Simple-ELAN

Version 1.2

The Language Archive, MPI for Psycholinguistics, Nijmegen, The Netherlands

This manual was last updated on 2017-06-06.

The latest version can be downloaded from: https://tla.mpi.nl/tools/tla-tools/elan/.


Introduction

Simple-ELAN is an annotation tool with reduced functionality and a companion tool to the annotation tool ELAN. Its main purpose is to support easy, keyboard driven segmentation and labelling of media (video and/or audio) recordings. The segmentation process is part of the transcription process: pressing the Enter key not only moves the caret to the next line but also creates a new segment, a new annotation, into which the next line of the transcript can be typed. The main purpose is to transcribe recordings on the level of scenes, to transcribe a single speaker or multiple speakers on a single layer (tier).

Internally Simple-ELAN uses the file format of ELAN so that files created in Simple-ELAN can be opened in ELAN (and vice versa) without conversion.

Getting started


After opening Simple-ELAN, go to FileNew... ( Ctrl+N ), select one video and/or a wave file. In the new transcription one tier is created and in the top-right area one "segment" is visible representing the entire duration of the media. Clicking the segment activates the text field and places the play head at the beginning of the file (if it wasn't there already). CTRL+Space can be used to start and pause playback and you can start entering text. Pressing Enter creates a new segment (annotation) that contains the current text and the cursor will move to the next (still empty) segment. The end time of the newly created segment is the current media time and this is also the start time of the next segment.

In general the Enter key splits an existing annotation or a "gap" into two, based on the position of the media play-head and the position of the caret in the text. If the play-head is at the beginning of a segment or outside the segment that is being edited, Enter will only commit the text.

A new transcript with a video, the wave view and the first annotation spanning the entire media file.

When FileOpen... ( Ctrl+O ) is used to open an existing transcription, the first top-level tier will be shown. All annotations will be shown with "empty annotations" inserted where there are gaps (so that these can be edited as well). There can only be one document open at a time in this application, so opening a file means closing the current document (if any).

FileClose ( Ctrl+W ) closes the current document without exiting the application. An empty window without a document remains open.

FileSave ( Ctrl+S ) saves the file if it is an existing file or shows a save as dialog if it is a new file.

FileSave As... ( Ctrl+Shift+S ) shows a save as dialog.

FileExit ( Ctrl+Q ) exits the application. If there is a document open and there are unsaved changes a prompt will appear whether or not the file should be saved.

The annotation view


The annotation viewer shows annotations in a vertical layout.

The annotations (and empty segments) are visualized in a vertical layout with the time line running from top-to-bottom. Some color coding is in place:

  1. a dark blue border marks the active annotation. Keyboard shortcuts that e.g. delete or merge an annotation operate on the active annotation. A blinking caret indicates whether or not the annotation has the keyboard focus.

  2. a light blue bar on the left side indicates that the selected time interval overlaps this annotation. The time selection can cover multiple annotations/segments (e.g.. when the selection has been set by dragging with the mouse in the Waveform viewer).

  3. a red marker indicates the current position of the media playhead. The location of the red rectangle in the bar to the left roughly indicates the position of media time within the annotation's interval.

  4. annotations are shown with a white background, empty segments (gaps) are painted with a light green background.

The main editing and navigation actions and their keyboard shortcuts

Note: in this section the terms "annotation" and "segment" are mostly used as equivalents


  • Split/Create an annotation - Enter.

    Creating a new annotation, a new segment, can only be done by splitting an existing annotation or an existing "gap". When starting annotation in a new document, the entire recording (from 0 to the end of the media) is considered as one "virtual" segment. New annotations can be created by splitting a segment into two. When activating an annotation (e.g.. by a mouse click) the text edit field becomes active and the media playhead jumps to the begin time of the annotation. A keyboard shortcut can be used to start playback of a fragment and text can be typed simultaneously. Pressing Enter splits the segment at the current media location, saves the text and activates the next segment.

    Creation of the first annotation: left just before hitting Enter, right just after. The current media time determines the position of the new boundary.
  • Modify the text of an annotation - Enter.

    The Enter key is also used to confirm changes to the text of an existing annotation. If the media playhead is exactly at the start or end of the annotation or even outside the interval of the annotation, Enter commits the text to the annotation and activates the next annotation.

  • Merge an annotation with the annotation before it - Control+Shift+Up.

    When an annotation is merged with the preceding annotation the resulting annotation will span from the preceding annotation's start time to the current annotation's end time and the text of both annotations will be concatenated.

    Merging an annotation with the annotation before it.
  • Merge an annotation with the next annotation following it Control+Shift+Down.

    Similar to the previous option only here the merging is done with the annotation following in the list.

