Voicewriting

When I tell people that I am working toward a career change to court reporting the natural question I’m asked, “Will you be the person typing on that machine in the courtroom?”  My answer is “yes” and “no.”  I will be that person in the courtroom; however, I will not be typing on a machine – I will be doing voicewriting.  To clear up the puzzled look on their face, I always have to explain what a voicewriter does.


Voicewriters use speech recognition software to translate spoken word to text by speaking into a microphone connected to a computer that has the speech recognition software installed.  The software “listens” to the sounds and displays text almost as fast as the speaker can speak. 
    
I will spend many hours “training” the speech recognition software to effectively recognize my voice pattern.  When I communicate with others and use phrases such as “whaddya think” or “I dunno,” the person I am speaking to can easily translate and understand.  Speech recognition software will struggle to differentiate these idiosyncrasies to produce accurate text.  Good voicing technique equals good translation.
    
Mistranslates (errors in deciphering what was spoken and what the software determined was correct) can happen if the software did not understand the spoken word or if the word is not in the vocabulary of the software.  The software does part of the training automatically to distinguish variations in your voice which can lead to mistranslates.  It is impossible to speak each word exactly the same each time – inflection, emotion, illness, or the slightest variation of any kind could create a mistranslate. 
    
To ensure accuracy, the software records several pronunciations of each fraction of a syllable to make the best possible choice.  Speech recognition software uses artificial intelligence to choose the correct homonym.  The software calculates the phonetic possibilities and chooses the word with the highest statistical probability of being combined with the previous two words. 
    
I’ve also been asked why television programs do not utilize speech recognition for closed captioning.  To make text that is readable, it needs to be punctuated – periods, commas, question marks, etc.  Without punctuation, the translation would be difficult to comprehend since it would just be all words with no pause indicators.  I do not think listeners would like to hear the broadcaster say part of a sentence “comma” and then the rest of the sentence “period.”  Inserting proper punctuation is also done by the voicewriter using voice commands.
    
A successful attribute for this career is the ability to listen and repeat what is said while enunciating clearly.  This learned skill takes time and practice to master.  I should be ready for employment in this field in 6-12 months - depending on how much time I devote to practicing.  As a trained voicewriter, in addition to court reporting, I will be qualified to look for employment in fields such as closed captioning or providing services for deaf and/or hard of hearing in classrooms, seminars, meeting environments, and public events.