jump to navigation

Voice to Text March 14, 2006

Posted by Michael McVey in Online Tools, Writing.
trackback

Voice to text   Yesterday I finally purchased a decent voice to text software program in the hopes that it might assist me in dictating a short novel. I will not mention the name of the software until I am satisfied it is up to the task. Several years ago, I was using a free version of the software that came bundled with my new Toshiba laptop. It had a lot of potential but I am afraid that I was not an adequate spokesperson for the software. 

After training the software to recognize my voice, I thought it was time to perform a demonstration for my students. When the day of the presentation came, I was suffering from a terrible head cold and at the time, I had no idea of the damage that would cause to the voice recognition patterns that I had built up over the last few weeks.

I began my demonstration explaining to students how wonderful the software was and how it was able to capture my words exactly and turn them into text. Most of my students at that time had never seen anything quite like it back in 1999. As you can imagine, with a head full of mucus my ability to speak clearly was severely compromised. As a result, what appeared on the screen in the way of text was almost gobbledygook. It bore no similarity to what I had spoken into the computer and as a result, my students were severely disappointed in this new software. So was I. Now, six years later, I believe the software might be up to the task and I am willing to try it.

The first thing I noticed was that it took significantly less time for me to teach the software to recognize my voice and I appreciate that. This difficulty in training the software to recognize the voice of the speaker or the user is what caused a number of teachers to become frustrated with the research project using voice to text software I was helping to conduct among a population of severely dysgraphic adolescent in a high school.

The software requires the user to read a complicated passage into the microphone while it correlates speech patterns with the words of the text. This activity posed a great deal of difficulty for my high schoolers who could not read very well. One solution we came up with was to have teachers’ assistants read short sections of the text to the students who would then speak the sections into the microphone.

Another solution was to see if it was possible to have the students memorize one sentence at a time then speak that sentence into the microphone. Both of these attempts were abandoned because of the time required of the teachers’ aides and the students, which, in hindsight, was quite shortsighted. The goal of the research was to determine whether or not students with access to this voice to text software could not only find the means with which to fairly assess what they are learning, but also to strengthen their literacy skills by providing immediate feedback on the computer screen.

The amount of time required to ensure the voice recognition software was accurately recording what it was hearing proved too much for the research study and we abandoned it. It is my hope that the newer versions of the software will provide faster and better recognition of the spoken word of the severely disk graphic students. Now it is my job to convince the professor who originally wanted to do the study that she should attempted again. I am very optimistic about my chances.

The fun thing about this software is that it is allowing me to perform my usual household tasks while composing at the same time. In fact, I dictated this entire piece while ironing clothing, albeit it was a small pile of laundry. This software has added an entire new dimension to my household cleaning routine. 

Advertisements

Comments»

1. Ricardo - April 24, 2006

Hello,

I’m very interested in experiences of people using voice-to-text software. If you have more comments on the program you are using, it’d be great if you tell about it in your blog.

I’ve never used this kind of software, but I often hear a lot of bad comments on them (lengthy training and poor accuracy). Wtih respect to the accuracy, I’m not interested in on-line processing. So, if you leave enough (off-line) processing time, do you get more accurate transcriptions?.

Well, thanks for the blog,

Ricardo.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s