Sunday, January 17, 2016

Linking audio to text

NOTE: Despite years (decades) of trying, I have not been able to achieve this function with any word processing program other than WordPerfect for Windows.  This is not to say it can't be done.

In 1981 I did my first audio transcription of an oral history interview.  This led me to becoming a legal secretary, and finally a court reporter, all jobs which involved me listening to audio recordings and typing them into a computer-readable form.  This is known as audio-transcription.

This article not about audio transcription per se, but about linking the audio file to the transcript in such a way that a user can do a keyword search in the text of the transcript and arrive almost instantly at the section of audio and play it back.

There are several separate stages. 

Firstly, saving the audio file in small clips.

Secondly, placing links within the text file in appropriate locations.

Third, merging the audio and text files.   This is an optional step.

Software: As headlined, I have only managed to make this work with WordPerfect for Windows (WPW).  I think I have done it with versions as early as WP8.  For audio work the most basic (and free) software I have used is Audacity, although I think almost any audio editing software could be used, as long as it has the function to select portions of the audio track and save them as *.wav file.

Let's assume I have transcribed a short (5 minute) interview and have created a native WPW text file with an extension of *.wpd.

The audio file should be saved in a different directory to the original audio file to minimise the possibility of corrupting the original file.  The file is open in Audacity (or a suitable audio editing suite.) 

Size of audio clip.  The larger the audio clip, the fewer clips there will be in the text file, but it will take longer to listen, so audio size will be at the discretion of the author.  Let's assume I will use clips of 15 second duration, 4 clips per minute, so 20 clips for the entire recording.

Using the "select" function (usually this will be in the "edit" menu) highlight 15 seconds of audio.  In Audacity the cursor is placed at the 15 second time on the audio screen and Edit/ Select/ Track Start to Cursor.  This will highlight the first 15 seconds of audio.  Then export the audio (as *.wav) into a newly created folder with a number as the name.  Eg 0001.wav.  Using this format up 9,999 files can be saved in numerical order eg 0001, 0002, 0003 ... 9998, 9999.

After you have saved each 15 second clip, delete the clip (did you remember to save the original file elsewhere?)  and you will find your cursor at the beginning of the track.  Move it to 15 seconds and repeat the steps above. 

If this seems like a lot of work, it probably is.  There is no doubt software around which can automate this task, but I like to do it the "old fashioned" way. 

Having saved the files, the next step is to link the audio file to the text file.  Since we are starting at the beginning of the text file, move the insertion point to the beginning of the text and click Insert/ Sound.  This will take you to a screen "sound clips in document".  Click on the folder icon on the right which will enable you to go to the appropriate directory where the 15 second clips are stored.  Select the first clip and you will be able to number the clip eg #0001, select the appropriate audio file from the subdirectory, and decide whether you want to merge the text and audio files or keep them as separate. 

If you choose the first option, you will need to have sufficient disk space to save the entire audio file.  If you choose the second option, you will have a much smaller text file, but will need to copy the audio clips if you are going to transfer the file to another directory or computer, say if you send it to a colleague to examine.  (Your colleague will need to be using WPW.)

To see where to place the second clip, insert it at the end of the text at the end of the first, then play it and at the end you will see the next section of text at the end of the second clip etc.

Once you have inserted links into the text document at appropriate places, it is a simple matter to search through the text to find a keyword.  You will need to have the document view set to "draft" or have the WPW "reveal codes" view active.  When you arrive at the keyword(s) in the document move your cursor up towards the beginning until you reach a marker "Sound:Clip# XX".  If you click your mouse on the marker the sound file will begin to play immediately without stopping.  If you select "Insert/Sound" menu you will have the normal audio player commands, play, stop, Ffwd etc.

To find out more about high speed audio transcription check the following links.

http://proword-transcription.blogspot.com.au/
http://mostergonomickeyboard.blogspot.com.au/
http://proword-keyboardlayoutefficiency.blogspot.com.au/
http://qwertymyth.blogspot.com.au/
 https://wpvsmsword.blogspot.com/