- - - Can NS Pro do this?

Thanks for the rundown on array microphones, I will not waste my time
with them. Probably a good noise cancelling headset might be better.

Right now, I am using a Parrott VXI headset on both my PC and my Mac
with good results, but find it _very_ inconvenient not being able to
unplug that headset from its cord, at the headset end. Also, the
microphone is difficult to adjust for proper distance from the corner
of my mouth.

I really would appreciate your opinion on the present status of
wireless headsets, keeping in mind that my first priority is for
accuracy of dictated text in the very first "raw text" version of
dictated speech. (before correction)

In other words, _if_ a "wired" headset or microphone was even
slightly better as regards accuracy, I would prefer that instead of a
wireless headset.

(I do a lot of dictation of entire books, and hate the correction
operation)

That is the main reason I prefer using DNS Pro 8.1 and a PC for
dictation, rather than my main platform, which is a Mac with iListen.

The correction features seem to be more advanced with DNS Pro, and
there is less "pain" in the text correction process, at least in my
opinion.

One other thing, opinion wise of course :)

I wish Scansoft would throw more money at improving DNS Pro, such as
incorporating more Artificial Intellegence features into the handling
of speech homonyms, which neither Dragon or iListen is very good at
processing, at this moment in time.

MacSpeech's software "iListen" is making good strides in the accuracy
area, just as good as DNS Pro is IMO, so Scansoft is likely to lose
market share if they do not invest more actively in improving their
software.

Just my 2 cents worth.

Mark-

There is no good array microphone for speech recognition. This is
because they try to bend the laws of physics and it just does not work
that way. This is because the further away you are from a microphone,
less pressure from your voice will be exerted on the microphone
element. Also, the more chance for background noise to enter the
microphone pickup pattern with enough pressure to cause extraneous
words to appear. That is why you do not see array microphones as a
generally accepted speech recognition device.

--
Martin Markoe, eMicrophones, Inc.
The best microphones for Speech Recognition
See us at: #
Read, "Key Steps to High Speech Recognition Accuracy" at:
#

Windows Sound Recorder has a limitation of only recording for 60
seconds. This limits its usefulness beyond a good way to test one's
microphone and sound card.

--
Martin Markoe, eMicrophones, Inc.
The best microphones for Speech Recognition
See us at: #
Read, "Key Steps to High Speech Recognition Accuracy" at:
#

Yes, I thought of that approach, however I hesitate to do it that way
for several reasons.

1) Recorder introduces another "layer" where its
(minor) distortion can enter the picture, as contrasted
with just dictating directly into the computer.

2) Present day digital recorders supported by Scansoft
are not as tiny and sexy as the older Sony model ICD-MS1
that they used to support years ago. I could literally put
that excellent recorder in my shirt pocket and have it
"disappear" there without hardly a trace. Quality and
sensitivety were excellent.

I have not really looked at all the features of the most recent version
of Dragon Naturally Speaking Pro, to see if what I want to do is
supported or not.

I know NS Pro supports delayed correction, and I _think_ it also
supports delayed conversion of audio input to text - - - but am not
certain about that.

Think I will try a quick email to Scansoft tech' support to try to
resolve that issue, one way or the other

You bet, happy to do that.

If that feature is supported, it means that we could use a tiny
"pocket" PC and a good array microphone to record the initial wave
audio file - - - then later move the wave file into a bigger PC for
processing and correction.

Should result in higher quality results, than the recorder approach.

Mark-

I think you could just dictate into sound recorder. I'm not sure
whether the speech recognition can handle input from a wave file,
but it probably can since it is able to translate from a recorder.
Speaking of... maybe you could use a compatible voice recorder.

Please report back on how things go, if you actually do any of the
experimentation you have talked about.