Audiamus  Versions 1 and 2

A tool for building corpora of linked transcripts and digitised media.

Nick Thieberger

November 2009


Contents

Overview

Functions

Recent Developments

System requirements

Licence

Inputs

Outputs

Using the standalone version of Audiamus

Using the playlist


Please let me know if you have found Audiamus to be useful, and, if so, what you are using it for. Thanks.

Click below to play a video introducing features of Audiamus (best in Firefox >3.5).

Overview

Audiamus is a tool developed in the course of writing a grammatical description of South Efate. The need for a special tool arose in the absence of a simple method to work interactively with digitised ethnographic field tapes via their transcripts. It is designed with the key principles of reusability of and accessibility to the data, and with the basic premise that every example quoted in the grammar should be
provenanced to an archival source if possible. A sample workflow for using Audiamus is outlined below. It shows that media is time-aligned, then added to the Audiamus corpus from where it can be exported to Shoebox while maintaining timecodes.

Audiamus is not a transcription tool!

This guide to Audiamus is available as a pdf document here.


Functions


Audiamus instantiates the links to digitised media. It requires no segmentation of the sound/video file. Currently there is no limit to the size of the media file or the number of transcripts. Each 'card' of the current model represents a single transcript (typically a complete side of a cassette). Time-aligned transcripts, as produced for example by SoundIndex or Transcriber are the input for Audiamus. The transcripts in Audiamus are plain text and can be edited, as can the timecodes. Thus the data in Audiamus is the master copy of the transcript that is improved incrementally with use. To avoid the problem of data being locked up in proprietary formats there is a mass export function that dumps all linked text and timecodes to plain text files, or to a format the user selects (one of QT, XML, Shoebox). A concordance of all words in the dataset provides one way of navigating through the data, and there is also a normal text search function and a regular expression search (using standard PERL expressions).

 

In Version 1, a concordance of individual lexical items based on Mark Zimmerman's Free Text/ Conc / qinxr.c / Texas (inverted index) provides a point of entry to the data. In version 2.1 a concordance function is built in to Audiamus. Audiamus allows the user to select an example and to clip either the time codes or both the text and the time codes to the clipboard for pasting into a document. The timecodes are specified as follows: (audio filename, starttime, endtime) or (98002b,1413.9999, 1419.3600)

They can also be clipped in a Shoebox compatible format:

\aud 98002b
\as 1413.9999
\ae 1419.3600

Examples can be added to a playlist, for use in a presentation for example. The playlist itself can be stored for future use while another playlist is constructed. Audiamus permits the researcher to check fieldtapes with speakers by having instant access to material that needs querying.

Recent Developments,

Version 2.6 (November 2009)

Added a 'Play all' function that plays from the selected chunk forward, highlighting the text as it plays.

Version 2.5 (November 2007)

Added an export to HTML format to allow streaming access to the media. This implies that the media is located in an Annodex server (as discussed on the EthnoER pages [Note that as of April 2011 this is no longer supported as Eopas has a better way of playing media than Annodex].

Version 2.4 (May 2007)

Added an export to Transcriber and export to Audacity label format.

Version 2.3 (October 2006)

Added a repeat function to loop over selected segments. Added a variable speed slider to allow slow playback.

Version 2.2 (May 2006)

Larger font size used in interface to facilitate legibility. A problem with a persistent window in the MS Windows version has been resolved. The concordance function has been speeded up. Unicode is still problematic and so not yet supported.


System requirements

Version 1: Macintosh computer running HyperCard (known to work from sys 8 to sys 9.2). No longer supported.

Version 2: Macintosh OS X, classic, Windows, Linux (standalone built in Revolution).


Licence

Audiamus is distributed for free under the terms of the GPL licence. While the standalone version distributed here is free, you will need to purchase Revolution if you want to edit the software. I would like to hear from you if you do edit Audiamus and ask that you post any amended versions to a free download site linked to this site.


Inputs

Time-aligned transcripts in tab-delimited format: text/ starttime/ endtime.

Time-aligned transcripts in Limsi label format (exported from Transcriber) starttime/ endtime.

