Upgrading -- what processor needed for fast voice-to-text recognition?

S

Steven O.

I have had my current computer for about three or four years. I have
the ATX K7T266 Pro (MS-6380) motherboard, running at 900 MHz with an
AMD Athlon processor (obviously from 3 or 4 years back, I don't know
the model number). I also have 500 MHz of RAM.

The system is fine for almost everything I do, except voice
recognition. I have Dragon Naturally Speaking 7.2, and the
performance is sluggish. When I set DNS to it's best speech
recognition mode (as opposed to fastest), it lags a good 30 to 60
seconds in translating my speech. I need translation that is much,
much faster -- ideally keeping up with my speech, or at least lagging
no more than 2 to 5 seconds.

So, I think it's time to just tear out the old motherboard and put in
a new one. However, I have a few questions:

1. Instead of getting a new motherboard, would extra RAM (1 Gig)
influence the voice-to-text translation speed?

2. Assuming I go AMD, what's the minimum system I need for my
purposes? Will a fast Sempron processor do the trick, or do I need
the newest Athlon? Will, say 2000 MHz processor do the trick, or do I
need 2800 MHz? Will a 1600 bus be sufficient, or do I need the 2000?

Basically, I'm trying to be frugal with my money -- I'd rather spend
$400 to $500 on the motherboard/memory/processor, rather than $1000 --
but obviously if I'm going to upgrade I might as well get the
performance I'm seeking. Suggestions for a minimum configuration for
effective speech-to-text processing would be appreciated.

Thanks in advance for all replies.

Steve O.


"Spying On The College Of Your Choice" -- How to pick the college that is the Best Match for a high school student's needs.
www.SpyingOnTheCollegeOfYourChoice.com
 
M

Martin Markoe

Steven said:
The system is fine for almost everything I do, except voice
recognition. I have Dragon Naturally Speaking 7.2, and the
performance is sluggish. When I set DNS to it's best speech
recognition mode (as opposed to fastest), it lags a good 30 to 60
seconds in translating my speech. I need translation that is much,
much faster -- ideally keeping up with my speech, or at least lagging
no more than 2 to 5 seconds.
1. Instead of getting a new motherboard, would extra RAM (1 Gig)
influence the voice-to-text translation speed?
Adding another 512 MB of RAM will probably do the trick for Dragon
seven. Most people find when they put the slider about 2/3 of the way
towards fastest it is not detrimental to recognition accuracy

2. Assuming I go AMD, what's the minimum system I need for my
purposes? Will a fast Sempron processor do the trick, or do I need
the newest Athlon? Will, say 2000 MHz processor do the trick, or do I
need 2800 MHz? Will a 1600 bus be sufficient, or do I need the 2000?

If you're planning to use speech recognition in the future, you should
get the fastest machine possible. New versions of speech recognition
are likely to take advantage of dual core processors. If you never plan
on upgrading your speech recognition software, then any Pentium four
equivalent AMD with 1 GB of RAM will be plenty sufficient.

Marty Markoe, eMicrophones, Inc.
See us at: http://www.emicrophones.com
 
R

Richard Owlett

Martin said:
Steven O. wrote:

The system is fine for almost everything I do, except voice
recognition. I have Dragon Naturally Speaking 7.2, and the
performance is sluggish. [*SNIP*]


2. Assuming I go AMD, what's the minimum system I need for my
purposes? Will a fast Sempron processor do the trick, or do I need
the newest Athlon? Will, say 2000 MHz processor do the trick, or do I
need 2800 MHz? Will a 1600 bus be sufficient, or do I need the 2000?


If you're planning to use speech recognition in the future, you should
get the fastest machine possible. New versions of speech recognition
are likely to take advantage of dual core processors. If you never plan
on upgrading your speech recognition software, then any Pentium four
equivalent AMD with 1 GB of RAM will be plenty sufficient.

I'd raise a secondary question.
How tightly does DNS take advantage of Intel *specific* microcode?
Is there any significant loss using "generic processor"?
 
M

Michael Hawes

Richard Owlett said:
Martin said:
Steven O. wrote:

The system is fine for almost everything I do, except voice
recognition. I have Dragon Naturally Speaking 7.2, and the
performance is sluggish. [*SNIP*]


2. Assuming I go AMD, what's the minimum system I need for my
purposes? Will a fast Sempron processor do the trick, or do I need
the newest Athlon? Will, say 2000 MHz processor do the trick, or do I
need 2800 MHz? Will a 1600 bus be sufficient, or do I need the 2000?


If you're planning to use speech recognition in the future, you should
get the fastest machine possible. New versions of speech recognition
are likely to take advantage of dual core processors. If you never plan
on upgrading your speech recognition software, then any Pentium four
equivalent AMD with 1 GB of RAM will be plenty sufficient.

