different keyboard mapping. This method is implemented using the GATE Unicode Kit developed at Sheffield University, Department of Computer Science.
If you select a character set that is based on GUK, a visual representation of a keyboard appears on the screen, which informs you about the implemented keyboard mapping. The following illustrations show the mappings of “ipa-96 (SAMPA)” and “Arabic (WINDOWS)”:
To enter a character, do one the following:
The visual representation has the layout of the standard UK keyboard. If you do not have a UK keyboard, there may be mismatches between the characters and their visual representation.
For example, the IPA character “ə;” is matched to the key “@”, i.e., in order to get “ə”, you have to type “@”. On a standard UK keyboard, the key “@” is located to the left of the key “enter” (see the illustration above). On other keyboards, however, “@” may be located on a different key. In such cases, if you press the key to the left of “enter”, you will not get the character “ə”. To get “ə”, you have to search for the location of “@” on your keyboard, and then press that key. (Note that these mismatches only arise if you use the physical keyboard, but not if you use the visual representation on the screen.)
The character set “ipa-96 (SAMPA)” can be used to enter IPA characters. However, the current version of ELAN only supports SAMPA, but not X-SAMPA. As a consequence, some of the characters that you require may not be available yet (see http://www.phon.ucl.ac.uk/home/sampa for further information).
IPA characters can be entered using the RTR input method. This method is based on the following principle: whenever you type a character, all typographically similar characters are displayed in a lookup list, as shown in the following illustration:
To select a character from the lookup list, do the following:
Do not use the mouse within the lookup window. If you do, the window will disappear.
The input of IPA characters is restricted to the official IPA-96 character set. Withdrawn or superseded characters are not supported.
The RTR-mapping is generally obvious, but please pay attention to the following features:
Table 2.4. Diacritics
key | position of diacritic in relation to character |
---|---|
“ | (quotation mark) above |
‘ | (apostrophe) in the upper right |
: | (colon) to the right |
` | (grave accent) in the lower right |
^ | (circumflex accent) under |
‘ | (tilde) on |
For example:
On Windows 2000, if you use an international keyboard, you have to type SPACE after typing the quotation mark (“) or one of the accents (‘, `, ^).
It is possible to stack diacritics, but there may be problems displaying them.
Chinese characters, both traditional and simplified, are entered using the Pinyin method. Characters are selected by starting to type Roman characters. Candidates are shown in a lookup window while the user types along. The desired character is selected with the UP and DOWN arrow keys, e.g.:
Use the following procedure:
Do not use the mouse within the lookup window. If you do, the window will disappear.
On Windows 98, you cannot display both Chinese and IPA characters.
If the options above don’t fulfil your needs (e.g. the character set is not supported or you don’t want to use the on-screen display for a large amount of annotations) you might want to look for a third-party solution. Such a tool provides a mean to remap your keyboard to the desired input character set. For details, we refer to the following programs:
Keyman (Windows), http://www.tavultesoft.com/
Ukelele (Mac), http://scripts.sil.org/ukelele
If you are using a third-party keyboarding solution like Keyman, make sure to select the default system language as input language for the tier to be edited (e.g. Dutch if your system language is set to Dutch).
In ELAN it is possible to associate an individual annotation with a data category of the ISO Data Category Repository (DCR). To do so first select an annotation. Then click Annotation in the main menu of ELAN and select Modify Annotation Data Category (alternatively press ALT+SHIFT+M). In the next window (Figure 2.94, “Local Data Category Selection: Annotation”), select a data category from the list and click on Apply to associate the selected data category to the annotation.
More information about the ISO DCR and how to use it can be found in Data categories.