  • Delete the current annotation - Ctrl+Shift+D.

    When an annotation is deleted it is replaced by a "gap" segment. In case the preceding or following segment represents a gap, the interval of the annotation is added to that gap. A gap segment can not be deleted.

    An annotation is deleted, leaving a gap, an empty space between two existing annotations.
  • Remove the text of the current annotation - Ctrl+Delete.

    This option clears the contents of the current active annotation without modifying time boundaries.

  • Modify an annotation's start or end time - Control+Enter.

    It is possible to correct the alignment of an annotation by activating the annotation, moving the media playhead to the new target start or end time outside the current interval and pressing Control+Enter. E.g. in order to change the start time of an annotation to an earlier time, activate the annotation, move the playhead to a position before the current start time but within the boundaries of the preceding annotation and use the keyboard shortcut. The end time of the preceding annotation is also updated to match the new begin time of the active annotation.

    Modifying the boundary between two annotations: on the left the media playhead is shown outside the segment of the active annotation. After the key combination the boundary jumps to the media position.
  • Enter a line break in the current annotation - Shift+Enter.

    Although it is not advised to used line breaks in annotations, it is also not prevented. Since the function of the Enter key has been re-assigned, this combination has been implemented to make it possible to insert line breaks.

  • Activate the annotation before (up) - Control+Up when a text editor is active, or Alt+Left when no text editor is active.

  • Activate the next annotation (down) - Control+Down when a text editor is active, or Alt+Right when no text editor is active.

  • Activate the annotation at the current media time - Control+Slash.

  • Cancel editing, deactivate the current annotation - Escape.

  • Play/Pause with a text editor active - Tab

  • Place the media playhead at the start of the selection - Control+[.

  • Place the media playhead at the end of the selection - Control+].

  • Place the media playhead in the center of the selection - Control+Alt+Slash.

  • See the tables below for other keyboard shortcuts

Transcription and translation, transcription of multiple speakers


Simple-ELAN is in principle designed for transcribing and segmenting on a single tier or layer. There are two options to facilitate simultaneous transcription and translation and transcription for multiple speakers. This can be done in a single annotation by using special markers (character combinations).

  • Combined transcription and translation

    When entering the transcription and a translation in the same annotation and separated by the special marker combination "//", they can later be split into two annotations, a transcription and a dependent translation annotation. The "marker" will be highlighted in the user interface with a background color.

    An annotation containing a transcription and a translation.
  • Transcription of multiple speakers

    It is possible to indicate the name or code of a speaker by inserting a special speaker ID marker in front of a part of the transcription. This marker has the form of "mmm@nnn:". The "mmm" part is optional and will become part of the tier name after export. If it is there it should contain no white spaces. The "@" is mandatory. The "nnn" part is also mandatory, should consist of at least one character and should contain no white spaces. The ":" is mandatory and terminates the marker. There can be multiple speaker markers per annotation and these markers will also be highlighted when detected. Examples of valid speaker markers are "@Sp1:", "@Sp2:", "txt@A:" etc.

    An annotation containing speaker ID markers.
  • Exporting to a "regular" ELAN .eaf file via FileExport as Regular Multi-tier EAF

    This export function creates a new file in which transcription and translation are separated and where the translation tier is a "Symbolic Association" dependent tier of the transcription tier. If speaker ID markers have been used, for each speaker a top level tier will be generated of the same type as the original tier. If there are also translation markers, for each speaker tier a translation will be created.

Other options


Resizing Viewers

  • Use the mouse to drag the split-pane divider up/down or left/right to increase/decrease the size of the corresponding Viewers.

  • Click on the up/down or left/right arrows in the split-pane dividers to hide one of the sides completely and give all the space to the other side.

Setting the tier to work on (the active tier)

  • In a new document there will only be one tier, labelled "transcription", this is the active tier.

  • If an existing ELAN .eaf file has been opened or if a new file is created based on a template, all top-level tiers (tiers without a parent) are listed in the dropdown box above the annotation viewer. Selecting a tier in this box makes it the active tier and its annotations are loaded and displayed in the viewer.

Accessing points in the media time

  • Click on any location in the Waveform or Annotation Density Viewer to jump to this point in time.

  • Click on one of the media controls to go forward/backward in time.

  • Click on the time code box and enter a time code to jump to this point in time.

  • Click on an annotation in the Annotation Viewer to jump to the start time of that annotation

Making a selection

  • Click with the mouse on any point in the Waveform Viewer, and drag it to another point to select a time interval.

  • Click on an annotation in the Annotation Viewer to make it active and to set the selection to its interval.

Activating continuous playback mode

  • By default the player stops when creating a new segment/annotation. When the option Continue playback on segmentation (Ctrl+K) is checked, the player continues to play while the next segment is activated and is ready to be typed in.

