The thing that I learned over the past week is highly significant when it comes to Dragon NaturallySpeaking. It has to do with maintaining the effectiveness of the user profile. Now the way that Dragon NaturallySpeaking advertises their product and the way they explained their product to the consumer is a bit different from what I have just discovered.
Dragon NaturallySpeaking takes the point of view that using auto tuning will actually make your speech recognition a lot better. But what I have found is that statement is only true if your speech does not change. Meaning, if the sound of your voice or the quality of your voice does not changed hardly at all from one day to the next then, auto tuning will in fact be extremely effective.
But if your speech changes, as a result of environmental conditions or weather conditions or some other factor. Under those circumstances auto tuning is not the best option. And the reason for that is because when you're using Dragon NaturallySpeaking and you are making corrections as you are dictating every correction you make is in fact recorded. But it is in fact recorded with the quality of voice that you have at that particular moment. And that is where it becomes somewhat of an issue. Because if in fact your voice is changing as a result of whatever factor that may be present. That means that some of the corrections you are making may not in fact be consistent with regard to the type of speech or the level or the volume or any other factor that in fact is present when you actually re-train the profile.
So basically what that means is, that if your speech pattern or the quality of your speech changes in any way the best way to make Dragon more effective is not to do auto tuning. But in fact, to simply retrain the user profile. Because by doing that you are actually getting everything consistent in Dragon NaturallySpeaking, pursuant to the quality of your voice at that moment.
Now the other thing I discovered is even more important, it's almost critical. And that is that when you are actually setting up your microphone the acoustic part of setting up your microphone to see whether or not your microphone is good enough to be used with Dragon NaturallySpeaking that particular test really does not have any significant basis on the effectiveness of how you train your user profile. For example, what I'm talking about, is that, when I set up my microphone. And I go through that qualification process to see if my microphone qualifies. I always speak really loudly and really clearly to make sure the microphone passes.
But when I do my training I talked exactly like I want to talk when I do my dictations. Which of course is a much lower speech and much softer speech. Because I have a speech impediment and because I have had throat cancer. And that is critical. Because most people believe that when they are training their profile that they actually have to maintain the same volume that they had when they were qualified the microphone which is not true. And for that reason, I recommend turning off the low-volume notification in Dragon NaturallySpeaking 11.5. Because, for example, in my case, having that turned on would actually prohibit me from being able to use Dragon NaturallySpeaking at the level of IM that I am comfortable talking with and talking at as a result of having my left lung. Mostly collapsed.
And when you do that. Amazingly, Dragon NaturallySpeaking actually responds perfectly. Because I am right now using the desktop microphone. And I specifically trained it, meaning I just trained my user profile based on the fact that I now have this head cold. So my voice is literally in the basement. And yet, Dragon NaturallySpeaking is picking up every single word perfectly. And it is because I'm not using the auto tuning feature. Because when I notice imperfections in my Dragon NaturallySpeaking. I'm actually using the retraining method in order to make everything be as consistent as possible.
And that's why in the options for Dragon NaturallySpeaking, on the data tab, where it says. Incremental adaptation in general training. You should not have that box selected. Because if you do, then Dragon NaturallySpeaking will only update your profile based on just the training session and will not take into account any of the other day that you have already previously experienced were built up in Dragon NaturallySpeaking. So, by not selecting that option. When you are doing general training Dragon NaturallySpeaking will actually take into account every single thing you have said that is part of your user profile. Which of course, makes Dragon NaturallySpeaking, a great deal more effective.
These are the two things I have discovered that I think they are extremely important. Because a lot of people are under the misconception that in actuality auto tuning is the best way to go. But as I said, auto tuning only works if your voice really never changes that much. But if you are like me. Or if you're living in an area of the country where whether really does have a direct impact on your speaking ability or your sinuses or anything like that. I would highly recommend simply retraining the profile rather than using auto tuning.
But that is what I personally have found and am experiencing. And that's why I am mentioning it. Not because it is the final answer. But because it is just what I am finding works for me on my computer system.
Thank you very much for listening.Technorati Tags: System,Update,Dragon,user,product,consumer,auto,speech,recognition,statement,fact,result,factor,Under,option,correction,moment,basis,example,qualification,impediment,throat,cancer,notification,basement,word,method,data,Incremental,adaptation,session,account,misconception,area,impact,computer,Thank,corrections,dictations,imperfections,options,whether