Speech Recognition Software and You
Today I and testing some speech recognition software that comes prepackaged in Windows 7. I’ve also played around with programs like Dragon NaturallySpeaking to see if I can improve my speed in which I make my posts. It’s always been difficult and often I would rather just type it and speak it. But technology is getting better.
What you can do with speech recognition is quite amazing once you’ve trained at, but you really must sit down and do some talking into the microphone and whatever you decide to do take time to do the training – lots of training! You do have to wait a moment after you turn on the microphone before you can dictate because it takes a second for this massive amount of software to get around the ones in the zeros involved in the process.
It used to be that in 20 minutes of thinking and timing, I could type about 600-700 words. That’s only 35 words per minute. I just finished writing and article that was 2300 words, and it took me 20 minutes – that’s 115 words a minute! The accuracy was quite good and in that 2300 words I had to make about 10-12 edits for words like “there” / “their”, “too / “to”, etc. Not bad – MUCH faster than I could type coherently with my finger tips. For fast typers I can see why they would like to keep typing, and typing is fun. As a matter of fact, I miss typing if I have not done so in a few days. Just touching the computer is something I would miss.
By the way, I’m not writing this article where you may think I am. I’m sure you imagine some guy sitting at his laptop, or sitting at his desktop, in a bedroom office, or a basement office. You may think this is more mobile and I am using a laptop while I am traveling, and that does happen often. But the truth is I am not in any of those pictures or examples. I am laying in bed watching the Good, the Bad and the Ugly. I’m lying in bed getting ready for a small nap after a hard days work of shoveling snow (I live in the north) and working on my websites. I’m using a small dictation tool made by Olympus.
It’s a digital voice recorder and to be exact, the model number is the WS-500M. This little unit cost me $120 and is made to take dictation. I simply speak into a microphone, preferably through a headset style as I am now, or just speaking into the microphone at the top of the unit. To be honest, I’ve tried both ways, and you want to use a headset with a boom style mic if you want any kind of accuracy at all. One of the handy things about a little digital voice recorder is that it comes equipped with a popout USB connector so you can plug your voice recorder into your laptop USB connection and navigate to your dictation files on your computer and let it run. All your audio files can be imported to your speech recognition software transcribed in text.
If for some reason you have to stop speaking and recording dictation into your voice recorder, simply press the record button once and it will pause the recording and allow you to take a small break, and in that case for me it is usually to answer a phone, speak with my wife, answer the door, go to the bathroom, or play with my dogs.
Of course it takes time at first to get your speech recognition software working smoothly. You do have to train thoroughly, and by this I mean do lots of talking. Just keep on talking and talking and talking, and eventually your speech recognition software will learn your voice so well that your accuracy levels will be stunning. One of the handy aspects of using one of these little voice recorders to do your speaking is that you can hear your voice in your earphones and you will have a better idea how clearly you are dictating. Of course it is very handy to be remotely located in a car, on the beach, or walking down the street (of course you look a little silly talking to yourself in a headset while you’re walking down the street), or doing what I am right now, and just relaxing in bed.
Now one of the things that you have to choose is which kind of speech recognition software you want. You might have a Mac computer and for a Mac operating system you need to use Mac dictate software which uses the Dragon NaturallySpeaking engine, which I believe to be the best on the market so far. But there are other speech recognition software packages coming on the market which will be giving Dragon NaturallySpeaking a run for its money. One of these is the speech recognition software that comes packaged with Windows 7 and with Vista, which are Windows-based 64-bit operating systems.
To try and compare Dragon NaturallySpeaking, and the Windows 7 onboard speech software I have used both of them in the post that you are reading right now. Both of them worked well, because I have trained them both extensively, but I will be using Dragon naturally in the future because the way it can handle the import of audio files enabling what I am doing right now by being away from a computer to do my dictation.
With Dragon NaturallySpeaking Preferred (Version 10) which I am using right now there is the ability to import WMA files, Wav files, and MP3 files. At one time Dragon NaturallySpeaking would only accept mono 16 bit Wav files, and I quickly abandoned their software because of this. Now that I can use any digital voice recorder in concert with Dragon, and along with the fact that DNS is even more accurate than it used to be five years ago, I have returned to their product. When I can more than triple my writing speed I have to take advantage of this technology. This particular website is only one of the many that I write on, and without such speed I would never be able to keep up with my quota of writing assignments each week.
I know that some people have been frustrated with speech recognition software in the past, and if you are one of those I suggest you give it another shot. Software and technology has greatly improved, and the use of 64-bit operating systems is going to make speech recognition software even more viable in the future.
There are certain things you have to get used to when you start dictating by speaking into a microphone if you want any kind of accuracy at all. It helps if you have English as your main language, and it helps when the English you speak isn’t with an accent. What they used to call non-regional diction is what I am talking about. In my case it has been very easy, because my English is very clear and I don’t have a Southern accent, or in Eastern accent, or a New York accent, or any kind of European accent. Some of the speech recognition software is so accurate now that even people who have fairly strong accents can do quite well.
If you are one of those people who has an accent, I would try using DNS, or Mac dictate. One of the things you have to do while speaking is speak clear enough and slowly enough so that the speech recognition software can accurately take down your words. There is no question that when I take my little voice recorder here downstairs and plug it into my computer that there will be edits I have to perform. Especially if you are getting tired or lazy, you will find that you may slur your words together and this creates havoc on speech recognition technology. If you have too many edits to carry out when you see all of your words populating the screen you will be discouraged and go back to typing. But I assure you it is worth it to learn how to speak clearly and concisely for future writings.
The Use of Speech Recognition for the Disabled
I hired a writer one time whose business was creating content for websites. Him and his wife both typed out articles for companies who needed content. His wife was ill and she could no longer type with any kind of comfort anymore. I suggested he try getting her some speech recognition software, and begin practicing dictation by voice. She did, and it dramatically changed her life’s work. She produces more than her husband ever can while he is simply typing the old-fashioned way. One of the advantages that comes with using speech recognition software is “the more you use it the better it gets”. The more you use your software the more accurate it gets and the faster you can speak.
One of the interesting things that I learned from using this software is that what we think we are saying is completely different than what actually comes out of our mouth. After a while speaking into a microphone and hearing your voice back into your headset you are will be surprised at how many words you think you have said clearly, when in fact those weren’t the words that came out of your mouth. I am always amazed at how much of a mumbler I have always been. My friends and family would tell you that I speak very clearly and very concisely, but I know for a fact when I hear my voice I am a mumbler. This has taken a little bit of time to get over as I have trained my speech recognition software to know my voice, but it is the only hurdle I have had.
I can just imagine if I had any thick accent of some sort or if English wasn’t my main language. Of course you can get speech recognition software for different languages. I should have stated that earlier in this article. Even with the right language software you have to be concious of accents. An accent in any language makes it difficult to get along with speech recognition software. If you have a really thick accent you will basically have to learn how to speak all over again. This is unfortunate, and it may force you to either learn to speak again or learn to type. Either way it’s just a decision you will have to make, and depending on what your business is, and what you require in the way of speed, you may not need such tools in your life at all.
If I didn’t have to write for a living, I certainly wouldn’t go to the effort of buying digital voice recorders (because I have had 3 now at and expensive of over $500) and you don’t want to purchase expensive software which can range between $100 and $350 USD, unless you really find it extremely useful.
Well I think that just about covers it, and I am getting a little bit tired, and Clint Eastwood is getting ready for the big gunfight.
Tags: Dragon NaturallySpeaking, speech recognition software, testing, Windows 7
You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.





Leave a Reply