Audio / video files saved as Quicktime movies. Also .wav files (Version 2)

Audiamus imports files in tab-delimited or Limsi label format, and there is a mass import function that allows batches of files in either tab or Limsi format to be imported. From version 2.1, Audiamus builds a concordance of all files in the corpus, with access to the media associated with any word in context via a mouseclick.


Outputs

Export text and time codes as:

• Tab-delimited text

{SPEAKER = HT}Ipiatlak nmatu iskei,  ^t0.0000 ^t 2.3000

nmatu tiawi iskei ipiatlak tesa inru rana rato panpan. ^t 2.9201 ^t 7.8799

Go, tesa nra nen nagier karu kin ipi Ririel go karu ipi Ririal ^t 8.1200 ^t 13.5599

 

• Transcriber

Creates a valid Transcriber file to allow realigning of time segments in Transcriber if required.

 

• Audacity Label

A format that allows the timecodes and text to be read into Audacity to create labels in the Audacity file (ueful for subsequent export to segmented sound files).

 

• XML (currently using Lacito's DTD)

<?xml version="1.0" encoding="ISO-8859-1" ?>

<TEXT>

<HEADER>

<SOUNDFILE href="98003b" />

</HEADER>

<BODY>

<S id="s1"><TRANSCR>{SPEAKER = HT}Ipiatlak nmatu iskei,</TRANSCR><AUDIO start="0.0000" end="2.3000"></AUDIO></S>

<S id="s2"><TRANSCR>nmatu tiawi iskei ipiatlak tesa inru rana rato panpan.</TRANSCR><AUDIO start="2.9201" end="7.8799"></AUDIO></S>

[.....]

 

• Toolbox format (to allow glossing of texts while maintaining the time-alignment). This model is based on the EOPAS Toolbox format (described here, with a sample text, lexicon and .typ files available here.

\aud [medianame] [starttime] [endtime]

\tx [text]

 

• Quicktime format (for use in subtitling Quicktime movies)

{QTtext}{font:Geneva}{plain}{size:12}{textColor: 65535, 65535, 65535}{backColor: 0, 0, 0}{justify:center}{timeScale:1000}{width:160}{height:48}{timeStamps:absolute}{language:0}{textEncoding:0}

[00:00:00.]

{SPEAKER = HT}Ipiatlak nmatu iskei,

[00:00:02.920]

nmatu tiawi iskei ipiatlak tesa inru rana rato panpan.

[00:00:08.120]

 

Using the standalone version of Audiamus

To import a tab-delimited text file of the form 'text' tab 'start time' tab 'end time', select 'File menu/Import'. Browse until you locate your file and then select it. It will be imported and you will be prompted to type the audio file name in the field at the top left of the screen. Now click the button that says 'lock' to make it all work.

Select 'File menu/Save' to save your work. Audiamus does not save automatically.

To add another transcript, select 'File menu/New'.

To make the index pop-up menu work, you need to option-click on it. this will make it load in the audionames on each card and present them as a menu. Selecting one of them will take you to that card.

To delete a card select 'File menu/Delete'

Using the playlist

To use the playlist you need to click on the button 'go to playlist'. Each playlist is a new 'card' that you can name as you wish. A playlist is your own collection of sentences, utterances or whatever, each playable with the same time references as are in the main Audiamus view of the data, but this time they are on a line-by-line basis, not part of a whole cassette or sound file. Of course the sound is still in exactly the same place, but the reference to it is configured a little differently in the playlist.

To create a new playlist go to the 'File menu' in the playlist screen and select 'New'. Then go back to Audiamus and click on 'add to playlist' to place the selected chunk into the currently open playlist.

Future development

Ideally the transcripts referenced in Audiamus would be native XML files that would be read by Audiamus for interactive use (much as Shoebox does with standard format text files). Currently in development as part of the EthnoER project is an online browser of multiple time-aligned files, with a concordance function, based on Michel Jacobson's work at LACITO.

We have established a schema for interlinear text and for simple (one-line) transcripts called EOPAS (described in Schroeter, Ronald & Nicholas Thieberger. 2006. EOPAS, the EthnoER online representation of interlinear text. Barwick, Linda and Nicholas Thieberger. (eds.) 2006. Sustainable Data from Digital Fieldwork Sydney: Sydney University Press. 99-124.