Changing the font for the annotation content

The Simple-ELAN application has no graphical interface for setting the font of the viewer. There are two alternative ways to specify the font for the viewer displaying the contents of annotations:

  • Using an ELAN template. If the template specifies a font for certain tiers, this font is used by the Simple-ELAN viewer.

  • Specifying the font in a properties file named "elan.properties", which is also supported by Simple-ELAN. One of the keys in that file is labelled "ELAN.Tiers.DefaultFontName" and if the value to the right of the "=" sign is the (full) name of a font that is installed on the system, that font is used.

    Result of having the following line in elan.properties: ELAN.Tiers.DefaultFontName=Bauhaus 93

The shortcut keys

Note: on MacOS the Command key is used instead of the CTRL key (with the exception of the combination CTRL+SPACE). And the Option key is used instead of ALT.


Table 1. File options

CTRL+SSave the current transcription
CTRL+SHIFT+SSave as… with a different name and/or in a different location
CTRL+WClose the current transcription
CTRL+QExit the application
CTRL+OOpen a document
CTRL+NCreate a new document


Table 2. Playing and navigating the media (some combinations only work if no text editor is active)

CTRL+SPACEStarts or stops the Media Player
CTRL+BGo to the beginning of the video/audio recording
CTRL+PAGE UPGo to the previous scroll view
SHIFT+LEFTGo back one second
CTRL+LEFTGo back one frame
CTRL+SHIFT+LEFTGo back one “pixel” in the Waveform Viewer
CTRL+SHIFT+RIGHTGo to the next “pixel” in the Waveform Viewer
CTRL+RIGHTGo to the next frame
SHIFT+RIGHTGo to the next second
CTRL+PAGE DOWNGo to the next scroll view
CTRL+EGo to the end of the media
CTRL+GGo to time...
CTRL+KActivate or de-activate the continuous playback mode


Table 3. Working with annotations

CTRL+DeleteDeletes the value of the selected annotation
CTRL+SHIFT+DDeletes the selected annotation
ALT+LEFTGo to previous Annotation (when no text editor is active)
ALT+RIGHTGo to next Annotation (when no text editor is active)
CTRL+UPGo annotation upward (when a text editor is active)
CTRL+DOWNGo annotation downward (when a text editor is active)
CTRL+ENTERChange the start or end time of the active annotation to match the current media time (when a text editor is active)
ESCExits the text editor without saving the changes
SHIFT+ENTERInsert a line break in the current annotation
CTRL+SHIFT+UPMerge the current annotation with the annotation before it
CTRL+SHIFT+DOWNMerge the current annotation with the annotation after it


Table 4. Working with selections

SHIFT+SPACEPlay selection
CTRL+SHIFT+SPACEPlay around selection
TABPlay selection (in the text editor)
ALT+SHIFT+CClear selection
CTRL+[Move the crosshair to the left of the selection
CTRL+]Move the crosshair to the right of the selection
CTRL+ALT+/Move the crosshair to the center of the selection
CTRL+LToggle loop mode


Table 5. General

CTRL+ZUndo
CTRL+YRedo
CTRL+HShow this Help
CTRL+=Zoom in in the Waveform viewer or increase the font size of the annotations if there is no Waveform viewer. Increase font size in the Help window.
CTRL+-Zoom out in the Waveform viewer or decrease the font size of the annotations if there is no Waveform viewer. Decrease font size in the Help window.
CTRL+0Zoom to the default zoomlevel in the Waveform viewer or set the font size of the annotations to the default value. In the Help window set the font size to the default value.


Key concepts


Annotation file (*.eaf)

An annotation file is the document that contains all the information about tiers, annotations and time alignments. EAF is a document in XML format.

Media file (*.mpg, *.mp4, *.wav etc.)

A media file contains a digitized recording. It can contain both video and audio (e.g. *.mpg) or only the audio (*.wav) data. It determines the time axis to which the annotations are linked. Visualization of the Waveform requires a *.wav as one of the linked files.

Annotation

An annotation is text (e.g. a transcription, a translation etc.) that is linked to a selected time interval of the video/audio file (e.g. the time interval corresponding to an utterance of a speaker).

Tier

A tier is a container for annotations that usually have something in common e.g. they all contain utterances of one speaker or all contain a description of a scene in the media. Annotations on the same tier can not overlap in time. In Simple-ELAN it is not possible to create new tiers but the application can read (and write) ELAN files, so it is possible to use an ELAN template e.g. with separate tiers per speaker.

By default a new transcription document in Simple-ELAN contains only one tier on which all annotations are created. A future release might support a formalized way to handle transcription + translation and/or multiple speakers per annotation.