I'd raise a secondary question.
How tightly does DNS take advantage of Intel *specific* microcode?
Is there any significant loss using "generic processor"?

Your present system should run the program fine.Min spec for v8 is PIII
500 and 256 Mb of RAM.
See
http://www.digitalriver.com/dr/sat5...&API4=Search&API5=www.google.co.uk&CACHE_ID=0

What soundcard are you using?
Mike
 
J

John Doe

Steven O. said:
I have had my current computer for about three or four years. I
have the ATX K7T266 Pro (MS-6380) motherboard, running at 900 MHz
with an AMD Athlon processor (obviously from 3 or 4 years back, I
don't know the model number). I also have 500 MHz of RAM.

Thanks for the details. Assuming that mainboard is v2.0.

http://www.msicomputer.com/product/detail_spec/K7T266pro2.htm

What type of RAM do you have?

It says:
"Chip Integrated (2 channel S/W audio)"
I'm not sure if they are saying there is a separate audio chip or
that the audio is integrated into some standard chip.

For what it's worth, better lingo might be "Athlon running at 900
MHz" and "500 MB of RAM".
The system is fine for almost everything I do, except voice
recognition. I have Dragon Naturally Speaking 7.2, and the
performance is sluggish. When I set DNS to it's best speech
recognition mode (as opposed to fastest), it lags a good 30 to 60
seconds in translating my speech.

In WordPad, on the best speech recognition mode, when I dictate "Now
is the time for all good men to come to the aid of their country"
taking five seconds to do so, it prints within two seconds. If I set
NATSPEAK.EXE to above average priority in Windows XP, it prints in
less than one second. For command recognition, bumping up NATSPEAK's
priority while playing full-screen games makes a world of difference
here. So far, I haven't attributed any errors to raising the
program priority one notch in Windows XP.

My system:
MSI K7N2 Delta2, 200 MHz FSB (doubled for CPU and RAM)
Athlon XP 3000+
1 GB RAM
10,000 rpm hard drive
fast video card, USB sound card, USB microphone
I need translation that is much,
much faster -- ideally keeping up with my speech, or at least
lagging no more than 2 to 5 seconds.

So, I think it's time to just tear out the old motherboard and put
in a new one. However, I have a few questions:

1. Instead of getting a new motherboard, would extra RAM (1 Gig)
influence the voice-to-text translation speed?

I think at least 1 GB of RAM is useful.
2. Assuming I go AMD, what's the minimum system I need for my
purposes? Will a fast Sempron processor do the trick, or do I
need the newest Athlon? Will, say 2000 MHz processor do the
trick, or do I need 2800 MHz? Will a 1600 bus be sufficient,

I would hope so, unless that 1600 figure is misleading. I think
maybe most important is CPU power and memory bandwidth. Hard drive
speed shouldn't be overlooked (even if it seems boring).

While dictating, sometimes my system pauses for some strange reason.
Maybe faster would be better, or maybe it's much more efficiently
remedied with a system change/modification of some sort. That problem
here ain't easily understood.
or do I need the 2000?

Basically, I'm trying to be frugal with my money -- I'd rather
spend $400 to $500 on the motherboard/memory/processor, rather
than $1000 -- but obviously if I'm going to upgrade I might as
well get the performance I'm seeking. Suggestions for a minimum
configuration for effective speech-to-text processing would be
appreciated.

If you are willing to spend $500, unless you are just as interested
in home building your PC, you might consider buying a no-frills Dell
or whatever prebuilt system. You could keep your current system
intact and get Windows XP with the new system. You also get a new
power supply.

You might not get a highly upgradable system, but that doesn't
happen when replacing components in your current system either.

I'm not sure about this, but maybe ideal would be a quality
bare-bones system that is upgradable and that you can move some of
your current components into. I wouldn't know what model that might
be.

I highly recommend Windows XP, especially with a speech-recognition
interface.

Good luck.
 
J

John Doe

"Michael Hawes" <michael.hawes1remove tiscali.co.uk> wrote:

....
Your present system should run the program fine.

Min spec for v8 is PIII
500 and 256 Mb of RAM.

Don't advise people based on Minimum System Requirements.

Running NaturallySpeaking 8 in Windows XP with a 500 MHz CPU and 256
MB RAM would be persistently painful.

That might be the minimum but only if you must use a computer with
speech recognition and cannot afford a better system.
 
M

marbatis

"If I set
NATSPEAK.EXE to above average priority in Windows XP, it "

How do you do that? How do you increase natspeak.exe priority in xp?

Thanks!
 
M

Martin Markoe

How do you do that? How do you increase natspeak.exe priority in xp

If you have a fast enough computer with 1 GB of RAM, setting the
priority does not seem to really make a difference.

If you want to try this, go to task manager. The easiest way to do this
is to right click on the blank area of the taskbar that goes across the
bottom of your windows main screen. Then choose task manager. Then
choose the Processes tab. Find the file "Natspeak.exe" and right click
on it. Then click on set priority, and choose high.

Martin Markoe, eMicrophones, Inc.
See us at: http://www.emicrophones.com
 
J

John Doe

Martin Markoe said:
If you have a fast enough computer with 1 GB of RAM, setting the
priority does not seem to really make a difference.

It does matter as previously explained.
If you want to try this, go to task manager. The easiest way to do
this is to right click on the blank area of the taskbar that goes
across the bottom of your windows main screen.

The easiest way is to press Ctrl+Alt+Delete once.
Then choose task manager. Then choose the Processes tab. Find the
file "Natspeak.exe" and right click on it. Then click on set
priority, and choose high.

Do not choose high priority. The higher the priority, the more
likely the program is going to interfere with system processes.

If you want to mess with it, just choose "AboveNormal" priority and
watch for Windows/system errors that might be caused by doing so.

Otherwise, have fun.
 
M

Martin Markoe

"If I set NATSPEAK.EXE to above average priority in Windows XP, it "

Forget to mention something. If you follow my excellent directions
based on 13 years of speech recognition experience:
1. You will maximize the priority for Dragon NaturallySpeaking.
2. You will have to reset the priority every time you use Dragon
NaturallySpeaking, if you find it makes an improvement. This is because
it does not hold the high priority setting.

--
Martin Markoe, eMicrophones, Inc.
The best microphones for Speech Recognition
See us at: http://www.eMicrophones.com/index.asp
Read, "Key Steps to High Speech Recognition Accuracy" at:
http://www.emicrophones.com/docDetails.asp?DocumentID=38
 
J

John Doe

Martin Markoe said:
Forget to mention something. If you follow my excellent directions
based on 13 years of speech recognition experience:

Thirteen years of spamming the speech users groups hasn't taught
Martin Markoe about setting program priorities in Windows XP.

See the MicroSoft Developer Network (MSDN) library for more
information.

Whatever you do with your system, just be sure to always have
removable media copies of any important files.



<Snipped more bozo talk>
 
M

Martin Markoe

John said:
See the MicroSoft Developer Network (MSDN) library for more
information.

You really need to work on your communication skills. Why don't you
point to the link? Perhaps if you take your medicine as you were doing
last week, you would not be so nasty and could focus more on actually
helping people.

Sincerely,
Martin
 
R

Richard Owlett

John Doe Is ultimate spammer.

He hides behind anonymous id while others post using verifiable id.

Does he have any verifiable credentials? I doubt it!
 
J

John Doe

A bozo from the speech users group who is supposed to be a
programmer, but instead of attempting to provide clarification or
insight about setting program priorities in Windows XP, he trolls as
usual.


Path: newssvr29.news.prodigy.net!newsdbm05.news.prodigy.com!newsmst01b.news.prodigy.com!prodigy.com!newscon02.news.prodigy.com!prodigy.net!news.glorb.com!news-in.ntli.net!newsrout1-win.ntli.net!ntli.net!news.highwinds-media.com!sn-xt-sjc-02!sn-xt-sjc-11!sn-xt-sjc-07!sn-post-sjc-01!supernews.com!corp.supernews.com!not-for-mail
From: Richard Owlett <rowlett atlascomm.net>
Newsgroups: comp.speech.users,alt.comp.hardware.pc-homebuilt
Subject: Re: Upgrading -- what processor needed for fast voice-to-text recognition?
Date: Mon, 19 Jun 2006 15:51:03 -0500
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <129e3glepocbq24 corp.supernews.com>
Reply-To: rowlett atlascomm.net
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.12) Gecko/20050915
X-Accept-Language: en-us, en
MIME-Version: 1.0
References: <9h3b92lqu6n014oug2224thlp4cgb72i84 4ax.com> <Xns97E6B421845C90123456789 207.115.17.102> <Xns97E6D08039ADC0123456789 207.115.17.102> <1150727894.592921.58450 u72g2000cwu.googlegroups.com> <1150745505.225937.203950 c74g2000cwc.googlegroups.com> <Xns97E79C99ADBD10123456789 207.115.17.102>
In-Reply-To: <Xns97E79C99ADBD10123456789 207.115.17.102>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Complaints-To: abuse supernews.com
Lines: 6
Xref: prodigy.net comp.speech.users:13268 alt.comp.hardware.pc-homebuilt:468740

John Doe Is ultimate spammer.

He hides behind anonymous id while others post using verifiable id.

Does he have any verifiable credentials? I doubt it!
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top