Some people eat, sleep and chew gum, I do genealogy and write...

Tuesday, August 22, 2017

Where are we today with voice recognition software 2017

I made a reference to voice recognition software in a recent post about using the Bluetooth interface that came with my hearing aids to dictate text to my computer. I decided to expand on that topic and write about the current status of voice recognition software in general.

Not too long ago, my wife discovered that she could dictate text messages on her phone. As a result, the number of her text messages has increased exponentially. She has also found a tendency to insert commas and periods into conversations in other contexts. We have also been frequently amused by the misinterpretation of her speech by the voice recognition program built into her phone. 

About once a year, I write a post focusing on the issue of the status voice recognition software.

Despite the development of sophisticated voice recognition programs by Apple, Google and Microsoft, there is still a paucity of programs available for purchase. There are some very expensive, high-end programs for specialized applications such as medical transcriptions. My view of the program set a somewhat skewed by virtue of the fact that I am using exclusively Apple computers.

By the way, this entire post is being dictated using Dragon Dictate (Dragon Naturally Speaking for Windows) voice recognition software. Because of this fact, you might see some strange words and sentence constructions. I tried to go back through and edit the text but my editing isn't always complete. The need for careful editing usually cancels out any input speed gains. Any sudden increase in background noise or if someone in the room suddenly starts talking results in a garbled dictation.

So why continue to try to use voice recognition software? The answer is a little bit complex. I did find out a number of years ago that the Dragon Dictate program that finally gotten to the point where it produced more clean text than errors. Since that time, computers have increased in speed and the program has continued to improve. Unfortunately, the price is also increased. The reason for this is rather simple there are no other competitive products.

If I really wanted to use all of the features of the Dragon Dictate program, I could probably eliminate nearly all of my keyboard interaction with the computer. However, because there are limitations on the internal dictionary of the program and because of transcription errors, it is still necessary to occasionally use the keyboard to make corrections or add in words that the program simply cannot understand. As I noted briefly in my previous post, none of the programs seem to accommodate research or data entry, especially for genealogical research. I have yet to find a program that I would use to fill in a family tree entry for example.

All in all, the program works very well if you enunciate each word separately and distinctly. With a good microphone hooked up to a fast computer, you can dictate as fast as you can speak. But if you speak too quickly, the program will have a tendency to misinterpret some of the words more frequently.

If your typing skills are limited, I would certainly suggest that voice recognition software could be a viable alternative. I would suggest starting with the built in programs available on both Apple OS X and Microsoft Windows operating systems. If you decide you can use voice recognition, you will still have to decide if your usage would justify the cost of a dedicated program.

No comments:

Post a Comment