Example: Dragon NaturallySpeaking

I’ve spent more (I think much more) than 100 hours training Dragon. The following paragraph was dictated with Dragon NaturallySpeaking and appear here without any alterations:

until recently I spent all my time making up catching up. are image in my head of what life should be and it looked something like this: On week days – wake up around seven, get dressed and to the usuals morning. You need to work, work an eight-hour day, including and with a sociable lunchbreak with work colleagues, and commute home. I’m getting home, and to the daily household chores while cooking dinner have done a, a dinner do the washing-up, watch a little TV, go to bed, read a book, good on weekendson weekends, squeezed in a visit to family, at a dinner party with fragments, he take out a movie, a couple of shopping trips want to pick up groceries and one just fond, gardening, laundry, some quality time with loved ones or even a day trip from weekend away in,

And this is what it should have been:

Until recently I spent all my time catching up. I had an image in my head of what life should be and it looked something like this: On week days – wake up around seven, get dressed and do all the other usual morning things. Commute to work, work an eight-hour day, including and with a sociable lunchbreak with work colleagues, and commute home. On getting home, do the daily household chores while cooking dinner, have dinner, do the washing-up, watch a little TV, go to bed, read a book, go to sleep. On weekends, squeeze in a visit to family, a dinner party with friends, a take-out and movie, a couple of shopping trips (one to pick up groceries and one just for fun), some gardening, laundry, quality time with loved ones or even a day trip or weekend away.

I do have a non-american accent. I don’t have speech difficulties, but do sometimes struggle a little to enunciate sufficiently. I had great expectations for speech recognition post-training and feel rather let down. Maybe it’s Dragon that isn’t up for the task, maybe it’s me that’s not up for the task for this level of training and maybe speech recognition is just not quite there yet. Regardless of the reason, it means that my computer usage is still unrealiable in real-time and require a fair amount of revision. Not exactly ideal.

Advertisements

2 Comments

Add yours →

  1. Which version do you have? If it’s 10, it might be a little buggy.

    Sometimes it depends on your system and/or processor. Before my old computer died, I was running Dragon 9 on it with only 512 MB of RAM. I wouldn’t recommend that; it took dictation excruciatingly slowly. I was able to remedy that somewhat by adjusting the speed vs. accuracy slider in the tools, but then you sacrifice one for the other. Now I’m running 9.5 on 2 GB of RAM and it likes the new system much better; it transcribes almost at the rate I speak, if not quite. I think the best thing I did, though, was to get a USB soundcard and connect my microphone to that. Both computers’ built-in soundcards were full of static, which made transcription worse.

    Sometimes selecting “require ‘click’ to select menus and controls” can help distinguish between commands and dictation without having to switch modes altogether. Hope this helps!

  2. I’m running 9.5 on 4GB of RAM and a quadcore processor. I do have a somewhat better-than-standard built-in sound card (i’ve got a gaming specced pc with improved sound and graphics cards) and assumed that it’d be adequate particularly as the system requirements list a sound card supporting 16-bit recording and my built-in sound card is a step or two above that.

    Sampling latency or noise/static could be the issue and if it is, it would’ve been nice to know that when I initially bought the software. I do have the ‘require click’ function turned on, but in an ideal world, I would have preferred it as an option rather than a requirement as it isn’t always suitable. :)

    Thanks for the tips input. Think it may be time to either upgrade to 10 or find a different program.

Comments are closed.

%d bloggers like this: