|
|
QUICK TIPS
Work Habits
|
Patience,
Perseverance, Persistence |
Speech
recognition is not plug-and-play.
Even though the technology has
improved so much, and accuracy
levels are much higher, it is not
perfectly accurate. Using
speech recognition effectively
requires lots of work and
practice. In the initial
stages of learning how to use speech
recognition and training your voice
files, it is important to be
patient and to make corrections as
often as needed. On an ongoing
basis, even once you have become
familiar with the program and have
learned how to use it well, you need
to persist and continue to make
corrections where you see errors of
recognition, so that you can
continue to improve your voice file.
Successful use of speech input
depends strongly on how patient you
are willing to be. |
top
Background noise
You CAN use
Naturally
Speaking in a 'noisy' environment
(within reason...) |
Currently, speech-recognition
programs together with the use of
good noise canceling microphones,
performed extremely well even in
noisy or busy environments. As long
as the "Audio settings-microphone
volume" has been run (under the
tools-accuracy menu on Dragon
NaturallySpeaking) with in the
environment, you should be able to
dictate with good levels of accuracy
even if there is a lot of ambient
noise in the environment. The only
time that loud noises will be picked
up, is when they are individual, or
excessively loud noises such as the
sound of a telephone ring close by
or otherwise a loud school bell or
overhead announcement. In general, a
regular classroom environment with
other students talking should not
cause any interference while using
speech recognition. |
top
Accuracy of
dictation and how to BUILD your voice files
|
Always correct
errors |
The only
way that your voice file will
continue to improve is if you always
make corrections while dictating.
Dictating words and allowing errors
to remain on the screen, will result
in a reduction in accuracy in the
voice file. |
|
Only SAVE voice
files once you've made corrections |
Once you
have corrected errors on the page,
so that the text appearing on the
page reflects exactly what you said,
then saving the voice file
(NaturallySpeaking-Save User Files)
will result in improvement of the
voice file each time you save it
with corrected errors on the screen.
If you have errors remaining on the
screen and then save the voice file,
this will result in a gradual
decrease in accuracy of the voice
file. On days in which you have
produced a document and have not had
the time or ability to correct all
the errors, it is suggested that you
do not save your voice files on that
day. It doesn’t matter
if there are grammar or punctuation
mistakes. The corrections that are
important for the voice file are
those that have to do with how well
the computer recognized your words. |
|
Speak very
clearly, say each word and make sure
that you say the end of every word |
Even though speech recognition is
continuous, and allows you to speak
almost normally at a good rate, it
is still highly dependent on words
being said as clearly as possible.
Any kind of mumbling or slurring of
words will result in lower levels of
accuracy and higher levels of
frustration for the user. If the
user imagines that the computer is
like a young child, or otherwise
talks as if talking to someone who
doesn't understand English very
well, then user will most likely
slow down and talk more clearly.
Your speech files grow and become
better at each and every work
session. People who become
sloppy - mumbling, making sounds
while the microphone is on end up
with bad speech files and then think
that the program isn’t working !!!
Remember - every session is a
training session for the computer.
Good speech recognition demands
consistent, clear speech |
|
Make sure the
text on the screen matches exactly
what you have said |
The most
important strategy for building your
voice files and making them stronger
over time, is to make sure that you
see exactly what you dictated. Using
the "Spell That", "Correct That" and
"Select ___" commands, you can
change the text that is shown on the
screen so that it is the same as
what you said. Only then should you
save your voice file
(NaturallySpeaking- Save User
Files). |
top
Back-ups and Saving
|
Save your:
Work and
Voice File |
There
are two things you need to save:
1. Text (the text you created while
dictating in a program such as
DragonPad or Word)
2. Voice files (remember to only
save voice files after a good
session - with good recognition and
corrected text) (NaturallySpeaking-
Save User Files) |
|
Back up voice
files |
Back up
your voice files often - onto
another place on your hard drive AND
onto an external drive / zip drive /
network. |
... |
|
|