- - - Can NS Pro do this? - comp.speech.users

Message from discussion - - - Can NS Pro do this?

Joe

View profile

More options Oct 16 2005, 7:56 pm

Newsgroups: comp.speech.users

From: Joe <joehnos...@ospamwarwick.net>

Date: Sun, 16 Oct 2005 23:56:05 GMT

Local: Sun, Oct 16 2005 7:56 pm

Subject: Re: - - - Can NS Pro do this?

  Reply to author
  |
  Forward
  | Print
  | View thread
  | Show original
  | Report this message
  | Find messages by this author
  

Martin Markoe wrote:
> Joe wrote:

>>Regarding Martin's comments concernig array microphones; he is
>>misleading you, whether from ignorance or commercial bias I don't know.
>> There are several examples of array microphone technology on the
>>market at this time. Consider all of the telematics applications out there.

> Considering you do not use a full name and do not reference your
> expertise, your comments are "misleading."

Joseph Hofffman Email address included in header.

In high technology since 1962, Univac, CDC, Alliant and Sun. I doubt
that you have any clue as to who the first three are/were.

I am in the business and my firm will me marketing products based on
array technology in the near future.

Your reference to "bending the laws of physics" was the misleading
point. You were, I have to assume, referring to the Inverse Square Law.
While that is true, perhaps you could explain array telescopes, array
radar systems or the military application of array microphones in ASW
work. These larger scale devices all use arrays to obtain higher
resolution and noise rejection.

In terms telematics which

> use constrained vocabulary speech recognition arrays sometimes work
> just fine. A small constrained vocabulary is only looking for a very
> limited number of possible responses, usually under 100. Dragon and IBM
> ViaVoice word count is over one hundred thousand. The larger the
> possibility of words to parse, the greater the necessity for untainted
> acoustic information.

Correct.
Since this thread concerns large speaker

> dependent vocabularies, you are incorrect.

It does not follow.

It would be a disservice and

> misleading if someone were to purchase an array microphone (we have
> them on our web site) whilst looking for the greatest accuracy with the
> software being discussed.

SPAM You do/did sell a low level device from Andrea. It happens to be
a fair device but is now several years behind the curve.

>>My opinion on Scansoft is that they will be slow in makinging
>>improvements. It seems they have taken a page from the Bill Gates
>>handbook and let others do the work and buy or kill them. Just my opinion.

> Could it be that Speech Recognition is a complex technology and the
> improvements you and others desire are just not feasible?

Wrong! Prima facie!

As I have

> been in the Speech Recognition field for 12 years, and have been
> invited inside, I speak with knowledge and authority, not speculation.

You MAY have been inside but I HAVE been and also speak from a position
of knowledge. More importantly I am not here touting my business. A
position which makes anything you say suspect, doesn't it.

As far as Scansoft goes, I am sure that the senior staff shared their
vision and direction and strategy for the company's development. If
they haven't, then you are also speculating. I simply put forward my
opinion, since I am not an insider, just a knowledgeable observer.

I have been watching this group for a long time and you have managed to
turn it into the eMicrophones, Inc. mail list. Almost no one with
expertise posts here any longer.

> Martin Markoe, eMicrophones, Inc.

Joe

Reply to author Forward