| From | qiuyi_01 |
|---|---|
| Posted | 2003-07-28 02:27:06 UTC ago |
| Subject | A newbie's test of MS VR with Atmar's sample text |
Hi,
I've been playing around with the MS VR for few days, but only
tonight have I finished all the training session.
Even in the training session, I've found some irritants, like it
never recognizes my "THE" that leads first on any training passage
screen. It also has very poor accuracy whenever it comes to wh, h.
Whenever I run into "which", "had" etc, the cursor stoped moving and
it stops recognizing my dictation:I had to read several times, word
by word the whole sentence that contains "which". It also has great
diffculty with recognizing "know" or "no".
After the training, I read Itamar Even-Zoha's sample, with better
results, probably because I read faster out of frustration.
Maybe I was too slow and careful.
I use the VR in Word 2002, with WinXP. My Dell Dimension 7250 has
the following configuration:
My system specifications are a Pentium IV, 2.4GHz, 512 Meg RAM,
level 2 cache, Creative sb Live!(WDM) sound card (not on-board
chip). CA Ac-633 microphone, Swap file of ??? MB, 80 gig HD,
Windows XP home with SP-1. I use MS Word 2002 and 2 and IE 6. No
Norton Anti-Virus. Start up system resources are ? %.
Following is the result of my test:
a sample dictation with the Microsoft speech recognition
Version 6.1, a available under office 2,003 pet
Presently there has been much talk about speech recognition engines
and similar topics
On the one hand people are very skeptical about the possibility of
dictating to a machine, but on the other hand, because of the high a
misty thicket shin connected with machines in general and computers
in particular, people believe there is no difficulty in a cheating
set bold.
Unfortunately, because of the lack of popularity of to the
dedication, it in a mall native speakers of the language for which
they are speech with again nation at the kitchens available, the the
street doesn¡¯t consider it profitable in of the invest in
developing these applications and bringing them to a state where
more and more people would be able to joining and the intro or a
speaking to their computers instead of producing tax by typing.
In short, this is a vicious circle from which I see no queasy exit.
They generally. Most that first I directed into manufacturing of
products for the inflation and rich, while coal the other languages
lack considerably behind
.
Fortunately, however, most the other relevant languages have a farce
in her son¡¯s Patton, which makes the relevant products highly
functional.
I¡¯m referring mostly to Italian and Spanish but if I believe that
the two men and perhaps touch beyond to this category.
and here is Itamar Even-Zoha's original text :
A SAMPLE DICTATION WITH MICROSOFT SPEECH RECOGNITION
(Version 6.1, available under Office 2003 Beta)
Recently, there has been much talk about speech-recognition engines
and similar topics. On the one hand people are very skeptic about
the possibility of dictating to a machine, but on the other hand,
because of the high mystification connected with machines in general
and computers in particular, people believe there is no difficulty
in achieving such a goal. Unfortunately, because of the lack of
popularity of dictation, even among native speakers of the languages
for which there are speech recognition applications available, the
industry doesn¡¯t consider it profitable enough to invest in
developing these applications and bringing them to a state where
more and more people would be able to join in and enjoy speaking to
their computers instead of producing texts by typing. In short,
this is a vicious circle from which I see no easy exit. Naturally,
most efforts are directed into manufacturing of products for the
English language, while all the other languages lag considerably
behind. Fortunately, however, may so the other relevant languages
have a far simpler sounds pattern, which makes the relevant products
highly functional. I am referring mostly to Italian and Spanish,
but I believe that German and perhaps Dutch belong to this category.
Itamar Even-Zohar
[This text was dictated with Microsoft SR 2003 Beta and Telex M-560
Microphone]
