Not a fan of Windows Speech Recognition or the Dictate function in Word 365

 

I was hopeful. But after 3 months of use, I can say that I’m not a fan of either Windows Speech Recognition or the Dictate function in Microsoft Word. I was hoping that the two programs together could be a cheap substitute for Dragon. How wrong I was.

I’ll first start with Windows Speech Recognition. It’s better than nothing, but that’s the most favorable thing that can be said. Dictating with this app can be wildly inaccurate. The app’s accuracy is nowhere close to Dragon’s. And it is not possible to dictate quickly without a number of errors. I have to speak a lot slower than my normal rate of speech and very carefully enunciate each word. So annoying. While it is possible to train words and add them to the vocabulary in WSR, the app seems to randomly choose which words can be trained. The app’s navigation isn’t great. It frequently gets navigation commands confused with text dictation and vice versa. I found that using “MouseGrid” or “Show Numbers” commands to be decent workarounds for the app’s frequent confusion regarding the navigation commands and text dictation. An explanation of the “ MouseGrid” command can be found under the heading “Commands for using the mouse.” [It’s the first item listed under that heading.] An explanation of the “Show Numbers” command can be found under the heading “Commands for Windows and apps.” [It’s the sixth item listed under that heading.] Windows Speech Recognition is also not particularly great for dictating numbers or a combination of numbers and letters. I had a very hard time getting WSR to format my numbers correctly. I had a lot of unnecessary spaces and, often times, had the number spelled out when I wanted to dictate solely numbers.

I also experimented with the Dictate function in Microsoft Word 365. Actually, the accuracy of the Dictate function is pretty good, significantly better than WSR. However, I had issues with navigation in Word using the Dictate function. So, I tried using WSR for its navigation commands then switching back to the Dictate function for dictating text. This was way too slow. If I had the ability to use some type of analog hands-free mouse such as a head mouse for navigation, the Dictate function might have been doable for me.

Overall, I much prefer Dragon to using Windows Speech Recognition or Microsoft Word 365’s Dictate function. [That’s what I’m using to write this post.] Dictation in Dragon is significantly faster than those apps. Additionally, browsing the web is better with Dragon because I can write custom commands that quickly take me to the sites that I want. The ergonomics of editing in Dragon are better for me than in WSR or Word 365. To choose an option with Dragon, I can say “Choose [number]” which is much easier on my vocal cords than saying “[Number] OK” in WSR or Word 365. Saying the hard consonant “K” in the word “OK” multiple times starts to feel like coughing and tires my vocal cords. In spite of all the issues that I had with the speech recognition in both of those apps, I would recommend them to someone with a temporary injury such as a sprained or broken arm. Learning Dragon due to a short-term condition would be too big of a lift.

What is speech recognition software?

 

For those of you who are new to this rabbit hole of hands-free typing, speech recognition software is defined as assistive technology that allows you to control a computer with voice commands. The apps can either be third-party or built into a computer operating system. Let’s start with the two built in apps first: Windows Speech Recognition (PC) and Voice Control (Mac). Both can be used to open programs, menus, dictate text, and click objects on a computer screen by voice. These built in apps will also give some limited positioning of the mouse using the Grid commands. However, grid commands are very slow for positioning a mouse. I personally hate those commands in any type of app.

There are a lot of speech to text apps available. However, the 4 third-party apps that allow the most complete control of the computer are as follows: Dragon, Talon Voice, Dragonfly, and Serenade. There are people who use more than one of these apps depending on their workflow. Serenade and Dragonfly are open source and free. Talon is proprietary and free although you can become a patron and get beta access to new features and a VIP Slack channel. Talon and Serenade seem to be the most popular of the coding required apps. I have not used these 3 apps because I prefer not to code. So, I can’t speak from experience. I use Dragon, a more “out-of-the-box” app that does not necessarily require coding. An individual license for Dragon ranges anywhere from $150-$699 depending on the software version and type of license. A subscription to Dragon Anywhere starts at $150 per year. I paid $500 for my Dragon Professional license (version 15.61). Let’s say I had some sticker shock. Prior to my injury, I bought the cheapest software and computer that I could find. Those days are over. Between my current software, computer, and other assistive tech, I’ve definitely gone into the thousands.

*Update 4/7/23. A fabulous reader pointed out that Talon is proprietary, not open-source. Thanks Spud! I have made the correction and also per Spud, I am adding Numen, an open source Linux based speech recognition app, to this list.